(DEPRECATED) Apache Flink Mailing List archive.

Checkpointing under backpressure

Classic

List

Threaded

31 messages Options

Thomas Weise

Checkpointing under backpressure

Hi,

One of the major operational difficulties we observe with Flink are
checkpoint timeouts under backpressure. I'm looking for both confirmation
of my understanding of the current behavior as well as pointers for future
improvement work:

Prior to introduction of credit based flow control in the network stack [1]
[2], checkpoint barriers would back up with the data for all logical
channels due to TCP backpressure. Since Flink 1.5, the buffers are
controlled per channel, and checkpoint barriers are only held back for
channels that have backpressure, while others can continue processing
normally. However, checkpoint barriers still cannot "overtake data",
therefore checkpoint alignment remains affected for the channel with
backpressure, with the potential for slow checkpointing and timeouts.
Albeit the delay of barriers would be capped by the maximum in-transit
buffers per channel, resulting in an improvement compared to previous
versions of Flink. Also, the backpressure based checkpoint alignment can
help the barrier advance faster on the receiver side (by suspending
channels that have already delivered the barrier). Is that accurate as of
Flink 1.8?

What appears to be missing to completely unblock checkpointing is a
mechanism for checkpoints to overtake the data. That would help in
situations where the processing itself is the bottleneck and prioritization
in the network stack alone cannot address the barrier delay. Was there any
related discussion? One possible solution would be to drain incoming data
till the barrier and make it part of the checkpoint instead of processing
it. This is somewhat related to asynchronous processing, but I'm thinking
more of a solution that is automated in the Flink runtime for the
backpressure scenario only.

Thanks,
Thomas

[1] https://flink.apache.org/2019/06/05/flink-network-stack.html
[2]
https://docs.google.com/document/d/1chTOuOqe0sBsjldA_r-wXYeSIhU2zRGpUaTaik7QZ84/edit#heading=h.pjh6mv7m2hjn

Zhijiang(wangzhijiang999)

Re: Checkpointing under backpressure

Hi Thomas,

Thanks for proposing this concern. The barrier alignment takes long time in backpressure case which could cause several problems:
1. Checkpoint timeout as you mentioned.
2. The recovery cost is high once failover, because much data needs to be replayed.
3. The delay for commit-based sink is high in exactly-once.

For credit-based flow control from release-1.5, the amount of in-flighting buffers before barrier alignment is reduced, so we could get a bit
benefits from speeding checkpoint aspect.

In release-1.8, I guess we did not suspend the channels which already received the barrier in practice. But actually we ever did the similar thing
to speed barrier alighment before. I am not quite sure that release-1.8 covers this feature. There were some relevant discussions under jira [1].

For release-1.10, the community is now discussing the feature of unaligned checkpoint which is mainly for resolving above concerns. The basic idea
is to make barrier overtakes the output/input buffer queue to speed alignment, and snapshot the input/output buffers as part of checkpoint state. The
details have not confirmed yet and is still under discussion. Wish we could make some improvments for the release-1.10.

[1] https://issues.apache.org/jira/browse/FLINK-8523

Best,
Zhijiang
------------------------------------------------------------------
From:Thomas Weise <[hidden email]>
Send Time:2019年8月12日(星期一) 21:38
To:dev <[hidden email]>
Subject:Checkpointing under backpressure

Hi,

One of the major operational difficulties we observe with Flink are
checkpoint timeouts under backpressure. I'm looking for both confirmation
of my understanding of the current behavior as well as pointers for future
improvement work:

Prior to introduction of credit based flow control in the network stack [1]
[2], checkpoint barriers would back up with the data for all logical
channels due to TCP backpressure. Since Flink 1.5, the buffers are
controlled per channel, and checkpoint barriers are only held back for
channels that have backpressure, while others can continue processing
normally. However, checkpoint barriers still cannot "overtake data",
therefore checkpoint alignment remains affected for the channel with
backpressure, with the potential for slow checkpointing and timeouts.
Albeit the delay of barriers would be capped by the maximum in-transit
buffers per channel, resulting in an improvement compared to previous
versions of Flink. Also, the backpressure based checkpoint alignment can
help the barrier advance faster on the receiver side (by suspending
channels that have already delivered the barrier). Is that accurate as of
Flink 1.8?

What appears to be missing to completely unblock checkpointing is a
mechanism for checkpoints to overtake the data. That would help in
situations where the processing itself is the bottleneck and prioritization
in the network stack alone cannot address the barrier delay. Was there any
related discussion? One possible solution would be to drain incoming data
till the barrier and make it part of the checkpoint instead of processing
it. This is somewhat related to asynchronous processing, but I'm thinking
more of a solution that is automated in the Flink runtime for the
backpressure scenario only.

Thanks,
Thomas

[1] https://flink.apache.org/2019/06/05/flink-network-stack.html
[2]
https://docs.google.com/document/d/1chTOuOqe0sBsjldA_r-wXYeSIhU2zRGpUaTaik7QZ84/edit#heading=h.pjh6mv7m2hjn

Piotr Nowojski-3

Re: Checkpointing under backpressure

Hi Thomas,

As Zhijiang has responded, we are now in the process of discussing how to address this issue and one of the solution that we are discussing is exactly what you are proposing: checkpoint barriers overtaking the in flight data and make the in flight data part of the checkpoint.

If everything works well, we will be able to present result of our discussions on the dev mailing list soon.

Piotrek

> On 12 Aug 2019, at 23:23, zhijiang <[hidden email]> wrote:
>
> Hi Thomas,
>
> Thanks for proposing this concern. The barrier alignment takes long time in backpressure case which could cause several problems:
> 1. Checkpoint timeout as you mentioned.
> 2. The recovery cost is high once failover, because much data needs to be replayed.
> 3. The delay for commit-based sink is high in exactly-once.
>
> For credit-based flow control from release-1.5, the amount of in-flighting buffers before barrier alignment is reduced, so we could get a bit
> benefits from speeding checkpoint aspect.
>
> In release-1.8, I guess we did not suspend the channels which already received the barrier in practice. But actually we ever did the similar thing
> to speed barrier alighment before. I am not quite sure that release-1.8 covers this feature. There were some relevant discussions under jira [1].
>
> For release-1.10, the community is now discussing the feature of unaligned checkpoint which is mainly for resolving above concerns. The basic idea
> is to make barrier overtakes the output/input buffer queue to speed alignment, and snapshot the input/output buffers as part of checkpoint state. The
> details have not confirmed yet and is still under discussion. Wish we could make some improvments for the release-1.10.
>
> [1] https://issues.apache.org/jira/browse/FLINK-8523
>
> Best,
> Zhijiang
> ------------------------------------------------------------------
> From:Thomas Weise <[hidden email]>
> Send Time:2019年8月12日(星期一) 21:38
> To:dev <[hidden email]>
> Subject:Checkpointing under backpressure
>
> Hi,
>
> One of the major operational difficulties we observe with Flink are
> checkpoint timeouts under backpressure. I'm looking for both confirmation
> of my understanding of the current behavior as well as pointers for future
> improvement work:
>
> Prior to introduction of credit based flow control in the network stack [1]
> [2], checkpoint barriers would back up with the data for all logical
> channels due to TCP backpressure. Since Flink 1.5, the buffers are
> controlled per channel, and checkpoint barriers are only held back for
> channels that have backpressure, while others can continue processing
> normally. However, checkpoint barriers still cannot "overtake data",
> therefore checkpoint alignment remains affected for the channel with
> backpressure, with the potential for slow checkpointing and timeouts.
> Albeit the delay of barriers would be capped by the maximum in-transit
> buffers per channel, resulting in an improvement compared to previous
> versions of Flink. Also, the backpressure based checkpoint alignment can
> help the barrier advance faster on the receiver side (by suspending
> channels that have already delivered the barrier). Is that accurate as of
> Flink 1.8?
>
> What appears to be missing to completely unblock checkpointing is a
> mechanism for checkpoints to overtake the data. That would help in
> situations where the processing itself is the bottleneck and prioritization
> in the network stack alone cannot address the barrier delay. Was there any
> related discussion? One possible solution would be to drain incoming data
> till the barrier and make it part of the checkpoint instead of processing
> it. This is somewhat related to asynchronous processing, but I'm thinking
> more of a solution that is automated in the Flink runtime for the
> backpressure scenario only.
>
> Thanks,
> Thomas
>
> [1] https://flink.apache.org/2019/06/05/flink-network-stack.html
> [2]
> https://docs.google.com/document/d/1chTOuOqe0sBsjldA_r-wXYeSIhU2zRGpUaTaik7QZ84/edit#heading=h.pjh6mv7m2hjn
>

Paris Carbone-3

Re: Checkpointing under backpressure

Interesting problem! Thanks for bringing it up Thomas.

Ignore/Correct me if I am wrong but I believe Chandy-Lamport snapshots [1] would help out solve this problem more elegantly without sacrificing correctness.
- They do not need alignment, only (async) logging for in-flight records between the time the first barrier is processed until the last barrier arrives in a task.
- They work fine for failure recovery as long as logged records are replayed on startup.

Flink’s “alligned” savepoints would probably be still necessary for transactional sink commits + any sort of reconfiguration (e.g., rescaling, updating the logic of operators to evolve an application etc.).

I don’t completely understand the “overtaking” approach but if you have a concrete definition I would be happy to check it out and help if I can!
Mind that Chandy-Lamport essentially does this by logging things in pending channels in a task snapshot before the barrier arrives.

-Paris

[1] https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm <https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm>

> On 13 Aug 2019, at 10:27, Piotr Nowojski <[hidden email]> wrote:
>
> Hi Thomas,
>
> As Zhijiang has responded, we are now in the process of discussing how to address this issue and one of the solution that we are discussing is exactly what you are proposing: checkpoint barriers overtaking the in flight data and make the in flight data part of the checkpoint.
>
> If everything works well, we will be able to present result of our discussions on the dev mailing list soon.
>
> Piotrek
>
>> On 12 Aug 2019, at 23:23, zhijiang <[hidden email]> wrote:
>>
>> Hi Thomas,
>>
>> Thanks for proposing this concern. The barrier alignment takes long time in backpressure case which could cause several problems:
>> 1. Checkpoint timeout as you mentioned.
>> 2. The recovery cost is high once failover, because much data needs to be replayed.
>> 3. The delay for commit-based sink is high in exactly-once.
>>
>> For credit-based flow control from release-1.5, the amount of in-flighting buffers before barrier alignment is reduced, so we could get a bit
>> benefits from speeding checkpoint aspect.
>>
>> In release-1.8, I guess we did not suspend the channels which already received the barrier in practice. But actually we ever did the similar thing
>> to speed barrier alighment before. I am not quite sure that release-1.8 covers this feature. There were some relevant discussions under jira [1].
>>
>> For release-1.10, the community is now discussing the feature of unaligned checkpoint which is mainly for resolving above concerns. The basic idea
>> is to make barrier overtakes the output/input buffer queue to speed alignment, and snapshot the input/output buffers as part of checkpoint state. The
>> details have not confirmed yet and is still under discussion. Wish we could make some improvments for the release-1.10.
>>
>> [1] https://issues.apache.org/jira/browse/FLINK-8523
>>
>> Best,
>> Zhijiang
>> ------------------------------------------------------------------
>> From:Thomas Weise <[hidden email]>
>> Send Time:2019年8月12日(星期一) 21:38
>> To:dev <[hidden email]>
>> Subject:Checkpointing under backpressure
>>
>> Hi,
>>
>> One of the major operational difficulties we observe with Flink are
>> checkpoint timeouts under backpressure. I'm looking for both confirmation
>> of my understanding of the current behavior as well as pointers for future
>> improvement work:
>>
>> Prior to introduction of credit based flow control in the network stack [1]
>> [2], checkpoint barriers would back up with the data for all logical
>> channels due to TCP backpressure. Since Flink 1.5, the buffers are
>> controlled per channel, and checkpoint barriers are only held back for
>> channels that have backpressure, while others can continue processing
>> normally. However, checkpoint barriers still cannot "overtake data",
>> therefore checkpoint alignment remains affected for the channel with
>> backpressure, with the potential for slow checkpointing and timeouts.
>> Albeit the delay of barriers would be capped by the maximum in-transit
>> buffers per channel, resulting in an improvement compared to previous
>> versions of Flink. Also, the backpressure based checkpoint alignment can
>> help the barrier advance faster on the receiver side (by suspending
>> channels that have already delivered the barrier). Is that accurate as of
>> Flink 1.8?
>>
>> What appears to be missing to completely unblock checkpointing is a
>> mechanism for checkpoints to overtake the data. That would help in
>> situations where the processing itself is the bottleneck and prioritization
>> in the network stack alone cannot address the barrier delay. Was there any
>> related discussion? One possible solution would be to drain incoming data
>> till the barrier and make it part of the checkpoint instead of processing
>> it. This is somewhat related to asynchronous processing, but I'm thinking
>> more of a solution that is automated in the Flink runtime for the
>> backpressure scenario only.
>>
>> Thanks,
>> Thomas
>>
>> [1] https://flink.apache.org/2019/06/05/flink-network-stack.html
>> [2]
>> https://docs.google.com/document/d/1chTOuOqe0sBsjldA_r-wXYeSIhU2zRGpUaTaik7QZ84/edit#heading=h.pjh6mv7m2hjn
>>
>

Piotr Nowojski-3

Re: Checkpointing under backpressure

Thanks for the input. Regarding the Chandy-Lamport snapshots don’t you still have to wait for the “checkpoint barrier” to arrive in order to know when have you already received all possible messages from the upstream tasks/operators? So instead of processing the “in flight” messages (as the Flink is doing currently), you are sending them to an “observer”?

In that case, that’s sounds similar to “checkpoint barriers overtaking in flight records” (aka unaligned checkpoints). Just for us, the observer is a snapshot state.

Piotrek

> On 13 Aug 2019, at 13:14, Paris Carbone <[hidden email]> wrote:
>
> Interesting problem! Thanks for bringing it up Thomas.
>
> Ignore/Correct me if I am wrong but I believe Chandy-Lamport snapshots [1] would help out solve this problem more elegantly without sacrificing correctness.
> - They do not need alignment, only (async) logging for in-flight records between the time the first barrier is processed until the last barrier arrives in a task.
> - They work fine for failure recovery as long as logged records are replayed on startup.
>
> Flink’s “alligned” savepoints would probably be still necessary for transactional sink commits + any sort of reconfiguration (e.g., rescaling, updating the logic of operators to evolve an application etc.).
>
> I don’t completely understand the “overtaking” approach but if you have a concrete definition I would be happy to check it out and help if I can!
> Mind that Chandy-Lamport essentially does this by logging things in pending channels in a task snapshot before the barrier arrives.
>
> -Paris
>
> [1] https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm <https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm>
>
>> On 13 Aug 2019, at 10:27, Piotr Nowojski <[hidden email]> wrote:
>>
>> Hi Thomas,
>>
>> As Zhijiang has responded, we are now in the process of discussing how to address this issue and one of the solution that we are discussing is exactly what you are proposing: checkpoint barriers overtaking the in flight data and make the in flight data part of the checkpoint.
>>
>> If everything works well, we will be able to present result of our discussions on the dev mailing list soon.
>>
>> Piotrek
>>
>>> On 12 Aug 2019, at 23:23, zhijiang <[hidden email]> wrote:
>>>
>>> Hi Thomas,
>>>
>>> Thanks for proposing this concern. The barrier alignment takes long time in backpressure case which could cause several problems:
>>> 1. Checkpoint timeout as you mentioned.
>>> 2. The recovery cost is high once failover, because much data needs to be replayed.
>>> 3. The delay for commit-based sink is high in exactly-once.
>>>
>>> For credit-based flow control from release-1.5, the amount of in-flighting buffers before barrier alignment is reduced, so we could get a bit
>>> benefits from speeding checkpoint aspect.
>>>
>>> In release-1.8, I guess we did not suspend the channels which already received the barrier in practice. But actually we ever did the similar thing
>>> to speed barrier alighment before. I am not quite sure that release-1.8 covers this feature. There were some relevant discussions under jira [1].
>>>
>>> For release-1.10, the community is now discussing the feature of unaligned checkpoint which is mainly for resolving above concerns. The basic idea
>>> is to make barrier overtakes the output/input buffer queue to speed alignment, and snapshot the input/output buffers as part of checkpoint state. The
>>> details have not confirmed yet and is still under discussion. Wish we could make some improvments for the release-1.10.
>>>
>>> [1] https://issues.apache.org/jira/browse/FLINK-8523
>>>
>>> Best,
>>> Zhijiang
>>> ------------------------------------------------------------------
>>> From:Thomas Weise <[hidden email]>
>>> Send Time:2019年8月12日(星期一) 21:38
>>> To:dev <[hidden email]>
>>> Subject:Checkpointing under backpressure
>>>
>>> Hi,
>>>
>>> One of the major operational difficulties we observe with Flink are
>>> checkpoint timeouts under backpressure. I'm looking for both confirmation
>>> of my understanding of the current behavior as well as pointers for future
>>> improvement work:
>>>
>>> Prior to introduction of credit based flow control in the network stack [1]
>>> [2], checkpoint barriers would back up with the data for all logical
>>> channels due to TCP backpressure. Since Flink 1.5, the buffers are
>>> controlled per channel, and checkpoint barriers are only held back for
>>> channels that have backpressure, while others can continue processing
>>> normally. However, checkpoint barriers still cannot "overtake data",
>>> therefore checkpoint alignment remains affected for the channel with
>>> backpressure, with the potential for slow checkpointing and timeouts.
>>> Albeit the delay of barriers would be capped by the maximum in-transit
>>> buffers per channel, resulting in an improvement compared to previous
>>> versions of Flink. Also, the backpressure based checkpoint alignment can
>>> help the barrier advance faster on the receiver side (by suspending
>>> channels that have already delivered the barrier). Is that accurate as of
>>> Flink 1.8?
>>>
>>> What appears to be missing to completely unblock checkpointing is a
>>> mechanism for checkpoints to overtake the data. That would help in
>>> situations where the processing itself is the bottleneck and prioritization
>>> in the network stack alone cannot address the barrier delay. Was there any
>>> related discussion? One possible solution would be to drain incoming data
>>> till the barrier and make it part of the checkpoint instead of processing
>>> it. This is somewhat related to asynchronous processing, but I'm thinking
>>> more of a solution that is automated in the Flink runtime for the
>>> backpressure scenario only.
>>>
>>> Thanks,
>>> Thomas
>>>
>>> [1] https://flink.apache.org/2019/06/05/flink-network-stack.html
>>> [2]
>>> https://docs.google.com/document/d/1chTOuOqe0sBsjldA_r-wXYeSIhU2zRGpUaTaik7QZ84/edit#heading=h.pjh6mv7m2hjn
>>>
>>
>

Paris Carbone-3

Re: Checkpointing under backpressure

yes! It’s quite similar I think. Though mind that the devil is in the details, i.e., the temporal order actions are taken.

To clarify, let us say you have a task T with two input channels I1 and I2.
The Chandy Lamport execution flow is the following:

1) T receives barrier from I1 and...
2) ...the following three actions happen atomically
I ) T snapshots its state T*
II) T forwards marker to its outputs
III) T starts logging all events of I2 (only) into a buffer M*
- Also notice here that T does NOT block I1 as it does in aligned snapshots -
3) Eventually T receives barrier from I2 and stops recording events. Its asynchronously captured snapshot is now complete: {T*,M*}.
Upon recovery all messages of M* should be replayed in FIFO order.

With this approach alignment does not create a deadlock situation since anyway 2.II happens asynchronously and messages can be logged as well asynchronously during the process of the snapshot. If there is back-pressure in a pipeline the cause is most probably not this algorithm.

Back to your observation, the answer : yes and no. In your network model, I can see the logic of “logging” and “committing” a final snapshot being provided by the channel implementation. However, do mind that the first barrier always needs to go “all the way” to initiate the Chandy Lamport algorithm logic.

The above flow has been proven using temporal logic in my phd thesis in case you are interested about the proof.
I hope this helps a little clarifying things. Let me know if there is any confusing point to disambiguate. I would be more than happy to help if I can.

Paris

> On 13 Aug 2019, at 13:28, Piotr Nowojski <[hidden email]> wrote:
>
> Thanks for the input. Regarding the Chandy-Lamport snapshots don’t you still have to wait for the “checkpoint barrier” to arrive in order to know when have you already received all possible messages from the upstream tasks/operators? So instead of processing the “in flight” messages (as the Flink is doing currently), you are sending them to an “observer”?
>
> In that case, that’s sounds similar to “checkpoint barriers overtaking in flight records” (aka unaligned checkpoints). Just for us, the observer is a snapshot state.
>
> Piotrek
>
>> On 13 Aug 2019, at 13:14, Paris Carbone <[hidden email]> wrote:
>>
>> Interesting problem! Thanks for bringing it up Thomas.
>>
>> Ignore/Correct me if I am wrong but I believe Chandy-Lamport snapshots [1] would help out solve this problem more elegantly without sacrificing correctness.
>> - They do not need alignment, only (async) logging for in-flight records between the time the first barrier is processed until the last barrier arrives in a task.
>> - They work fine for failure recovery as long as logged records are replayed on startup.
>>
>> Flink’s “alligned” savepoints would probably be still necessary for transactional sink commits + any sort of reconfiguration (e.g., rescaling, updating the logic of operators to evolve an application etc.).
>>
>> I don’t completely understand the “overtaking” approach but if you have a concrete definition I would be happy to check it out and help if I can!
>> Mind that Chandy-Lamport essentially does this by logging things in pending channels in a task snapshot before the barrier arrives.
>>
>> -Paris
>>
>> [1] https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm <https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm>
>>
>>> On 13 Aug 2019, at 10:27, Piotr Nowojski <[hidden email]> wrote:
>>>
>>> Hi Thomas,
>>>
>>> As Zhijiang has responded, we are now in the process of discussing how to address this issue and one of the solution that we are discussing is exactly what you are proposing: checkpoint barriers overtaking the in flight data and make the in flight data part of the checkpoint.
>>>
>>> If everything works well, we will be able to present result of our discussions on the dev mailing list soon.
>>>
>>> Piotrek
>>>
>>>> On 12 Aug 2019, at 23:23, zhijiang <[hidden email]> wrote:
>>>>
>>>> Hi Thomas,
>>>>
>>>> Thanks for proposing this concern. The barrier alignment takes long time in backpressure case which could cause several problems:
>>>> 1. Checkpoint timeout as you mentioned.
>>>> 2. The recovery cost is high once failover, because much data needs to be replayed.
>>>> 3. The delay for commit-based sink is high in exactly-once.
>>>>
>>>> For credit-based flow control from release-1.5, the amount of in-flighting buffers before barrier alignment is reduced, so we could get a bit
>>>> benefits from speeding checkpoint aspect.
>>>>
>>>> In release-1.8, I guess we did not suspend the channels which already received the barrier in practice. But actually we ever did the similar thing
>>>> to speed barrier alighment before. I am not quite sure that release-1.8 covers this feature. There were some relevant discussions under jira [1].
>>>>
>>>> For release-1.10, the community is now discussing the feature of unaligned checkpoint which is mainly for resolving above concerns. The basic idea
>>>> is to make barrier overtakes the output/input buffer queue to speed alignment, and snapshot the input/output buffers as part of checkpoint state. The
>>>> details have not confirmed yet and is still under discussion. Wish we could make some improvments for the release-1.10.
>>>>
>>>> [1] https://issues.apache.org/jira/browse/FLINK-8523
>>>>
>>>> Best,
>>>> Zhijiang
>>>> ------------------------------------------------------------------
>>>> From:Thomas Weise <[hidden email]>
>>>> Send Time:2019年8月12日(星期一) 21:38
>>>> To:dev <[hidden email]>
>>>> Subject:Checkpointing under backpressure
>>>>
>>>> Hi,
>>>>
>>>> One of the major operational difficulties we observe with Flink are
>>>> checkpoint timeouts under backpressure. I'm looking for both confirmation
>>>> of my understanding of the current behavior as well as pointers for future
>>>> improvement work:
>>>>
>>>> Prior to introduction of credit based flow control in the network stack [1]
>>>> [2], checkpoint barriers would back up with the data for all logical
>>>> channels due to TCP backpressure. Since Flink 1.5, the buffers are
>>>> controlled per channel, and checkpoint barriers are only held back for
>>>> channels that have backpressure, while others can continue processing
>>>> normally. However, checkpoint barriers still cannot "overtake data",
>>>> therefore checkpoint alignment remains affected for the channel with
>>>> backpressure, with the potential for slow checkpointing and timeouts.
>>>> Albeit the delay of barriers would be capped by the maximum in-transit
>>>> buffers per channel, resulting in an improvement compared to previous
>>>> versions of Flink. Also, the backpressure based checkpoint alignment can
>>>> help the barrier advance faster on the receiver side (by suspending
>>>> channels that have already delivered the barrier). Is that accurate as of
>>>> Flink 1.8?
>>>>
>>>> What appears to be missing to completely unblock checkpointing is a
>>>> mechanism for checkpoints to overtake the data. That would help in
>>>> situations where the processing itself is the bottleneck and prioritization
>>>> in the network stack alone cannot address the barrier delay. Was there any
>>>> related discussion? One possible solution would be to drain incoming data
>>>> till the barrier and make it part of the checkpoint instead of processing
>>>> it. This is somewhat related to asynchronous processing, but I'm thinking
>>>> more of a solution that is automated in the Flink runtime for the
>>>> backpressure scenario only.
>>>>
>>>> Thanks,
>>>> Thomas
>>>>
>>>> [1] https://flink.apache.org/2019/06/05/flink-network-stack.html
>>>> [2]
>>>> https://docs.google.com/document/d/1chTOuOqe0sBsjldA_r-wXYeSIhU2zRGpUaTaik7QZ84/edit#heading=h.pjh6mv7m2hjn
>>>>
>>>
>>
>

Zhijiang(wangzhijiang999)

Re: Checkpointing under backpressure

Hi Paris,

Thanks for the detailed sharing. And I think it is very similar with the way of overtaking we proposed before.

There are some tiny difference:
The way of overtaking might need to snapshot all the input/output queues. Chandy Lamport seems only need to snaphost (n-1) input channels after the first barrier arrives, which might reduce the state sizea bit. But normally there should be less buffers for the first input channel with barrier.
The output barrier still follows with regular data stream in Chandy Lamport, the same way as current flink. For overtaking way, we need to pay extra efforts to make barrier transport firstly before outque queue on upstream side, and change the way of barrier alignment based on receiving instead of current reading on downstream side.
In the backpressure caused by data skew, the first barrier in almost empty input channel should arrive much eariler than the last heavy load input channel, so the Chandy Lamport could benefit well. But for the case of all balanced heavy load input channels, I mean the first arrived barrier might still take much time, then the overtaking way could still fit well to speed up checkpoint.
Anyway, your proposed suggestion is helpful on my side, especially considering some implementation details .

Best,
Zhijiang
------------------------------------------------------------------
From:Paris Carbone <[hidden email]>
Send Time:2019年8月13日(星期二) 14:03
To:dev <[hidden email]>
Cc:zhijiang <[hidden email]>
Subject:Re: Checkpointing under backpressure

yes! It’s quite similar I think. Though mind that the devil is in the details, i.e., the temporal order actions are taken.

To clarify, let us say you have a task T with two input channels I1 and I2.
The Chandy Lamport execution flow is the following:

1) T receives barrier from I1 and...
2) ...the following three actions happen atomically
I ) T snapshots its state T*
II) T forwards marker to its outputs
III) T starts logging all events of I2 (only) into a buffer M*
- Also notice here that T does NOT block I1 as it does in aligned snapshots -
3) Eventually T receives barrier from I2 and stops recording events. Its asynchronously captured snapshot is now complete: {T*,M*}.
Upon recovery all messages of M* should be replayed in FIFO order.

With this approach alignment does not create a deadlock situation since anyway 2.II happens asynchronously and messages can be logged as well asynchronously during the process of the snapshot. If there is back-pressure in a pipeline the cause is most probably not this algorithm.

Back to your observation, the answer : yes and no. In your network model, I can see the logic of “logging” and “committing” a final snapshot being provided by the channel implementation. However, do mind that the first barrier always needs to go “all the way” to initiate the Chandy Lamport algorithm logic.

The above flow has been proven using temporal logic in my phd thesis in case you are interested about the proof.
I hope this helps a little clarifying things. Let me know if there is any confusing point to disambiguate. I would be more than happy to help if I can.

Paris

Thomas Weise

Re: Checkpointing under backpressure

Great discussion! I'm excited that this is already under consideration! Are
there any JIRAs or other traces of discussion to follow?

Paris, if I understand correctly, then the proposal is to not block any
input channel at all, but only log data from the backpressured channel (and
make it part of the snapshot) until the barrier arrives? This is
intriguing. But probably there is also a benefit of to not continue reading
I1 since that could speed up retrieval from I2. Also, if the user code is
the cause of backpressure, this would avoid pumping more data into the
process function.

Thanks,
Thomas

On Tue, Aug 13, 2019 at 8:02 AM zhijiang <[hidden email]>
wrote:

> Hi Paris,
>
> Thanks for the detailed sharing. And I think it is very similar with the
> way of overtaking we proposed before.
>
> There are some tiny difference:
> The way of overtaking might need to snapshot all the input/output queues.
> Chandy Lamport seems only need to snaphost (n-1) input channels after the
> first barrier arrives, which might reduce the state sizea bit. But normally
> there should be less buffers for the first input channel with barrier.
> The output barrier still follows with regular data stream in Chandy
> Lamport, the same way as current flink. For overtaking way, we need to pay
> extra efforts to make barrier transport firstly before outque queue on
> upstream side, and change the way of barrier alignment based on receiving
> instead of current reading on downstream side.
> In the backpressure caused by data skew, the first barrier in almost empty
> input channel should arrive much eariler than the last heavy load input
> channel, so the Chandy Lamport could benefit well. But for the case of all
> balanced heavy load input channels, I mean the first arrived barrier might
> still take much time, then the overtaking way could still fit well to speed
> up checkpoint.
> Anyway, your proposed suggestion is helpful on my side, especially
> considering some implementation details .
>
> Best,
> Zhijiang
> ------------------------------------------------------------------
> From:Paris Carbone <[hidden email]>
> Send Time:2019年8月13日(星期二) 14:03
> To:dev <[hidden email]>
> Cc:zhijiang <[hidden email]>
> Subject:Re: Checkpointing under backpressure
>
> yes! It’s quite similar I think. Though mind that the devil is in the
> details, i.e., the temporal order actions are taken.
>
> To clarify, let us say you have a task T with two input channels I1 and I2.
> The Chandy Lamport execution flow is the following:
>
> 1) T receives barrier from I1 and...
> 2) ...the following three actions happen atomically
> I ) T snapshots its state T*
> II) T forwards marker to its outputs
> III) T starts logging all events of I2 (only) into a buffer M*
> - Also notice here that T does NOT block I1 as it does in aligned
> snapshots -
> 3) Eventually T receives barrier from I2 and stops recording events. Its
> asynchronously captured snapshot is now complete: {T*,M*}.
> Upon recovery all messages of M* should be replayed in FIFO order.
>
> With this approach alignment does not create a deadlock situation since
> anyway 2.II happens asynchronously and messages can be logged as well
> asynchronously during the process of the snapshot. If there is
> back-pressure in a pipeline the cause is most probably not this algorithm.
>
> Back to your observation, the answer : yes and no. In your network model,
> I can see the logic of “logging” and “committing” a final snapshot being
> provided by the channel implementation. However, do mind that the first
> barrier always needs to go “all the way” to initiate the Chandy Lamport
> algorithm logic.
>
> The above flow has been proven using temporal logic in my phd thesis in
> case you are interested about the proof.
> I hope this helps a little clarifying things. Let me know if there is any
> confusing point to disambiguate. I would be more than happy to help if I
> can.
>
> Paris
>
> > On 13 Aug 2019, at 13:28, Piotr Nowojski <[hidden email]> wrote:
> >
> > Thanks for the input. Regarding the Chandy-Lamport snapshots don’t you
> still have to wait for the “checkpoint barrier” to arrive in order to know
> when have you already received all possible messages from the upstream
> tasks/operators? So instead of processing the “in flight” messages (as the
> Flink is doing currently), you are sending them to an “observer”?
> >
> > In that case, that’s sounds similar to “checkpoint barriers overtaking
> in flight records” (aka unaligned checkpoints). Just for us, the observer
> is a snapshot state.
> >
> > Piotrek
> >
> >> On 13 Aug 2019, at 13:14, Paris Carbone <[hidden email]>
> wrote:
> >>
> >> Interesting problem! Thanks for bringing it up Thomas.
> >>
> >> Ignore/Correct me if I am wrong but I believe Chandy-Lamport snapshots
> [1] would help out solve this problem more elegantly without sacrificing
> correctness.
> >> - They do not need alignment, only (async) logging for in-flight
> records between the time the first barrier is processed until the last
> barrier arrives in a task.
> >> - They work fine for failure recovery as long as logged records are
> replayed on startup.
> >>
> >> Flink’s “alligned” savepoints would probably be still necessary for
> transactional sink commits + any sort of reconfiguration (e.g., rescaling,
> updating the logic of operators to evolve an application etc.).
> >>
> >> I don’t completely understand the “overtaking” approach but if you have
> a concrete definition I would be happy to check it out and help if I can!
> >> Mind that Chandy-Lamport essentially does this by logging things in
> pending channels in a task snapshot before the barrier arrives.
> >>
> >> -Paris
> >>
> >> [1] https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm <
> https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm>
> >>
> >>> On 13 Aug 2019, at 10:27, Piotr Nowojski <[hidden email]> wrote:
> >>>
> >>> Hi Thomas,
> >>>
> >>> As Zhijiang has responded, we are now in the process of discussing how
> to address this issue and one of the solution that we are discussing is
> exactly what you are proposing: checkpoint barriers overtaking the in
> flight data and make the in flight data part of the checkpoint.
> >>>
> >>> If everything works well, we will be able to present result of our
> discussions on the dev mailing list soon.
> >>>
> >>> Piotrek
> >>>
> >>>> On 12 Aug 2019, at 23:23, zhijiang <[hidden email]>
> wrote:
> >>>>
> >>>> Hi Thomas,
> >>>>
> >>>> Thanks for proposing this concern. The barrier alignment takes long
> time in backpressure case which could cause several problems:
> >>>> 1. Checkpoint timeout as you mentioned.
> >>>> 2. The recovery cost is high once failover, because much data needs
> to be replayed.
> >>>> 3. The delay for commit-based sink is high in exactly-once.
> >>>>
> >>>> For credit-based flow control from release-1.5, the amount of
> in-flighting buffers before barrier alignment is reduced, so we could get a
> bit
> >>>> benefits from speeding checkpoint aspect.
> >>>>
> >>>> In release-1.8, I guess we did not suspend the channels which already
> received the barrier in practice. But actually we ever did the similar thing
> >>>> to speed barrier alighment before. I am not quite sure that
> release-1.8 covers this feature. There were some relevant discussions under
> jira [1].
> >>>>
> >>>> For release-1.10, the community is now discussing the feature of
> unaligned checkpoint which is mainly for resolving above concerns. The
> basic idea
> >>>> is to make barrier overtakes the output/input buffer queue to speed
> alignment, and snapshot the input/output buffers as part of checkpoint
> state. The
> >>>> details have not confirmed yet and is still under discussion. Wish we
> could make some improvments for the release-1.10.
> >>>>
> >>>> [1] https://issues.apache.org/jira/browse/FLINK-8523
> >>>>
> >>>> Best,
> >>>> Zhijiang
> >>>> ------------------------------------------------------------------
> >>>> From:Thomas Weise <[hidden email]>
> >>>> Send Time:2019年8月12日(星期一) 21:38
> >>>> To:dev <[hidden email]>
> >>>> Subject:Checkpointing under backpressure
> >>>>
> >>>> Hi,
> >>>>
> >>>> One of the major operational difficulties we observe with Flink are
> >>>> checkpoint timeouts under backpressure. I'm looking for both
> confirmation
> >>>> of my understanding of the current behavior as well as pointers for
> future
> >>>> improvement work:
> >>>>
> >>>> Prior to introduction of credit based flow control in the network
> stack [1]
> >>>> [2], checkpoint barriers would back up with the data for all logical
> >>>> channels due to TCP backpressure. Since Flink 1.5, the buffers are
> >>>> controlled per channel, and checkpoint barriers are only held back for
> >>>> channels that have backpressure, while others can continue processing
> >>>> normally. However, checkpoint barriers still cannot "overtake data",
> >>>> therefore checkpoint alignment remains affected for the channel with
> >>>> backpressure, with the potential for slow checkpointing and timeouts.
> >>>> Albeit the delay of barriers would be capped by the maximum in-transit
> >>>> buffers per channel, resulting in an improvement compared to previous
> >>>> versions of Flink. Also, the backpressure based checkpoint alignment
> can
> >>>> help the barrier advance faster on the receiver side (by suspending
> >>>> channels that have already delivered the barrier). Is that accurate
> as of
> >>>> Flink 1.8?
> >>>>
> >>>> What appears to be missing to completely unblock checkpointing is a
> >>>> mechanism for checkpoints to overtake the data. That would help in
> >>>> situations where the processing itself is the bottleneck and
> prioritization
> >>>> in the network stack alone cannot address the barrier delay. Was
> there any
> >>>> related discussion? One possible solution would be to drain incoming
> data
> >>>> till the barrier and make it part of the checkpoint instead of
> processing
> >>>> it. This is somewhat related to asynchronous processing, but I'm
> thinking
> >>>> more of a solution that is automated in the Flink runtime for the
> >>>> backpressure scenario only.
> >>>>
> >>>> Thanks,
> >>>> Thomas
> >>>>
> >>>> [1] https://flink.apache.org/2019/06/05/flink-network-stack.html
> >>>> [2]
> >>>>
> https://docs.google.com/document/d/1chTOuOqe0sBsjldA_r-wXYeSIhU2zRGpUaTaik7QZ84/edit#heading=h.pjh6mv7m2hjn
> >>>>
> >>>
> >>
> >
>
>

Zhijiang(wangzhijiang999)

Re: Checkpointing under backpressure

Hi Thomas,

There are no Jira tickets or discussions at the moment. If there are any updates I would ping you.

I agree that there is a benefit to firstly read other channels without barrier in high priority, otherwise it seems
waste cpu resource to migrate blocked buffers to another cached queue. Actually we ever implemented
this improvement in Alibaba's private branch and it was running long time in real prodcution. It indeeds
speed up barrier alignment a bit if the blocked input channels have some buffers.

Best,
Zhijiang
------------------------------------------------------------------
From:Thomas Weise <[hidden email]>
Send Time:2019年8月14日(星期三) 06:00
To:dev <[hidden email]>; zhijiang <[hidden email]>
Cc:Paris Carbone <[hidden email]>
Subject:Re: Checkpointing under backpressure

Great discussion! I'm excited that this is already under consideration! Are
there any JIRAs or other traces of discussion to follow?

Paris, if I understand correctly, then the proposal is to not block any
input channel at all, but only log data from the backpressured channel (and
make it part of the snapshot) until the barrier arrives? This is
intriguing. But probably there is also a benefit of to not continue reading
I1 since that could speed up retrieval from I2. Also, if the user code is
the cause of backpressure, this would avoid pumping more data into the
process function.

Thanks,
Thomas

On Tue, Aug 13, 2019 at 8:02 AM zhijiang <[hidden email]>
wrote:

Piotr Nowojski-3

Re: Checkpointing under backpressure

In reply to this post by Thomas Weise

Hi,

Thomas:
There are no Jira tickets yet (or maybe there is something very old somewhere). First we want to discuss it, next present FLIP and at last create tickets :)

> if I understand correctly, then the proposal is to not block any
> input channel at all, but only log data from the backpressured channel (and
> make it part of the snapshot) until the barrier arrives

I would guess that it would be better to block the reads, unless we can already process the records from the blocked channel…

Paris:

Thanks for the explanation Paris. I’m starting to understand this more and I like the idea of snapshotting the state of an operator before receiving all of the checkpoint barriers - this would allow more things to happen at the same time instead of sequentially. As Zhijiang has pointed out there are some things not considered in your proposal: overtaking output buffers, but maybe those things could be incorporated together.

Another thing is that from the wiki description I understood that the initial checkpointing is not initialised by any checkpoint barrier, but by an independent call/message from the Observer. I haven’t played with this idea a lot, but I had some discussion with Nico and it seems that it might work:

1. JobManager sends and RPC “start checkpoint” to all tasks
2. Task (with two input channels l1 and l2) upon receiving RPC from 1., takes a snapshot of it's state and:
a) broadcast checkpoint barrier down the stream to all channels (let’s ignore for a moment potential for this barrier to overtake the buffer output data)
b) for any input channel for which it hasn’t yet received checkpoint barrier, the data are being added to the checkpoint
c) once a channel (for example l1) receives a checkpoint barrier, the Task blocks reads from that channel (?)
d) after all remaining channels (l2) receive checkpoint barriers, the Task first has to process the buffered data after that it can unblock the reads from the channels

Checkpoint barriers do not cascade/flow through different tasks here. Checkpoint barrier emitted from Task1, reaches only the immediate downstream Tasks. Thanks to this setup, total checkpointing time is not sum of checkpointing times of all Tasks one by one, but more or less max of the slowest Tasks. Right?

Couple of intriguing thoughts are:
3. checkpoint barriers overtaking the output buffers
4. can we keep processing some data (in order to not waste CPU cycles) after we have taking the snapshot of the Task. I think we could.

Piotrek

> On 14 Aug 2019, at 06:00, Thomas Weise <[hidden email]> wrote:
>
> Great discussion! I'm excited that this is already under consideration! Are
> there any JIRAs or other traces of discussion to follow?
>
> Paris, if I understand correctly, then the proposal is to not block any
> input channel at all, but only log data from the backpressured channel (and
> make it part of the snapshot) until the barrier arrives? This is
> intriguing. But probably there is also a benefit of to not continue reading
> I1 since that could speed up retrieval from I2. Also, if the user code is
> the cause of backpressure, this would avoid pumping more data into the
> process function.
>
> Thanks,
> Thomas
>
>
> On Tue, Aug 13, 2019 at 8:02 AM zhijiang <[hidden email]>
> wrote:
>
>> Hi Paris,
>>
>> Thanks for the detailed sharing. And I think it is very similar with the
>> way of overtaking we proposed before.
>>
>> There are some tiny difference:
>> The way of overtaking might need to snapshot all the input/output queues.
>> Chandy Lamport seems only need to snaphost (n-1) input channels after the
>> first barrier arrives, which might reduce the state sizea bit. But normally
>> there should be less buffers for the first input channel with barrier.
>> The output barrier still follows with regular data stream in Chandy
>> Lamport, the same way as current flink. For overtaking way, we need to pay
>> extra efforts to make barrier transport firstly before outque queue on
>> upstream side, and change the way of barrier alignment based on receiving
>> instead of current reading on downstream side.
>> In the backpressure caused by data skew, the first barrier in almost empty
>> input channel should arrive much eariler than the last heavy load input
>> channel, so the Chandy Lamport could benefit well. But for the case of all
>> balanced heavy load input channels, I mean the first arrived barrier might
>> still take much time, then the overtaking way could still fit well to speed
>> up checkpoint.
>> Anyway, your proposed suggestion is helpful on my side, especially
>> considering some implementation details .
>>
>> Best,
>> Zhijiang
>> ------------------------------------------------------------------
>> From:Paris Carbone <[hidden email]>
>> Send Time:2019年8月13日(星期二) 14:03
>> To:dev <[hidden email]>
>> Cc:zhijiang <[hidden email]>
>> Subject:Re: Checkpointing under backpressure
>>
>> yes! It’s quite similar I think. Though mind that the devil is in the
>> details, i.e., the temporal order actions are taken.
>>
>> To clarify, let us say you have a task T with two input channels I1 and I2.
>> The Chandy Lamport execution flow is the following:
>>
>> 1) T receives barrier from I1 and...
>> 2) ...the following three actions happen atomically
>> I ) T snapshots its state T*
>> II) T forwards marker to its outputs
>> III) T starts logging all events of I2 (only) into a buffer M*
>> - Also notice here that T does NOT block I1 as it does in aligned
>> snapshots -
>> 3) Eventually T receives barrier from I2 and stops recording events. Its
>> asynchronously captured snapshot is now complete: {T*,M*}.
>> Upon recovery all messages of M* should be replayed in FIFO order.
>>
>> With this approach alignment does not create a deadlock situation since
>> anyway 2.II happens asynchronously and messages can be logged as well
>> asynchronously during the process of the snapshot. If there is
>> back-pressure in a pipeline the cause is most probably not this algorithm.
>>
>> Back to your observation, the answer : yes and no. In your network model,
>> I can see the logic of “logging” and “committing” a final snapshot being
>> provided by the channel implementation. However, do mind that the first
>> barrier always needs to go “all the way” to initiate the Chandy Lamport
>> algorithm logic.
>>
>> The above flow has been proven using temporal logic in my phd thesis in
>> case you are interested about the proof.
>> I hope this helps a little clarifying things. Let me know if there is any
>> confusing point to disambiguate. I would be more than happy to help if I
>> can.
>>
>> Paris
>>
>>> On 13 Aug 2019, at 13:28, Piotr Nowojski <[hidden email]> wrote:
>>>
>>> Thanks for the input. Regarding the Chandy-Lamport snapshots don’t you
>> still have to wait for the “checkpoint barrier” to arrive in order to know
>> when have you already received all possible messages from the upstream
>> tasks/operators? So instead of processing the “in flight” messages (as the
>> Flink is doing currently), you are sending them to an “observer”?
>>>
>>> In that case, that’s sounds similar to “checkpoint barriers overtaking
>> in flight records” (aka unaligned checkpoints). Just for us, the observer
>> is a snapshot state.
>>>
>>> Piotrek
>>>
>>>> On 13 Aug 2019, at 13:14, Paris Carbone <[hidden email]>
>> wrote:
>>>>
>>>> Interesting problem! Thanks for bringing it up Thomas.
>>>>
>>>> Ignore/Correct me if I am wrong but I believe Chandy-Lamport snapshots
>> [1] would help out solve this problem more elegantly without sacrificing
>> correctness.
>>>> - They do not need alignment, only (async) logging for in-flight
>> records between the time the first barrier is processed until the last
>> barrier arrives in a task.
>>>> - They work fine for failure recovery as long as logged records are
>> replayed on startup.
>>>>
>>>> Flink’s “alligned” savepoints would probably be still necessary for
>> transactional sink commits + any sort of reconfiguration (e.g., rescaling,
>> updating the logic of operators to evolve an application etc.).
>>>>
>>>> I don’t completely understand the “overtaking” approach but if you have
>> a concrete definition I would be happy to check it out and help if I can!
>>>> Mind that Chandy-Lamport essentially does this by logging things in
>> pending channels in a task snapshot before the barrier arrives.
>>>>
>>>> -Paris
>>>>
>>>> [1] https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm <
>> https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm>
>>>>
>>>>> On 13 Aug 2019, at 10:27, Piotr Nowojski <[hidden email]> wrote:
>>>>>
>>>>> Hi Thomas,
>>>>>
>>>>> As Zhijiang has responded, we are now in the process of discussing how
>> to address this issue and one of the solution that we are discussing is
>> exactly what you are proposing: checkpoint barriers overtaking the in
>> flight data and make the in flight data part of the checkpoint.
>>>>>
>>>>> If everything works well, we will be able to present result of our
>> discussions on the dev mailing list soon.
>>>>>
>>>>> Piotrek
>>>>>
>>>>>> On 12 Aug 2019, at 23:23, zhijiang <[hidden email]>
>> wrote:
>>>>>>
>>>>>> Hi Thomas,
>>>>>>
>>>>>> Thanks for proposing this concern. The barrier alignment takes long
>> time in backpressure case which could cause several problems:
>>>>>> 1. Checkpoint timeout as you mentioned.
>>>>>> 2. The recovery cost is high once failover, because much data needs
>> to be replayed.
>>>>>> 3. The delay for commit-based sink is high in exactly-once.
>>>>>>
>>>>>> For credit-based flow control from release-1.5, the amount of
>> in-flighting buffers before barrier alignment is reduced, so we could get a
>> bit
>>>>>> benefits from speeding checkpoint aspect.
>>>>>>
>>>>>> In release-1.8, I guess we did not suspend the channels which already
>> received the barrier in practice. But actually we ever did the similar thing
>>>>>> to speed barrier alighment before. I am not quite sure that
>> release-1.8 covers this feature. There were some relevant discussions under
>> jira [1].
>>>>>>
>>>>>> For release-1.10, the community is now discussing the feature of
>> unaligned checkpoint which is mainly for resolving above concerns. The
>> basic idea
>>>>>> is to make barrier overtakes the output/input buffer queue to speed
>> alignment, and snapshot the input/output buffers as part of checkpoint
>> state. The
>>>>>> details have not confirmed yet and is still under discussion. Wish we
>> could make some improvments for the release-1.10.
>>>>>>
>>>>>> [1] https://issues.apache.org/jira/browse/FLINK-8523
>>>>>>
>>>>>> Best,
>>>>>> Zhijiang
>>>>>> ------------------------------------------------------------------
>>>>>> From:Thomas Weise <[hidden email]>
>>>>>> Send Time:2019年8月12日(星期一) 21:38
>>>>>> To:dev <[hidden email]>
>>>>>> Subject:Checkpointing under backpressure
>>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> One of the major operational difficulties we observe with Flink are
>>>>>> checkpoint timeouts under backpressure. I'm looking for both
>> confirmation
>>>>>> of my understanding of the current behavior as well as pointers for
>> future
>>>>>> improvement work:
>>>>>>
>>>>>> Prior to introduction of credit based flow control in the network
>> stack [1]
>>>>>> [2], checkpoint barriers would back up with the data for all logical
>>>>>> channels due to TCP backpressure. Since Flink 1.5, the buffers are
>>>>>> controlled per channel, and checkpoint barriers are only held back for
>>>>>> channels that have backpressure, while others can continue processing
>>>>>> normally. However, checkpoint barriers still cannot "overtake data",
>>>>>> therefore checkpoint alignment remains affected for the channel with
>>>>>> backpressure, with the potential for slow checkpointing and timeouts.
>>>>>> Albeit the delay of barriers would be capped by the maximum in-transit
>>>>>> buffers per channel, resulting in an improvement compared to previous
>>>>>> versions of Flink. Also, the backpressure based checkpoint alignment
>> can
>>>>>> help the barrier advance faster on the receiver side (by suspending
>>>>>> channels that have already delivered the barrier). Is that accurate
>> as of
>>>>>> Flink 1.8?
>>>>>>
>>>>>> What appears to be missing to completely unblock checkpointing is a
>>>>>> mechanism for checkpoints to overtake the data. That would help in
>>>>>> situations where the processing itself is the bottleneck and
>> prioritization
>>>>>> in the network stack alone cannot address the barrier delay. Was
>> there any
>>>>>> related discussion? One possible solution would be to drain incoming
>> data
>>>>>> till the barrier and make it part of the checkpoint instead of
>> processing
>>>>>> it. This is somewhat related to asynchronous processing, but I'm
>> thinking
>>>>>> more of a solution that is automated in the Flink runtime for the
>>>>>> backpressure scenario only.
>>>>>>
>>>>>> Thanks,
>>>>>> Thomas
>>>>>>
>>>>>> [1] https://flink.apache.org/2019/06/05/flink-network-stack.html
>>>>>> [2]
>>>>>>
>> https://docs.google.com/document/d/1chTOuOqe0sBsjldA_r-wXYeSIhU2zRGpUaTaik7QZ84/edit#heading=h.pjh6mv7m2hjn
>>>>>>
>>>>>
>>>>
>>>
>>
>>

Zhu Zhu

Re: Checkpointing under backpressure

Thanks Piotr and Zhijiang for sharing the thoughts on unaligned
checkpointing and the barrier overtaking.

I have a question about 2.d) in Piotr's last mail that states "the Task
first has to process the buffered data after that it can unblock the reads
from the channels".
Does this mean that we do not want checkpoint 2 to happen if the data
before checkpoint 1 barriers are not fully processed, as the barriers of
checkpoint 2 can be blocked?
If so, the checkpointing may get blocked once the job is overloaded.
If not, a overloaded job may lead to a larger and larger checkpoint.

And I agree a slot with point 4 in Piotr's last mail and think it's
necessary for a CP1 snapshotted task to process CP1 buffered data before
all CP1 barriers are are received in this task.
Otherwise it might be a processing performance regression compared to
current exactly-once checkpointing.

Thanks,
Zhu Zhu

Piotr Nowojski <[hidden email]> 于2019年8月14日周三下午3:08写道：

> Hi,
>
> Thomas:
> There are no Jira tickets yet (or maybe there is something very old
> somewhere). First we want to discuss it, next present FLIP and at last
> create tickets :)
>
> > if I understand correctly, then the proposal is to not block any
> > input channel at all, but only log data from the backpressured channel
> (and
> > make it part of the snapshot) until the barrier arrives
>
> I would guess that it would be better to block the reads, unless we can
> already process the records from the blocked channel…
>
> Paris:
>
> Thanks for the explanation Paris. I’m starting to understand this more and
> I like the idea of snapshotting the state of an operator before receiving
> all of the checkpoint barriers - this would allow more things to happen at
> the same time instead of sequentially. As Zhijiang has pointed out there
> are some things not considered in your proposal: overtaking output buffers,
> but maybe those things could be incorporated together.
>
> Another thing is that from the wiki description I understood that the
> initial checkpointing is not initialised by any checkpoint barrier, but by
> an independent call/message from the Observer. I haven’t played with this
> idea a lot, but I had some discussion with Nico and it seems that it might
> work:
>
> 1. JobManager sends and RPC “start checkpoint” to all tasks
> 2. Task (with two input channels l1 and l2) upon receiving RPC from 1.,
> takes a snapshot of it's state and:
> a) broadcast checkpoint barrier down the stream to all channels (let’s
> ignore for a moment potential for this barrier to overtake the buffer
> output data)
> b) for any input channel for which it hasn’t yet received checkpoint
> barrier, the data are being added to the checkpoint
> c) once a channel (for example l1) receives a checkpoint barrier, the
> Task blocks reads from that channel (?)
> d) after all remaining channels (l2) receive checkpoint barriers, the
> Task first has to process the buffered data after that it can unblock the
> reads from the channels
>
> Checkpoint barriers do not cascade/flow through different tasks here.
> Checkpoint barrier emitted from Task1, reaches only the immediate
> downstream Tasks. Thanks to this setup, total checkpointing time is not sum
> of checkpointing times of all Tasks one by one, but more or less max of the
> slowest Tasks. Right?
>
> Couple of intriguing thoughts are:
> 3. checkpoint barriers overtaking the output buffers
> 4. can we keep processing some data (in order to not waste CPU cycles)
> after we have taking the snapshot of the Task. I think we could.
>
> Piotrek
>
> > On 14 Aug 2019, at 06:00, Thomas Weise <[hidden email]> wrote:
> >
> > Great discussion! I'm excited that this is already under consideration!
> Are
> > there any JIRAs or other traces of discussion to follow?
> >
> > Paris, if I understand correctly, then the proposal is to not block any
> > input channel at all, but only log data from the backpressured channel
> (and
> > make it part of the snapshot) until the barrier arrives? This is
> > intriguing. But probably there is also a benefit of to not continue
> reading
> > I1 since that could speed up retrieval from I2. Also, if the user code is
> > the cause of backpressure, this would avoid pumping more data into the
> > process function.
> >
> > Thanks,
> > Thomas
> >
> >
> > On Tue, Aug 13, 2019 at 8:02 AM zhijiang <[hidden email]
> .invalid>
> > wrote:
> >
> >> Hi Paris,
> >>
> >> Thanks for the detailed sharing. And I think it is very similar with the
> >> way of overtaking we proposed before.
> >>
> >> There are some tiny difference:
> >> The way of overtaking might need to snapshot all the input/output
> queues.
> >> Chandy Lamport seems only need to snaphost (n-1) input channels after
> the
> >> first barrier arrives, which might reduce the state sizea bit. But
> normally
> >> there should be less buffers for the first input channel with barrier.
> >> The output barrier still follows with regular data stream in Chandy
> >> Lamport, the same way as current flink. For overtaking way, we need to
> pay
> >> extra efforts to make barrier transport firstly before outque queue on
> >> upstream side, and change the way of barrier alignment based on
> receiving
> >> instead of current reading on downstream side.
> >> In the backpressure caused by data skew, the first barrier in almost
> empty
> >> input channel should arrive much eariler than the last heavy load input
> >> channel, so the Chandy Lamport could benefit well. But for the case of
> all
> >> balanced heavy load input channels, I mean the first arrived barrier
> might
> >> still take much time, then the overtaking way could still fit well to
> speed
> >> up checkpoint.
> >> Anyway, your proposed suggestion is helpful on my side, especially
> >> considering some implementation details .
> >>
> >> Best,
> >> Zhijiang
> >> ------------------------------------------------------------------
> >> From:Paris Carbone <[hidden email]>
> >> Send Time:2019年8月13日(星期二) 14:03
> >> To:dev <[hidden email]>
> >> Cc:zhijiang <[hidden email]>
> >> Subject:Re: Checkpointing under backpressure
> >>
> >> yes! It’s quite similar I think. Though mind that the devil is in the
> >> details, i.e., the temporal order actions are taken.
> >>
> >> To clarify, let us say you have a task T with two input channels I1 and
> I2.
> >> The Chandy Lamport execution flow is the following:
> >>
> >> 1) T receives barrier from I1 and...
> >> 2) ...the following three actions happen atomically
> >> I ) T snapshots its state T*
> >> II) T forwards marker to its outputs
> >> III) T starts logging all events of I2 (only) into a buffer M*
> >> - Also notice here that T does NOT block I1 as it does in aligned
> >> snapshots -
> >> 3) Eventually T receives barrier from I2 and stops recording events. Its
> >> asynchronously captured snapshot is now complete: {T*,M*}.
> >> Upon recovery all messages of M* should be replayed in FIFO order.
> >>
> >> With this approach alignment does not create a deadlock situation since
> >> anyway 2.II happens asynchronously and messages can be logged as well
> >> asynchronously during the process of the snapshot. If there is
> >> back-pressure in a pipeline the cause is most probably not this
> algorithm.
> >>
> >> Back to your observation, the answer : yes and no. In your network
> model,
> >> I can see the logic of “logging” and “committing” a final snapshot being
> >> provided by the channel implementation. However, do mind that the first
> >> barrier always needs to go “all the way” to initiate the Chandy Lamport
> >> algorithm logic.
> >>
> >> The above flow has been proven using temporal logic in my phd thesis in
> >> case you are interested about the proof.
> >> I hope this helps a little clarifying things. Let me know if there is
> any
> >> confusing point to disambiguate. I would be more than happy to help if I
> >> can.
> >>
> >> Paris
> >>
> >>> On 13 Aug 2019, at 13:28, Piotr Nowojski <[hidden email]> wrote:
> >>>
> >>> Thanks for the input. Regarding the Chandy-Lamport snapshots don’t you
> >> still have to wait for the “checkpoint barrier” to arrive in order to
> know
> >> when have you already received all possible messages from the upstream
> >> tasks/operators? So instead of processing the “in flight” messages (as
> the
> >> Flink is doing currently), you are sending them to an “observer”?
> >>>
> >>> In that case, that’s sounds similar to “checkpoint barriers overtaking
> >> in flight records” (aka unaligned checkpoints). Just for us, the
> observer
> >> is a snapshot state.
> >>>
> >>> Piotrek
> >>>
> >>>> On 13 Aug 2019, at 13:14, Paris Carbone <[hidden email]>
> >> wrote:
> >>>>
> >>>> Interesting problem! Thanks for bringing it up Thomas.
> >>>>
> >>>> Ignore/Correct me if I am wrong but I believe Chandy-Lamport snapshots
> >> [1] would help out solve this problem more elegantly without sacrificing
> >> correctness.
> >>>> - They do not need alignment, only (async) logging for in-flight
> >> records between the time the first barrier is processed until the last
> >> barrier arrives in a task.
> >>>> - They work fine for failure recovery as long as logged records are
> >> replayed on startup.
> >>>>
> >>>> Flink’s “alligned” savepoints would probably be still necessary for
> >> transactional sink commits + any sort of reconfiguration (e.g.,
> rescaling,
> >> updating the logic of operators to evolve an application etc.).
> >>>>
> >>>> I don’t completely understand the “overtaking” approach but if you
> have
> >> a concrete definition I would be happy to check it out and help if I
> can!
> >>>> Mind that Chandy-Lamport essentially does this by logging things in
> >> pending channels in a task snapshot before the barrier arrives.
> >>>>
> >>>> -Paris
> >>>>
> >>>> [1] https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm <
> >> https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm>
> >>>>
> >>>>> On 13 Aug 2019, at 10:27, Piotr Nowojski <[hidden email]>
> wrote:
> >>>>>
> >>>>> Hi Thomas,
> >>>>>
> >>>>> As Zhijiang has responded, we are now in the process of discussing
> how
> >> to address this issue and one of the solution that we are discussing is
> >> exactly what you are proposing: checkpoint barriers overtaking the in
> >> flight data and make the in flight data part of the checkpoint.
> >>>>>
> >>>>> If everything works well, we will be able to present result of our
> >> discussions on the dev mailing list soon.
> >>>>>
> >>>>> Piotrek
> >>>>>
> >>>>>> On 12 Aug 2019, at 23:23, zhijiang <[hidden email]
> .INVALID>
> >> wrote:
> >>>>>>
> >>>>>> Hi Thomas,
> >>>>>>
> >>>>>> Thanks for proposing this concern. The barrier alignment takes long
> >> time in backpressure case which could cause several problems:
> >>>>>> 1. Checkpoint timeout as you mentioned.
> >>>>>> 2. The recovery cost is high once failover, because much data needs
> >> to be replayed.
> >>>>>> 3. The delay for commit-based sink is high in exactly-once.
> >>>>>>
> >>>>>> For credit-based flow control from release-1.5, the amount of
> >> in-flighting buffers before barrier alignment is reduced, so we could
> get a
> >> bit
> >>>>>> benefits from speeding checkpoint aspect.
> >>>>>>
> >>>>>> In release-1.8, I guess we did not suspend the channels which
> already
> >> received the barrier in practice. But actually we ever did the similar
> thing
> >>>>>> to speed barrier alighment before. I am not quite sure that
> >> release-1.8 covers this feature. There were some relevant discussions
> under
> >> jira [1].
> >>>>>>
> >>>>>> For release-1.10, the community is now discussing the feature of
> >> unaligned checkpoint which is mainly for resolving above concerns. The
> >> basic idea
> >>>>>> is to make barrier overtakes the output/input buffer queue to speed
> >> alignment, and snapshot the input/output buffers as part of checkpoint
> >> state. The
> >>>>>> details have not confirmed yet and is still under discussion. Wish
> we
> >> could make some improvments for the release-1.10.
> >>>>>>
> >>>>>> [1] https://issues.apache.org/jira/browse/FLINK-8523
> >>>>>>
> >>>>>> Best,
> >>>>>> Zhijiang
> >>>>>> ------------------------------------------------------------------
> >>>>>> From:Thomas Weise <[hidden email]>
> >>>>>> Send Time:2019年8月12日(星期一) 21:38
> >>>>>> To:dev <[hidden email]>
> >>>>>> Subject:Checkpointing under backpressure
> >>>>>>
> >>>>>> Hi,
> >>>>>>
> >>>>>> One of the major operational difficulties we observe with Flink are
> >>>>>> checkpoint timeouts under backpressure. I'm looking for both
> >> confirmation
> >>>>>> of my understanding of the current behavior as well as pointers for
> >> future
> >>>>>> improvement work:
> >>>>>>
> >>>>>> Prior to introduction of credit based flow control in the network
> >> stack [1]
> >>>>>> [2], checkpoint barriers would back up with the data for all logical
> >>>>>> channels due to TCP backpressure. Since Flink 1.5, the buffers are
> >>>>>> controlled per channel, and checkpoint barriers are only held back
> for
> >>>>>> channels that have backpressure, while others can continue
> processing
> >>>>>> normally. However, checkpoint barriers still cannot "overtake data",
> >>>>>> therefore checkpoint alignment remains affected for the channel with
> >>>>>> backpressure, with the potential for slow checkpointing and
> timeouts.
> >>>>>> Albeit the delay of barriers would be capped by the maximum
> in-transit
> >>>>>> buffers per channel, resulting in an improvement compared to
> previous
> >>>>>> versions of Flink. Also, the backpressure based checkpoint alignment
> >> can
> >>>>>> help the barrier advance faster on the receiver side (by suspending
> >>>>>> channels that have already delivered the barrier). Is that accurate
> >> as of
> >>>>>> Flink 1.8?
> >>>>>>
> >>>>>> What appears to be missing to completely unblock checkpointing is a
> >>>>>> mechanism for checkpoints to overtake the data. That would help in
> >>>>>> situations where the processing itself is the bottleneck and
> >> prioritization
> >>>>>> in the network stack alone cannot address the barrier delay. Was
> >> there any
> >>>>>> related discussion? One possible solution would be to drain incoming
> >> data
> >>>>>> till the barrier and make it part of the checkpoint instead of
> >> processing
> >>>>>> it. This is somewhat related to asynchronous processing, but I'm
> >> thinking
> >>>>>> more of a solution that is automated in the Flink runtime for the
> >>>>>> backpressure scenario only.
> >>>>>>
> >>>>>> Thanks,
> >>>>>> Thomas
> >>>>>>
> >>>>>> [1] https://flink.apache.org/2019/06/05/flink-network-stack.html
> >>>>>> [2]
> >>>>>>
> >>
> https://docs.google.com/document/d/1chTOuOqe0sBsjldA_r-wXYeSIhU2zRGpUaTaik7QZ84/edit#heading=h.pjh6mv7m2hjn
> >>>>>>
> >>>>>
> >>>>
> >>>
> >>
> >>
>
>

Paris Carbone-3

Re: Checkpointing under backpressure

In reply to this post by Piotr Nowojski-3

Now I see a little more clearly what you have in mind. Thanks for the explanation!
There are a few intermixed concepts here, some how to do with correctness some with performance.
Before delving deeper I will just enumerate a few things to make myself a little more helpful if I can.

I. Initiation
-------------

1. RPC to sources only is a less intrusive way to initiate snapshots since you utilize better pipeline parallelism (only a small subset of tasks is running progressively the protocol at a time, if snapshotting is async the overall overhead might not even be observable).

2. If we really want an RPC to all initiation take notice of the following implications:

a. (correctness) RPC calls are not guaranteed to arrive in every task before a marker from a preceding task.

b. (correctness) Either the RPC call OR the first arriving marker should initiate the algorithm. Whichever comes first. If you only do it per RPC call then you capture a "late" state that includes side effects of already logged events.

c. (performance) Lots of IO will be invoked at the same time on the backend store from all tasks. This might lead to high congestion in async snapshots.

II. Capturing State First
-------------------------

1. (correctness) Capturing state at the last marker sounds incorrect to me (state contains side effects of already logged events based on the proposed scheme). This results into duplicate processing. No?

III. Channel Blocking / "Alignment"
-----------------------------------

1. (performance?) What is the added benefit? We dont want a "complete" transactional snapshot, async snapshots are purely for failure-recovery. Thus, I dont see why this needs to be imposed at the expense of performance/throughput. With the proposed scheme the whole dataflow anyway enters snapshotting/logging mode so tasks more or less snapshot concurrently.

IV Marker Bypassing
-------------------

1. (correctness) This leads to equivalent in-flight snapshots so with some quick thinking correct. I will try to model this later and get back to you in case I find something wrong.

2. (performance) It also sounds like a meaningful optimisation! I like thinking of this as a push-based snapshot. i.e., the producing task somehow triggers forward a consumer/channel to capture its state. By example consider T1 -> |marker t1| -> T2.

V. Usage of "Async" Snapshots
---------------------

1. Do you see this as a full replacement of "full" aligned snapshots/savepoints? In my view async shanpshots will be needed from time to time but not as frequently. Yet, it seems like a valid approach solely for failure-recovery on the same configuration. Here's why:

a. With original snapshotting there is a strong duality between
a stream input (offsets) and committed side effects (internal states and external commits to transactional sinks). While in the async version, there are uncommitted operations (inflight records). Thus, you cannot use these snapshots for e.g., submitting sql queries with snapshot isolation. Also, the original snapshotting gives a lot of potential for flink to make proper transactional commits externally.

b. Reconfiguration is very tricky, you probably know that better. Inflight channel state is no longer valid in a new configuration (i.e., new dataflow graph, new operators, updated operator logic, different channels, different parallelism)

2. Async snapshots can also be potentially useful for monitoring the general health of a dataflow since they can be analyzed by the task manager about the general performance of a job graph and spot bottlenecks for example.

> On 14 Aug 2019, at 09:08, Piotr Nowojski <[hidden email]> wrote:
>
> Hi,
>
> Thomas:
> There are no Jira tickets yet (or maybe there is something very old somewhere). First we want to discuss it, next present FLIP and at last create tickets :)
>
>> if I understand correctly, then the proposal is to not block any
>> input channel at all, but only log data from the backpressured channel (and
>> make it part of the snapshot) until the barrier arrives
>
> I would guess that it would be better to block the reads, unless we can already process the records from the blocked channel…
>
> Paris:
>
> Thanks for the explanation Paris. I’m starting to understand this more and I like the idea of snapshotting the state of an operator before receiving all of the checkpoint barriers - this would allow more things to happen at the same time instead of sequentially. As Zhijiang has pointed out there are some things not considered in your proposal: overtaking output buffers, but maybe those things could be incorporated together.
>
> Another thing is that from the wiki description I understood that the initial checkpointing is not initialised by any checkpoint barrier, but by an independent call/message from the Observer. I haven’t played with this idea a lot, but I had some discussion with Nico and it seems that it might work:
>
> 1. JobManager sends and RPC “start checkpoint” to all tasks
> 2. Task (with two input channels l1 and l2) upon receiving RPC from 1., takes a snapshot of it's state and:
> a) broadcast checkpoint barrier down the stream to all channels (let’s ignore for a moment potential for this barrier to overtake the buffer output data)
> b) for any input channel for which it hasn’t yet received checkpoint barrier, the data are being added to the checkpoint
> c) once a channel (for example l1) receives a checkpoint barrier, the Task blocks reads from that channel (?)
> d) after all remaining channels (l2) receive checkpoint barriers, the Task first has to process the buffered data after that it can unblock the reads from the channels
>
> Checkpoint barriers do not cascade/flow through different tasks here. Checkpoint barrier emitted from Task1, reaches only the immediate downstream Tasks. Thanks to this setup, total checkpointing time is not sum of checkpointing times of all Tasks one by one, but more or less max of the slowest Tasks. Right?
>
> Couple of intriguing thoughts are:
> 3. checkpoint barriers overtaking the output buffers
> 4. can we keep processing some data (in order to not waste CPU cycles) after we have taking the snapshot of the Task. I think we could.
>
> Piotrek
>
>> On 14 Aug 2019, at 06:00, Thomas Weise <[hidden email]> wrote:
>>
>> Great discussion! I'm excited that this is already under consideration! Are
>> there any JIRAs or other traces of discussion to follow?
>>
>> Paris, if I understand correctly, then the proposal is to not block any
>> input channel at all, but only log data from the backpressured channel (and
>> make it part of the snapshot) until the barrier arrives? This is
>> intriguing. But probably there is also a benefit of to not continue reading
>> I1 since that could speed up retrieval from I2. Also, if the user code is
>> the cause of backpressure, this would avoid pumping more data into the
>> process function.
>>
>> Thanks,
>> Thomas
>>
>>
>> On Tue, Aug 13, 2019 at 8:02 AM zhijiang <[hidden email]>
>> wrote:
>>
>>> Hi Paris,
>>>
>>> Thanks for the detailed sharing. And I think it is very similar with the
>>> way of overtaking we proposed before.
>>>
>>> There are some tiny difference:
>>> The way of overtaking might need to snapshot all the input/output queues.
>>> Chandy Lamport seems only need to snaphost (n-1) input channels after the
>>> first barrier arrives, which might reduce the state sizea bit. But normally
>>> there should be less buffers for the first input channel with barrier.
>>> The output barrier still follows with regular data stream in Chandy
>>> Lamport, the same way as current flink. For overtaking way, we need to pay
>>> extra efforts to make barrier transport firstly before outque queue on
>>> upstream side, and change the way of barrier alignment based on receiving
>>> instead of current reading on downstream side.
>>> In the backpressure caused by data skew, the first barrier in almost empty
>>> input channel should arrive much eariler than the last heavy load input
>>> channel, so the Chandy Lamport could benefit well. But for the case of all
>>> balanced heavy load input channels, I mean the first arrived barrier might
>>> still take much time, then the overtaking way could still fit well to speed
>>> up checkpoint.
>>> Anyway, your proposed suggestion is helpful on my side, especially
>>> considering some implementation details .
>>>
>>> Best,
>>> Zhijiang
>>> ------------------------------------------------------------------
>>> From:Paris Carbone <[hidden email]>
>>> Send Time:2019年8月13日(星期二) 14:03
>>> To:dev <[hidden email]>
>>> Cc:zhijiang <[hidden email]>
>>> Subject:Re: Checkpointing under backpressure
>>>
>>> yes! It’s quite similar I think. Though mind that the devil is in the
>>> details, i.e., the temporal order actions are taken.
>>>
>>> To clarify, let us say you have a task T with two input channels I1 and I2.
>>> The Chandy Lamport execution flow is the following:
>>>
>>> 1) T receives barrier from I1 and...
>>> 2) ...the following three actions happen atomically
>>> I ) T snapshots its state T*
>>> II) T forwards marker to its outputs
>>> III) T starts logging all events of I2 (only) into a buffer M*
>>> - Also notice here that T does NOT block I1 as it does in aligned
>>> snapshots -
>>> 3) Eventually T receives barrier from I2 and stops recording events. Its
>>> asynchronously captured snapshot is now complete: {T*,M*}.
>>> Upon recovery all messages of M* should be replayed in FIFO order.
>>>
>>> With this approach alignment does not create a deadlock situation since
>>> anyway 2.II happens asynchronously and messages can be logged as well
>>> asynchronously during the process of the snapshot. If there is
>>> back-pressure in a pipeline the cause is most probably not this algorithm.
>>>
>>> Back to your observation, the answer : yes and no. In your network model,
>>> I can see the logic of “logging” and “committing” a final snapshot being
>>> provided by the channel implementation. However, do mind that the first
>>> barrier always needs to go “all the way” to initiate the Chandy Lamport
>>> algorithm logic.
>>>
>>> The above flow has been proven using temporal logic in my phd thesis in
>>> case you are interested about the proof.
>>> I hope this helps a little clarifying things. Let me know if there is any
>>> confusing point to disambiguate. I would be more than happy to help if I
>>> can.
>>>
>>> Paris
>>>
>>>> On 13 Aug 2019, at 13:28, Piotr Nowojski <[hidden email]> wrote:
>>>>
>>>> Thanks for the input. Regarding the Chandy-Lamport snapshots don’t you
>>> still have to wait for the “checkpoint barrier” to arrive in order to know
>>> when have you already received all possible messages from the upstream
>>> tasks/operators? So instead of processing the “in flight” messages (as the
>>> Flink is doing currently), you are sending them to an “observer”?
>>>>
>>>> In that case, that’s sounds similar to “checkpoint barriers overtaking
>>> in flight records” (aka unaligned checkpoints). Just for us, the observer
>>> is a snapshot state.
>>>>
>>>> Piotrek
>>>>
>>>>> On 13 Aug 2019, at 13:14, Paris Carbone <[hidden email]>
>>> wrote:
>>>>>
>>>>> Interesting problem! Thanks for bringing it up Thomas.
>>>>>
>>>>> Ignore/Correct me if I am wrong but I believe Chandy-Lamport snapshots
>>> [1] would help out solve this problem more elegantly without sacrificing
>>> correctness.
>>>>> - They do not need alignment, only (async) logging for in-flight
>>> records between the time the first barrier is processed until the last
>>> barrier arrives in a task.
>>>>> - They work fine for failure recovery as long as logged records are
>>> replayed on startup.
>>>>>
>>>>> Flink’s “alligned” savepoints would probably be still necessary for
>>> transactional sink commits + any sort of reconfiguration (e.g., rescaling,
>>> updating the logic of operators to evolve an application etc.).
>>>>>
>>>>> I don’t completely understand the “overtaking” approach but if you have
>>> a concrete definition I would be happy to check it out and help if I can!
>>>>> Mind that Chandy-Lamport essentially does this by logging things in
>>> pending channels in a task snapshot before the barrier arrives.
>>>>>
>>>>> -Paris
>>>>>
>>>>> [1] https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm <
>>> https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm>
>>>>>
>>>>>> On 13 Aug 2019, at 10:27, Piotr Nowojski <[hidden email]> wrote:
>>>>>>
>>>>>> Hi Thomas,
>>>>>>
>>>>>> As Zhijiang has responded, we are now in the process of discussing how
>>> to address this issue and one of the solution that we are discussing is
>>> exactly what you are proposing: checkpoint barriers overtaking the in
>>> flight data and make the in flight data part of the checkpoint.
>>>>>>
>>>>>> If everything works well, we will be able to present result of our
>>> discussions on the dev mailing list soon.
>>>>>>
>>>>>> Piotrek
>>>>>>
>>>>>>> On 12 Aug 2019, at 23:23, zhijiang <[hidden email]>
>>> wrote:
>>>>>>>
>>>>>>> Hi Thomas,
>>>>>>>
>>>>>>> Thanks for proposing this concern. The barrier alignment takes long
>>> time in backpressure case which could cause several problems:
>>>>>>> 1. Checkpoint timeout as you mentioned.
>>>>>>> 2. The recovery cost is high once failover, because much data needs
>>> to be replayed.
>>>>>>> 3. The delay for commit-based sink is high in exactly-once.
>>>>>>>
>>>>>>> For credit-based flow control from release-1.5, the amount of
>>> in-flighting buffers before barrier alignment is reduced, so we could get a
>>> bit
>>>>>>> benefits from speeding checkpoint aspect.
>>>>>>>
>>>>>>> In release-1.8, I guess we did not suspend the channels which already
>>> received the barrier in practice. But actually we ever did the similar thing
>>>>>>> to speed barrier alighment before. I am not quite sure that
>>> release-1.8 covers this feature. There were some relevant discussions under
>>> jira [1].
>>>>>>>
>>>>>>> For release-1.10, the community is now discussing the feature of
>>> unaligned checkpoint which is mainly for resolving above concerns. The
>>> basic idea
>>>>>>> is to make barrier overtakes the output/input buffer queue to speed
>>> alignment, and snapshot the input/output buffers as part of checkpoint
>>> state. The
>>>>>>> details have not confirmed yet and is still under discussion. Wish we
>>> could make some improvments for the release-1.10.
>>>>>>>
>>>>>>> [1] https://issues.apache.org/jira/browse/FLINK-8523
>>>>>>>
>>>>>>> Best,
>>>>>>> Zhijiang
>>>>>>> ------------------------------------------------------------------
>>>>>>> From:Thomas Weise <[hidden email]>
>>>>>>> Send Time:2019年8月12日(星期一) 21:38
>>>>>>> To:dev <[hidden email]>
>>>>>>> Subject:Checkpointing under backpressure
>>>>>>>
>>>>>>> Hi,
>>>>>>>
>>>>>>> One of the major operational difficulties we observe with Flink are
>>>>>>> checkpoint timeouts under backpressure. I'm looking for both
>>> confirmation
>>>>>>> of my understanding of the current behavior as well as pointers for
>>> future
>>>>>>> improvement work:
>>>>>>>
>>>>>>> Prior to introduction of credit based flow control in the network
>>> stack [1]
>>>>>>> [2], checkpoint barriers would back up with the data for all logical
>>>>>>> channels due to TCP backpressure. Since Flink 1.5, the buffers are
>>>>>>> controlled per channel, and checkpoint barriers are only held back for
>>>>>>> channels that have backpressure, while others can continue processing
>>>>>>> normally. However, checkpoint barriers still cannot "overtake data",
>>>>>>> therefore checkpoint alignment remains affected for the channel with
>>>>>>> backpressure, with the potential for slow checkpointing and timeouts.
>>>>>>> Albeit the delay of barriers would be capped by the maximum in-transit
>>>>>>> buffers per channel, resulting in an improvement compared to previous
>>>>>>> versions of Flink. Also, the backpressure based checkpoint alignment
>>> can
>>>>>>> help the barrier advance faster on the receiver side (by suspending
>>>>>>> channels that have already delivered the barrier). Is that accurate
>>> as of
>>>>>>> Flink 1.8?
>>>>>>>
>>>>>>> What appears to be missing to completely unblock checkpointing is a
>>>>>>> mechanism for checkpoints to overtake the data. That would help in
>>>>>>> situations where the processing itself is the bottleneck and
>>> prioritization
>>>>>>> in the network stack alone cannot address the barrier delay. Was
>>> there any
>>>>>>> related discussion? One possible solution would be to drain incoming
>>> data
>>>>>>> till the barrier and make it part of the checkpoint instead of
>>> processing
>>>>>>> it. This is somewhat related to asynchronous processing, but I'm
>>> thinking
>>>>>>> more of a solution that is automated in the Flink runtime for the
>>>>>>> backpressure scenario only.
>>>>>>>
>>>>>>> Thanks,
>>>>>>> Thomas
>>>>>>>
>>>>>>> [1] https://flink.apache.org/2019/06/05/flink-network-stack.html
>>>>>>> [2]
>>>>>>>
>>> https://docs.google.com/document/d/1chTOuOqe0sBsjldA_r-wXYeSIhU2zRGpUaTaik7QZ84/edit#heading=h.pjh6mv7m2hjn
>>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>>>
>

Piotr Nowojski-3

Re: Checkpointing under backpressure

Hi again,

Zhu Zhu let me think about this more. Maybe as Paris is writing, we do not need to block any channels at all, at least assuming credit base flow control. Regarding what should happen with the following checkpoint is another question. Also, should we support concurrent checkpoints and subsuming checkpoints as we do now? Maybe not…

Paris

Re
I. 2. a) and b) - yes, this would have to be taken into an account
I. 2. c) and IV. 2. - without those, end to end checkpoint time will probably be longer than it could be. It might affect external systems. For example Kafka, which automatically time outs lingering transactions, and for us, the transaction time is equal to the time between two checkpoints.

II 1. - I’m confused. To make things straight. Flink is currently snapshotting once it receives all of the checkpoint barriers from all of the input channels and only then it broadcasts the checkpoint barrier down the stream. And this is correct from exactly-once perspective.

As far as I understand, your proposal based on Chandy Lamport algorithm, is snapshotting the state of the operator on the first checkpoint barrier, which also looks correct to me.

III. 1. As I responded to Zhu Zhu, let me think a bit more about this.

V. Yes, we still need aligned checkpoints, as they are easier for state migration and upgrades.

Piotrek

> On 14 Aug 2019, at 11:22, Paris Carbone <[hidden email]> wrote:
>
> Now I see a little more clearly what you have in mind. Thanks for the explanation!
> There are a few intermixed concepts here, some how to do with correctness some with performance.
> Before delving deeper I will just enumerate a few things to make myself a little more helpful if I can.
>
> I. Initiation
> -------------
>
> 1. RPC to sources only is a less intrusive way to initiate snapshots since you utilize better pipeline parallelism (only a small subset of tasks is running progressively the protocol at a time, if snapshotting is async the overall overhead might not even be observable).
>
> 2. If we really want an RPC to all initiation take notice of the following implications:
>
> a. (correctness) RPC calls are not guaranteed to arrive in every task before a marker from a preceding task.
>
> b. (correctness) Either the RPC call OR the first arriving marker should initiate the algorithm. Whichever comes first. If you only do it per RPC call then you capture a "late" state that includes side effects of already logged events.
>
> c. (performance) Lots of IO will be invoked at the same time on the backend store from all tasks. This might lead to high congestion in async snapshots.
>
> II. Capturing State First
> -------------------------
>
> 1. (correctness) Capturing state at the last marker sounds incorrect to me (state contains side effects of already logged events based on the proposed scheme). This results into duplicate processing. No?
>
> III. Channel Blocking / "Alignment"
> -----------------------------------
>
> 1. (performance?) What is the added benefit? We dont want a "complete" transactional snapshot, async snapshots are purely for failure-recovery. Thus, I dont see why this needs to be imposed at the expense of performance/throughput. With the proposed scheme the whole dataflow anyway enters snapshotting/logging mode so tasks more or less snapshot concurrently.
>
> IV Marker Bypassing
> -------------------
>
> 1. (correctness) This leads to equivalent in-flight snapshots so with some quick thinking correct. I will try to model this later and get back to you in case I find something wrong.
>
> 2. (performance) It also sounds like a meaningful optimisation! I like thinking of this as a push-based snapshot. i.e., the producing task somehow triggers forward a consumer/channel to capture its state. By example consider T1 -> |marker t1| -> T2.
>
> V. Usage of "Async" Snapshots
> ---------------------
>
> 1. Do you see this as a full replacement of "full" aligned snapshots/savepoints? In my view async shanpshots will be needed from time to time but not as frequently. Yet, it seems like a valid approach solely for failure-recovery on the same configuration. Here's why:
>
> a. With original snapshotting there is a strong duality between
> a stream input (offsets) and committed side effects (internal states and external commits to transactional sinks). While in the async version, there are uncommitted operations (inflight records). Thus, you cannot use these snapshots for e.g., submitting sql queries with snapshot isolation. Also, the original snapshotting gives a lot of potential for flink to make proper transactional commits externally.
>
> b. Reconfiguration is very tricky, you probably know that better. Inflight channel state is no longer valid in a new configuration (i.e., new dataflow graph, new operators, updated operator logic, different channels, different parallelism)
>
> 2. Async snapshots can also be potentially useful for monitoring the general health of a dataflow since they can be analyzed by the task manager about the general performance of a job graph and spot bottlenecks for example.
>
>> On 14 Aug 2019, at 09:08, Piotr Nowojski <[hidden email]> wrote:
>>
>> Hi,
>>
>> Thomas:
>> There are no Jira tickets yet (or maybe there is something very old somewhere). First we want to discuss it, next present FLIP and at last create tickets :)
>>
>>> if I understand correctly, then the proposal is to not block any
>>> input channel at all, but only log data from the backpressured channel (and
>>> make it part of the snapshot) until the barrier arrives
>>
>> I would guess that it would be better to block the reads, unless we can already process the records from the blocked channel…
>>
>> Paris:
>>
>> Thanks for the explanation Paris. I’m starting to understand this more and I like the idea of snapshotting the state of an operator before receiving all of the checkpoint barriers - this would allow more things to happen at the same time instead of sequentially. As Zhijiang has pointed out there are some things not considered in your proposal: overtaking output buffers, but maybe those things could be incorporated together.
>>
>> Another thing is that from the wiki description I understood that the initial checkpointing is not initialised by any checkpoint barrier, but by an independent call/message from the Observer. I haven’t played with this idea a lot, but I had some discussion with Nico and it seems that it might work:
>>
>> 1. JobManager sends and RPC “start checkpoint” to all tasks
>> 2. Task (with two input channels l1 and l2) upon receiving RPC from 1., takes a snapshot of it's state and:
>> a) broadcast checkpoint barrier down the stream to all channels (let’s ignore for a moment potential for this barrier to overtake the buffer output data)
>> b) for any input channel for which it hasn’t yet received checkpoint barrier, the data are being added to the checkpoint
>> c) once a channel (for example l1) receives a checkpoint barrier, the Task blocks reads from that channel (?)
>> d) after all remaining channels (l2) receive checkpoint barriers, the Task first has to process the buffered data after that it can unblock the reads from the channels
>>
>> Checkpoint barriers do not cascade/flow through different tasks here. Checkpoint barrier emitted from Task1, reaches only the immediate downstream Tasks. Thanks to this setup, total checkpointing time is not sum of checkpointing times of all Tasks one by one, but more or less max of the slowest Tasks. Right?
>>
>> Couple of intriguing thoughts are:
>> 3. checkpoint barriers overtaking the output buffers
>> 4. can we keep processing some data (in order to not waste CPU cycles) after we have taking the snapshot of the Task. I think we could.
>>
>> Piotrek
>>
>>> On 14 Aug 2019, at 06:00, Thomas Weise <[hidden email]> wrote:
>>>
>>> Great discussion! I'm excited that this is already under consideration! Are
>>> there any JIRAs or other traces of discussion to follow?
>>>
>>> Paris, if I understand correctly, then the proposal is to not block any
>>> input channel at all, but only log data from the backpressured channel (and
>>> make it part of the snapshot) until the barrier arrives? This is
>>> intriguing. But probably there is also a benefit of to not continue reading
>>> I1 since that could speed up retrieval from I2. Also, if the user code is
>>> the cause of backpressure, this would avoid pumping more data into the
>>> process function.
>>>
>>> Thanks,
>>> Thomas
>>>
>>>
>>> On Tue, Aug 13, 2019 at 8:02 AM zhijiang <[hidden email]>
>>> wrote:
>>>
>>>> Hi Paris,
>>>>
>>>> Thanks for the detailed sharing. And I think it is very similar with the
>>>> way of overtaking we proposed before.
>>>>
>>>> There are some tiny difference:
>>>> The way of overtaking might need to snapshot all the input/output queues.
>>>> Chandy Lamport seems only need to snaphost (n-1) input channels after the
>>>> first barrier arrives, which might reduce the state sizea bit. But normally
>>>> there should be less buffers for the first input channel with barrier.
>>>> The output barrier still follows with regular data stream in Chandy
>>>> Lamport, the same way as current flink. For overtaking way, we need to pay
>>>> extra efforts to make barrier transport firstly before outque queue on
>>>> upstream side, and change the way of barrier alignment based on receiving
>>>> instead of current reading on downstream side.
>>>> In the backpressure caused by data skew, the first barrier in almost empty
>>>> input channel should arrive much eariler than the last heavy load input
>>>> channel, so the Chandy Lamport could benefit well. But for the case of all
>>>> balanced heavy load input channels, I mean the first arrived barrier might
>>>> still take much time, then the overtaking way could still fit well to speed
>>>> up checkpoint.
>>>> Anyway, your proposed suggestion is helpful on my side, especially
>>>> considering some implementation details .
>>>>
>>>> Best,
>>>> Zhijiang
>>>> ------------------------------------------------------------------
>>>> From:Paris Carbone <[hidden email]>
>>>> Send Time:2019年8月13日(星期二) 14:03
>>>> To:dev <[hidden email]>
>>>> Cc:zhijiang <[hidden email]>
>>>> Subject:Re: Checkpointing under backpressure
>>>>
>>>> yes! It’s quite similar I think. Though mind that the devil is in the
>>>> details, i.e., the temporal order actions are taken.
>>>>
>>>> To clarify, let us say you have a task T with two input channels I1 and I2.
>>>> The Chandy Lamport execution flow is the following:
>>>>
>>>> 1) T receives barrier from I1 and...
>>>> 2) ...the following three actions happen atomically
>>>> I ) T snapshots its state T*
>>>> II) T forwards marker to its outputs
>>>> III) T starts logging all events of I2 (only) into a buffer M*
>>>> - Also notice here that T does NOT block I1 as it does in aligned
>>>> snapshots -
>>>> 3) Eventually T receives barrier from I2 and stops recording events. Its
>>>> asynchronously captured snapshot is now complete: {T*,M*}.
>>>> Upon recovery all messages of M* should be replayed in FIFO order.
>>>>
>>>> With this approach alignment does not create a deadlock situation since
>>>> anyway 2.II happens asynchronously and messages can be logged as well
>>>> asynchronously during the process of the snapshot. If there is
>>>> back-pressure in a pipeline the cause is most probably not this algorithm.
>>>>
>>>> Back to your observation, the answer : yes and no. In your network model,
>>>> I can see the logic of “logging” and “committing” a final snapshot being
>>>> provided by the channel implementation. However, do mind that the first
>>>> barrier always needs to go “all the way” to initiate the Chandy Lamport
>>>> algorithm logic.
>>>>
>>>> The above flow has been proven using temporal logic in my phd thesis in
>>>> case you are interested about the proof.
>>>> I hope this helps a little clarifying things. Let me know if there is any
>>>> confusing point to disambiguate. I would be more than happy to help if I
>>>> can.
>>>>
>>>> Paris
>>>>
>>>>> On 13 Aug 2019, at 13:28, Piotr Nowojski <[hidden email]> wrote:
>>>>>
>>>>> Thanks for the input. Regarding the Chandy-Lamport snapshots don’t you
>>>> still have to wait for the “checkpoint barrier” to arrive in order to know
>>>> when have you already received all possible messages from the upstream
>>>> tasks/operators? So instead of processing the “in flight” messages (as the
>>>> Flink is doing currently), you are sending them to an “observer”?
>>>>>
>>>>> In that case, that’s sounds similar to “checkpoint barriers overtaking
>>>> in flight records” (aka unaligned checkpoints). Just for us, the observer
>>>> is a snapshot state.
>>>>>
>>>>> Piotrek
>>>>>
>>>>>> On 13 Aug 2019, at 13:14, Paris Carbone <[hidden email]>
>>>> wrote:
>>>>>>
>>>>>> Interesting problem! Thanks for bringing it up Thomas.
>>>>>>
>>>>>> Ignore/Correct me if I am wrong but I believe Chandy-Lamport snapshots
>>>> [1] would help out solve this problem more elegantly without sacrificing
>>>> correctness.
>>>>>> - They do not need alignment, only (async) logging for in-flight
>>>> records between the time the first barrier is processed until the last
>>>> barrier arrives in a task.
>>>>>> - They work fine for failure recovery as long as logged records are
>>>> replayed on startup.
>>>>>>
>>>>>> Flink’s “alligned” savepoints would probably be still necessary for
>>>> transactional sink commits + any sort of reconfiguration (e.g., rescaling,
>>>> updating the logic of operators to evolve an application etc.).
>>>>>>
>>>>>> I don’t completely understand the “overtaking” approach but if you have
>>>> a concrete definition I would be happy to check it out and help if I can!
>>>>>> Mind that Chandy-Lamport essentially does this by logging things in
>>>> pending channels in a task snapshot before the barrier arrives.
>>>>>>
>>>>>> -Paris
>>>>>>
>>>>>> [1] https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm <
>>>> https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm>
>>>>>>
>>>>>>> On 13 Aug 2019, at 10:27, Piotr Nowojski <[hidden email]> wrote:
>>>>>>>
>>>>>>> Hi Thomas,
>>>>>>>
>>>>>>> As Zhijiang has responded, we are now in the process of discussing how
>>>> to address this issue and one of the solution that we are discussing is
>>>> exactly what you are proposing: checkpoint barriers overtaking the in
>>>> flight data and make the in flight data part of the checkpoint.
>>>>>>>
>>>>>>> If everything works well, we will be able to present result of our
>>>> discussions on the dev mailing list soon.
>>>>>>>
>>>>>>> Piotrek
>>>>>>>
>>>>>>>> On 12 Aug 2019, at 23:23, zhijiang <[hidden email]>
>>>> wrote:
>>>>>>>>
>>>>>>>> Hi Thomas,
>>>>>>>>
>>>>>>>> Thanks for proposing this concern. The barrier alignment takes long
>>>> time in backpressure case which could cause several problems:
>>>>>>>> 1. Checkpoint timeout as you mentioned.
>>>>>>>> 2. The recovery cost is high once failover, because much data needs
>>>> to be replayed.
>>>>>>>> 3. The delay for commit-based sink is high in exactly-once.
>>>>>>>>
>>>>>>>> For credit-based flow control from release-1.5, the amount of
>>>> in-flighting buffers before barrier alignment is reduced, so we could get a
>>>> bit
>>>>>>>> benefits from speeding checkpoint aspect.
>>>>>>>>
>>>>>>>> In release-1.8, I guess we did not suspend the channels which already
>>>> received the barrier in practice. But actually we ever did the similar thing
>>>>>>>> to speed barrier alighment before. I am not quite sure that
>>>> release-1.8 covers this feature. There were some relevant discussions under
>>>> jira [1].
>>>>>>>>
>>>>>>>> For release-1.10, the community is now discussing the feature of
>>>> unaligned checkpoint which is mainly for resolving above concerns. The
>>>> basic idea
>>>>>>>> is to make barrier overtakes the output/input buffer queue to speed
>>>> alignment, and snapshot the input/output buffers as part of checkpoint
>>>> state. The
>>>>>>>> details have not confirmed yet and is still under discussion. Wish we
>>>> could make some improvments for the release-1.10.
>>>>>>>>
>>>>>>>> [1] https://issues.apache.org/jira/browse/FLINK-8523
>>>>>>>>
>>>>>>>> Best,
>>>>>>>> Zhijiang
>>>>>>>> ------------------------------------------------------------------
>>>>>>>> From:Thomas Weise <[hidden email]>
>>>>>>>> Send Time:2019年8月12日(星期一) 21:38
>>>>>>>> To:dev <[hidden email]>
>>>>>>>> Subject:Checkpointing under backpressure
>>>>>>>>
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> One of the major operational difficulties we observe with Flink are
>>>>>>>> checkpoint timeouts under backpressure. I'm looking for both
>>>> confirmation
>>>>>>>> of my understanding of the current behavior as well as pointers for
>>>> future
>>>>>>>> improvement work:
>>>>>>>>
>>>>>>>> Prior to introduction of credit based flow control in the network
>>>> stack [1]
>>>>>>>> [2], checkpoint barriers would back up with the data for all logical
>>>>>>>> channels due to TCP backpressure. Since Flink 1.5, the buffers are
>>>>>>>> controlled per channel, and checkpoint barriers are only held back for
>>>>>>>> channels that have backpressure, while others can continue processing
>>>>>>>> normally. However, checkpoint barriers still cannot "overtake data",
>>>>>>>> therefore checkpoint alignment remains affected for the channel with
>>>>>>>> backpressure, with the potential for slow checkpointing and timeouts.
>>>>>>>> Albeit the delay of barriers would be capped by the maximum in-transit
>>>>>>>> buffers per channel, resulting in an improvement compared to previous
>>>>>>>> versions of Flink. Also, the backpressure based checkpoint alignment
>>>> can
>>>>>>>> help the barrier advance faster on the receiver side (by suspending
>>>>>>>> channels that have already delivered the barrier). Is that accurate
>>>> as of
>>>>>>>> Flink 1.8?
>>>>>>>>
>>>>>>>> What appears to be missing to completely unblock checkpointing is a
>>>>>>>> mechanism for checkpoints to overtake the data. That would help in
>>>>>>>> situations where the processing itself is the bottleneck and
>>>> prioritization
>>>>>>>> in the network stack alone cannot address the barrier delay. Was
>>>> there any
>>>>>>>> related discussion? One possible solution would be to drain incoming
>>>> data
>>>>>>>> till the barrier and make it part of the checkpoint instead of
>>>> processing
>>>>>>>> it. This is somewhat related to asynchronous processing, but I'm
>>>> thinking
>>>>>>>> more of a solution that is automated in the Flink runtime for the
>>>>>>>> backpressure scenario only.
>>>>>>>>
>>>>>>>> Thanks,
>>>>>>>> Thomas
>>>>>>>>
>>>>>>>> [1] https://flink.apache.org/2019/06/05/flink-network-stack.html
>>>>>>>> [2]
>>>>>>>>
>>>> https://docs.google.com/document/d/1chTOuOqe0sBsjldA_r-wXYeSIhU2zRGpUaTaik7QZ84/edit#heading=h.pjh6mv7m2hjn
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>
>>>>
>>
>

Yun Gao

Re: Checkpointing under backpressure

Hi,
Very thanks for sharing the thoughts on the unaligned checkpoint !
Another question regarding I 2.C (Performance) by Paris is that do we always snapshot and broadcast the marks once the task receives the first mark from JM o? If so, then we will always need to snapshot all the records before the next barriers in all the input channels . However, I thinks users may be able to tolerate a fixed interval for the checkpointing, and we may postpone the snapshot and broadcast till a configurable fixed time passed. With such a postpone, jobs without back pressure could still avoid most IO operations and storage overhead, and jobs with back pressure could also be able to finish the checkpoint in a fixed interval with less IO overhead.

Best,
Yun

------------------------------------------------------------------
From:Piotr Nowojski <[hidden email]>
Send Time:2019 Aug. 14 (Wed.) 18:38
To:Paris Carbone <[hidden email]>
Cc:dev <[hidden email]>; zhijiang <[hidden email]>; Nico Kruber <[hidden email]>
Subject:Re: Checkpointing under backpressure

Hi again,

Zhu Zhu let me think about this more. Maybe as Paris is writing, we do not need to block any channels at all, at least assuming credit base flow control. Regarding what should happen with the following checkpoint is another question. Also, should we support concurrent checkpoints and subsuming checkpoints as we do now? Maybe not…

Paris

Re
I. 2. a) and b) - yes, this would have to be taken into an account
I. 2. c) and IV. 2. - without those, end to end checkpoint time will probably be longer than it could be. It might affect external systems. For example Kafka, which automatically time outs lingering transactions, and for us, the transaction time is equal to the time between two checkpoints.

II 1. - I’m confused. To make things straight. Flink is currently snapshotting once it receives all of the checkpoint barriers from all of the input channels and only then it broadcasts the checkpoint barrier down the stream. And this is correct from exactly-once perspective.

As far as I understand, your proposal based on Chandy Lamport algorithm, is snapshotting the state of the operator on the first checkpoint barrier, which also looks correct to me.

III. 1. As I responded to Zhu Zhu, let me think a bit more about this.

V. Yes, we still need aligned checkpoints, as they are easier for state migration and upgrades.

Piotrek

Stephan Ewen

Re: Checkpointing under backpressure

In reply to this post by Piotr Nowojski-3

Hi all!

Yes, the first proposal of "unaligend checkpoints" (probably two years back
now) drew a major inspiration from Chandy Lamport, as did actually the
original checkpointing algorithm.

"Logging data between first and last barrier" versus "barrier jumping over
buffer and storing those buffers" is pretty close same.
However, there are a few nice benefits of the proposal of unaligned
checkpoints over Chandy-Lamport.

*## Benefits of Unaligned Checkpoints*

(1) It is very similar to the original algorithm (can be seen an an
optional feature purely in the network stack) and thus can share lot's of
code paths.

(2) Less data stored. If we make the "jump over buffers" part timeout based
(for example barrier overtakes buffers if not flushed within 10ms) then
checkpoints are in the common case of flowing pipelines aligned without
in-flight data. Only back pressured cases store some in-flight data, which
means we don't regress in the common case and only fix the back pressure
case.

(3) Faster checkpoints. Chandy Lamport still waits for all barriers to
arrive naturally, logging on the way. If data processing is slow, this can
still take quite a while.

==> I think both these points are strong reasons to not change the
mechanism away from "trigger sources" and start with CL-style "trigger all".

*## Possible ways to combine Chandy Lamport and Unaligned Checkpoints*

We can think about something like "take state snapshot on first barrier"
and then store buffers until the other barriers arrive. Inside the network
stack, barriers could still overtake and persist buffers.
The benefit would be less latency increase in the channels which already
have received barriers.
However, as mentioned before, not prioritizing the inputs from which
barriers are still missing can also have an adverse effect.

*## Concerning upgrades*

I think it is a fair restriction to say that upgrades need to happen on
aligned checkpoints. It is a rare enough operation.

*## Concerning re-scaling (changing parallelism)*

We need to support that on unaligned checkpoints as well. There are several
feature proposals about automatic scaling, especially down scaling in case
of missing resources. The last snapshot might be a regular checkpoint, so
all checkpoints need to support rescaling.

*## Concerning end-to-end checkpoint duration and "trigger sources" versus
"trigger all"*

I think for the end-to-end checkpoint duration, an "overtake buffers"
approach yields faster checkpoints, as mentioned above (Chandy Lamport
logging still needs to wait for barrier to flow).

I don't see the benefit of a "trigger all tasks via RPC concurrently"
approach. Bear in mind that it is still a globally coordinated approach and
you need to wait for the global checkpoint to complete before committing
any side effects.
I believe that the checkpoint time is more determined by the state
checkpoint writing, and the global coordination and metadata commit, than
by the difference in alignment time between "trigger from source and jump
over buffers" versus "trigger all tasks concurrently".

Trying to optimize a few tens of milliseconds out of the network stack
sends (and changing the overall checkpointing approach completely for that)
while staying with a globally coordinated checkpoint will send us down a
path to a dead end.

To really bring task persistence latency down to 10s of milliseconds (so we
can frequently commit in sinks), we need to take an approach without any
global coordination. Tasks need to establish a persistent recovery point
individually and at their own discretion, only then can it be frequent
enough. To get there, they would need to decouple themselves from the
predecessor and successor tasks (via something like persistent channels).
This is a different discussion, though, somewhat orthogonal to this one
here.

Best,
Stephan

On Wed, Aug 14, 2019 at 12:37 PM Piotr Nowojski <[hidden email]> wrote:

> Hi again,
>
> Zhu Zhu let me think about this more. Maybe as Paris is writing, we do not
> need to block any channels at all, at least assuming credit base flow
> control. Regarding what should happen with the following checkpoint is
> another question. Also, should we support concurrent checkpoints and
> subsuming checkpoints as we do now? Maybe not…
>
> Paris
>
> Re
> I. 2. a) and b) - yes, this would have to be taken into an account
> I. 2. c) and IV. 2. - without those, end to end checkpoint time will
> probably be longer than it could be. It might affect external systems. For
> example Kafka, which automatically time outs lingering transactions, and
> for us, the transaction time is equal to the time between two checkpoints.
>
> II 1. - I’m confused. To make things straight. Flink is currently
> snapshotting once it receives all of the checkpoint barriers from all of
> the input channels and only then it broadcasts the checkpoint barrier down
> the stream. And this is correct from exactly-once perspective.
>
> As far as I understand, your proposal based on Chandy Lamport algorithm,
> is snapshotting the state of the operator on the first checkpoint barrier,
> which also looks correct to me.
>
> III. 1. As I responded to Zhu Zhu, let me think a bit more about this.
>
> V. Yes, we still need aligned checkpoints, as they are easier for state
> migration and upgrades.
>
> Piotrek
>
> > On 14 Aug 2019, at 11:22, Paris Carbone <[hidden email]> wrote:
> >
> > Now I see a little more clearly what you have in mind. Thanks for the
> explanation!
> > There are a few intermixed concepts here, some how to do with
> correctness some with performance.
> > Before delving deeper I will just enumerate a few things to make myself
> a little more helpful if I can.
> >
> > I. Initiation
> > -------------
> >
> > 1. RPC to sources only is a less intrusive way to initiate snapshots
> since you utilize better pipeline parallelism (only a small subset of tasks
> is running progressively the protocol at a time, if snapshotting is async
> the overall overhead might not even be observable).
> >
> > 2. If we really want an RPC to all initiation take notice of the
> following implications:
> >
> > a. (correctness) RPC calls are not guaranteed to arrive in every
> task before a marker from a preceding task.
> >
> > b. (correctness) Either the RPC call OR the first arriving marker
> should initiate the algorithm. Whichever comes first. If you only do it per
> RPC call then you capture a "late" state that includes side effects of
> already logged events.
> >
> > c. (performance) Lots of IO will be invoked at the same time on
> the backend store from all tasks. This might lead to high congestion in
> async snapshots.
> >
> > II. Capturing State First
> > -------------------------
> >
> > 1. (correctness) Capturing state at the last marker sounds incorrect to
> me (state contains side effects of already logged events based on the
> proposed scheme). This results into duplicate processing. No?
> >
> > III. Channel Blocking / "Alignment"
> > -----------------------------------
> >
> > 1. (performance?) What is the added benefit? We dont want a "complete"
> transactional snapshot, async snapshots are purely for failure-recovery.
> Thus, I dont see why this needs to be imposed at the expense of
> performance/throughput. With the proposed scheme the whole dataflow anyway
> enters snapshotting/logging mode so tasks more or less snapshot
> concurrently.
> >
> > IV Marker Bypassing
> > -------------------
> >
> > 1. (correctness) This leads to equivalent in-flight snapshots so with
> some quick thinking correct. I will try to model this later and get back
> to you in case I find something wrong.
> >
> > 2. (performance) It also sounds like a meaningful optimisation! I like
> thinking of this as a push-based snapshot. i.e., the producing task somehow
> triggers forward a consumer/channel to capture its state. By example
> consider T1 -> |marker t1| -> T2.
> >
> > V. Usage of "Async" Snapshots
> > ---------------------
> >
> > 1. Do you see this as a full replacement of "full" aligned
> snapshots/savepoints? In my view async shanpshots will be needed from time
> to time but not as frequently. Yet, it seems like a valid approach solely
> for failure-recovery on the same configuration. Here's why:
> >
> > a. With original snapshotting there is a strong duality between
> > a stream input (offsets) and committed side effects (internal
> states and external commits to transactional sinks). While in the async
> version, there are uncommitted operations (inflight records). Thus, you
> cannot use these snapshots for e.g., submitting sql queries with snapshot
> isolation. Also, the original snapshotting gives a lot of potential for
> flink to make proper transactional commits externally.
> >
> > b. Reconfiguration is very tricky, you probably know that better.
> Inflight channel state is no longer valid in a new configuration (i.e., new
> dataflow graph, new operators, updated operator logic, different channels,
> different parallelism)
> >
> > 2. Async snapshots can also be potentially useful for monitoring the
> general health of a dataflow since they can be analyzed by the task manager
> about the general performance of a job graph and spot bottlenecks for
> example.
> >
> >> On 14 Aug 2019, at 09:08, Piotr Nowojski <[hidden email]> wrote:
> >>
> >> Hi,
> >>
> >> Thomas:
> >> There are no Jira tickets yet (or maybe there is something very old
> somewhere). First we want to discuss it, next present FLIP and at last
> create tickets :)
> >>
> >>> if I understand correctly, then the proposal is to not block any
> >>> input channel at all, but only log data from the backpressured channel
> (and
> >>> make it part of the snapshot) until the barrier arrives
> >>
> >> I would guess that it would be better to block the reads, unless we can
> already process the records from the blocked channel…
> >>
> >> Paris:
> >>
> >> Thanks for the explanation Paris. I’m starting to understand this more
> and I like the idea of snapshotting the state of an operator before
> receiving all of the checkpoint barriers - this would allow more things to
> happen at the same time instead of sequentially. As Zhijiang has pointed
> out there are some things not considered in your proposal: overtaking
> output buffers, but maybe those things could be incorporated together.
> >>
> >> Another thing is that from the wiki description I understood that the
> initial checkpointing is not initialised by any checkpoint barrier, but by
> an independent call/message from the Observer. I haven’t played with this
> idea a lot, but I had some discussion with Nico and it seems that it might
> work:
> >>
> >> 1. JobManager sends and RPC “start checkpoint” to all tasks
> >> 2. Task (with two input channels l1 and l2) upon receiving RPC from 1.,
> takes a snapshot of it's state and:
> >> a) broadcast checkpoint barrier down the stream to all channels (let’s
> ignore for a moment potential for this barrier to overtake the buffer
> output data)
> >> b) for any input channel for which it hasn’t yet received checkpoint
> barrier, the data are being added to the checkpoint
> >> c) once a channel (for example l1) receives a checkpoint barrier, the
> Task blocks reads from that channel (?)
> >> d) after all remaining channels (l2) receive checkpoint barriers, the
> Task first has to process the buffered data after that it can unblock the
> reads from the channels
> >>
> >> Checkpoint barriers do not cascade/flow through different tasks here.
> Checkpoint barrier emitted from Task1, reaches only the immediate
> downstream Tasks. Thanks to this setup, total checkpointing time is not sum
> of checkpointing times of all Tasks one by one, but more or less max of the
> slowest Tasks. Right?
> >>
> >> Couple of intriguing thoughts are:
> >> 3. checkpoint barriers overtaking the output buffers
> >> 4. can we keep processing some data (in order to not waste CPU cycles)
> after we have taking the snapshot of the Task. I think we could.
> >>
> >> Piotrek
> >>
> >>> On 14 Aug 2019, at 06:00, Thomas Weise <[hidden email]> wrote:
> >>>
> >>> Great discussion! I'm excited that this is already under
> consideration! Are
> >>> there any JIRAs or other traces of discussion to follow?
> >>>
> >>> Paris, if I understand correctly, then the proposal is to not block any
> >>> input channel at all, but only log data from the backpressured channel
> (and
> >>> make it part of the snapshot) until the barrier arrives? This is
> >>> intriguing. But probably there is also a benefit of to not continue
> reading
> >>> I1 since that could speed up retrieval from I2. Also, if the user code
> is
> >>> the cause of backpressure, this would avoid pumping more data into the
> >>> process function.
> >>>
> >>> Thanks,
> >>> Thomas
> >>>
> >>>
> >>> On Tue, Aug 13, 2019 at 8:02 AM zhijiang <[hidden email]
> .invalid>
> >>> wrote:
> >>>
> >>>> Hi Paris,
> >>>>
> >>>> Thanks for the detailed sharing. And I think it is very similar with
> the
> >>>> way of overtaking we proposed before.
> >>>>
> >>>> There are some tiny difference:
> >>>> The way of overtaking might need to snapshot all the input/output
> queues.
> >>>> Chandy Lamport seems only need to snaphost (n-1) input channels after
> the
> >>>> first barrier arrives, which might reduce the state sizea bit. But
> normally
> >>>> there should be less buffers for the first input channel with barrier.
> >>>> The output barrier still follows with regular data stream in Chandy
> >>>> Lamport, the same way as current flink. For overtaking way, we need
> to pay
> >>>> extra efforts to make barrier transport firstly before outque queue on
> >>>> upstream side, and change the way of barrier alignment based on
> receiving
> >>>> instead of current reading on downstream side.
> >>>> In the backpressure caused by data skew, the first barrier in almost
> empty
> >>>> input channel should arrive much eariler than the last heavy load
> input
> >>>> channel, so the Chandy Lamport could benefit well. But for the case
> of all
> >>>> balanced heavy load input channels, I mean the first arrived barrier
> might
> >>>> still take much time, then the overtaking way could still fit well to
> speed
> >>>> up checkpoint.
> >>>> Anyway, your proposed suggestion is helpful on my side, especially
> >>>> considering some implementation details .
> >>>>
> >>>> Best,
> >>>> Zhijiang
> >>>> ------------------------------------------------------------------
> >>>> From:Paris Carbone <[hidden email]>
> >>>> Send Time:2019年8月13日(星期二) 14:03
> >>>> To:dev <[hidden email]>
> >>>> Cc:zhijiang <[hidden email]>
> >>>> Subject:Re: Checkpointing under backpressure
> >>>>
> >>>> yes! It’s quite similar I think. Though mind that the devil is in the
> >>>> details, i.e., the temporal order actions are taken.
> >>>>
> >>>> To clarify, let us say you have a task T with two input channels I1
> and I2.
> >>>> The Chandy Lamport execution flow is the following:
> >>>>
> >>>> 1) T receives barrier from I1 and...
> >>>> 2) ...the following three actions happen atomically
> >>>> I ) T snapshots its state T*
> >>>> II) T forwards marker to its outputs
> >>>> III) T starts logging all events of I2 (only) into a buffer M*
> >>>> - Also notice here that T does NOT block I1 as it does in aligned
> >>>> snapshots -
> >>>> 3) Eventually T receives barrier from I2 and stops recording events.
> Its
> >>>> asynchronously captured snapshot is now complete: {T*,M*}.
> >>>> Upon recovery all messages of M* should be replayed in FIFO order.
> >>>>
> >>>> With this approach alignment does not create a deadlock situation
> since
> >>>> anyway 2.II happens asynchronously and messages can be logged as well
> >>>> asynchronously during the process of the snapshot. If there is
> >>>> back-pressure in a pipeline the cause is most probably not this
> algorithm.
> >>>>
> >>>> Back to your observation, the answer : yes and no. In your network
> model,
> >>>> I can see the logic of “logging” and “committing” a final snapshot
> being
> >>>> provided by the channel implementation. However, do mind that the
> first
> >>>> barrier always needs to go “all the way” to initiate the Chandy
> Lamport
> >>>> algorithm logic.
> >>>>
> >>>> The above flow has been proven using temporal logic in my phd thesis
> in
> >>>> case you are interested about the proof.
> >>>> I hope this helps a little clarifying things. Let me know if there is
> any
> >>>> confusing point to disambiguate. I would be more than happy to help
> if I
> >>>> can.
> >>>>
> >>>> Paris
> >>>>
> >>>>> On 13 Aug 2019, at 13:28, Piotr Nowojski <[hidden email]>
> wrote:
> >>>>>
> >>>>> Thanks for the input. Regarding the Chandy-Lamport snapshots don’t
> you
> >>>> still have to wait for the “checkpoint barrier” to arrive in order to
> know
> >>>> when have you already received all possible messages from the upstream
> >>>> tasks/operators? So instead of processing the “in flight” messages
> (as the
> >>>> Flink is doing currently), you are sending them to an “observer”?
> >>>>>
> >>>>> In that case, that’s sounds similar to “checkpoint barriers
> overtaking
> >>>> in flight records” (aka unaligned checkpoints). Just for us, the
> observer
> >>>> is a snapshot state.
> >>>>>
> >>>>> Piotrek
> >>>>>
> >>>>>> On 13 Aug 2019, at 13:14, Paris Carbone <[hidden email]>
> >>>> wrote:
> >>>>>>
> >>>>>> Interesting problem! Thanks for bringing it up Thomas.
> >>>>>>
> >>>>>> Ignore/Correct me if I am wrong but I believe Chandy-Lamport
> snapshots
> >>>> [1] would help out solve this problem more elegantly without
> sacrificing
> >>>> correctness.
> >>>>>> - They do not need alignment, only (async) logging for in-flight
> >>>> records between the time the first barrier is processed until the last
> >>>> barrier arrives in a task.
> >>>>>> - They work fine for failure recovery as long as logged records are
> >>>> replayed on startup.
> >>>>>>
> >>>>>> Flink’s “alligned” savepoints would probably be still necessary for
> >>>> transactional sink commits + any sort of reconfiguration (e.g.,
> rescaling,
> >>>> updating the logic of operators to evolve an application etc.).
> >>>>>>
> >>>>>> I don’t completely understand the “overtaking” approach but if you
> have
> >>>> a concrete definition I would be happy to check it out and help if I
> can!
> >>>>>> Mind that Chandy-Lamport essentially does this by logging things in
> >>>> pending channels in a task snapshot before the barrier arrives.
> >>>>>>
> >>>>>> -Paris
> >>>>>>
> >>>>>> [1] https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm
> <
> >>>> https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm>
> >>>>>>
> >>>>>>> On 13 Aug 2019, at 10:27, Piotr Nowojski <[hidden email]>
> wrote:
> >>>>>>>
> >>>>>>> Hi Thomas,
> >>>>>>>
> >>>>>>> As Zhijiang has responded, we are now in the process of discussing
> how
> >>>> to address this issue and one of the solution that we are discussing
> is
> >>>> exactly what you are proposing: checkpoint barriers overtaking the in
> >>>> flight data and make the in flight data part of the checkpoint.
> >>>>>>>
> >>>>>>> If everything works well, we will be able to present result of our
> >>>> discussions on the dev mailing list soon.
> >>>>>>>
> >>>>>>> Piotrek
> >>>>>>>
> >>>>>>>> On 12 Aug 2019, at 23:23, zhijiang <[hidden email]
> .INVALID>
> >>>> wrote:
> >>>>>>>>
> >>>>>>>> Hi Thomas,
> >>>>>>>>
> >>>>>>>> Thanks for proposing this concern. The barrier alignment takes
> long
> >>>> time in backpressure case which could cause several problems:
> >>>>>>>> 1. Checkpoint timeout as you mentioned.
> >>>>>>>> 2. The recovery cost is high once failover, because much data
> needs
> >>>> to be replayed.
> >>>>>>>> 3. The delay for commit-based sink is high in exactly-once.
> >>>>>>>>
> >>>>>>>> For credit-based flow control from release-1.5, the amount of
> >>>> in-flighting buffers before barrier alignment is reduced, so we could
> get a
> >>>> bit
> >>>>>>>> benefits from speeding checkpoint aspect.
> >>>>>>>>
> >>>>>>>> In release-1.8, I guess we did not suspend the channels which
> already
> >>>> received the barrier in practice. But actually we ever did the
> similar thing
> >>>>>>>> to speed barrier alighment before. I am not quite sure that
> >>>> release-1.8 covers this feature. There were some relevant discussions
> under
> >>>> jira [1].
> >>>>>>>>
> >>>>>>>> For release-1.10, the community is now discussing the feature of
> >>>> unaligned checkpoint which is mainly for resolving above concerns. The
> >>>> basic idea
> >>>>>>>> is to make barrier overtakes the output/input buffer queue to
> speed
> >>>> alignment, and snapshot the input/output buffers as part of checkpoint
> >>>> state. The
> >>>>>>>> details have not confirmed yet and is still under discussion.
> Wish we
> >>>> could make some improvments for the release-1.10.
> >>>>>>>>
> >>>>>>>> [1] https://issues.apache.org/jira/browse/FLINK-8523
> >>>>>>>>
> >>>>>>>> Best,
> >>>>>>>> Zhijiang
> >>>>>>>> ------------------------------------------------------------------
> >>>>>>>> From:Thomas Weise <[hidden email]>
> >>>>>>>> Send Time:2019年8月12日(星期一) 21:38
> >>>>>>>> To:dev <[hidden email]>
> >>>>>>>> Subject:Checkpointing under backpressure
> >>>>>>>>
> >>>>>>>> Hi,
> >>>>>>>>
> >>>>>>>> One of the major operational difficulties we observe with Flink
> are
> >>>>>>>> checkpoint timeouts under backpressure. I'm looking for both
> >>>> confirmation
> >>>>>>>> of my understanding of the current behavior as well as pointers
> for
> >>>> future
> >>>>>>>> improvement work:
> >>>>>>>>
> >>>>>>>> Prior to introduction of credit based flow control in the network
> >>>> stack [1]
> >>>>>>>> [2], checkpoint barriers would back up with the data for all
> logical
> >>>>>>>> channels due to TCP backpressure. Since Flink 1.5, the buffers are
> >>>>>>>> controlled per channel, and checkpoint barriers are only held
> back for
> >>>>>>>> channels that have backpressure, while others can continue
> processing
> >>>>>>>> normally. However, checkpoint barriers still cannot "overtake
> data",
> >>>>>>>> therefore checkpoint alignment remains affected for the channel
> with
> >>>>>>>> backpressure, with the potential for slow checkpointing and
> timeouts.
> >>>>>>>> Albeit the delay of barriers would be capped by the maximum
> in-transit
> >>>>>>>> buffers per channel, resulting in an improvement compared to
> previous
> >>>>>>>> versions of Flink. Also, the backpressure based checkpoint
> alignment
> >>>> can
> >>>>>>>> help the barrier advance faster on the receiver side (by
> suspending
> >>>>>>>> channels that have already delivered the barrier). Is that
> accurate
> >>>> as of
> >>>>>>>> Flink 1.8?
> >>>>>>>>
> >>>>>>>> What appears to be missing to completely unblock checkpointing is
> a
> >>>>>>>> mechanism for checkpoints to overtake the data. That would help in
> >>>>>>>> situations where the processing itself is the bottleneck and
> >>>> prioritization
> >>>>>>>> in the network stack alone cannot address the barrier delay. Was
> >>>> there any
> >>>>>>>> related discussion? One possible solution would be to drain
> incoming
> >>>> data
> >>>>>>>> till the barrier and make it part of the checkpoint instead of
> >>>> processing
> >>>>>>>> it. This is somewhat related to asynchronous processing, but I'm
> >>>> thinking
> >>>>>>>> more of a solution that is automated in the Flink runtime for the
> >>>>>>>> backpressure scenario only.
> >>>>>>>>
> >>>>>>>> Thanks,
> >>>>>>>> Thomas
> >>>>>>>>
> >>>>>>>> [1] https://flink.apache.org/2019/06/05/flink-network-stack.html
> >>>>>>>> [2]
> >>>>>>>>
> >>>>
> https://docs.google.com/document/d/1chTOuOqe0sBsjldA_r-wXYeSIhU2zRGpUaTaik7QZ84/edit#heading=h.pjh6mv7m2hjn
> >>>>>>>>
> >>>>>>>
> >>>>>>
> >>>>>
> >>>>
> >>>>
> >>
> >
>
>

Paris Carbone-3

Re: Checkpointing under backpressure

In reply to this post by Yun Gao

Thanks for the responses. Starts getting a bit more clear for everyone now.

@Zhuzhu overlapping unaligned snapshots should be aborted/avoided imho.
@Piotr point II, it was a little too quickly written, sorry about that.
Simply put the two following approaches are equivalent for a valid checkpoint.

1. A -> S -> F
2. S -> F -> L

Other variants introduce issues. For example:

3. L -> S (duplicate side effects)
4. L -> A -> S (again duplicates)
5. S -> L -> A (valid but unnecessary channel blocking)

Abbreviations
—————————

S: Capture state
L: Log pending channels
F: Forward Marker
A: Align markers by blocking non-pending channels

What I like about the Chandy Lamport approach (2.) initiated from sources is that:
- Snapshotting imposes no modification to normal processing.
- Can exploit pipeline parallelism if started from the sources, following the natural processing order without concurrent operations to backends (think of long pipelines, e.g., 30 operators with high parallelism, e.g. 100)

I would love to see a design doc soon and discuss ideas there instead. Might not seem like it but I don’t fancy writing long and boring emails. :)

-Paris

> On 14 Aug 2019, at 13:10, Yun Gao <[hidden email]> wrote:
>
> Hi,
> Very thanks for sharing the thoughts on the unaligned checkpoint !
>
> Another question regarding I 2.C (Performance) by Paris is that do we always snapshot and broadcast the marks once the task receives the first mark from JM o? If so, then we will always need to snapshot all the records before the next barriers in all the input channels . However, I thinks users may be able to tolerate a fixed interval for the checkpointing, and we may postpone the snapshot and broadcast till a configurable fixed time passed. With such a postpone, jobs without back pressure could still avoid most IO operations and storage overhead, and jobs with back pressure could also be able to finish the checkpoint in a fixed interval with less IO overhead.
>
> Best,
> Yun
>
> ------------------------------------------------------------------
> From:Piotr Nowojski <[hidden email]>
> Send Time:2019 Aug. 14 (Wed.) 18:38
> To:Paris Carbone <[hidden email]>
> Cc:dev <[hidden email]>; zhijiang <[hidden email]>; Nico Kruber <[hidden email]>
> Subject:Re: Checkpointing under backpressure
>
> Hi again,
>
> Zhu Zhu let me think about this more. Maybe as Paris is writing, we do not need to block any channels at all, at least assuming credit base flow control. Regarding what should happen with the following checkpoint is another question. Also, should we support concurrent checkpoints and subsuming checkpoints as we do now? Maybe not…
>
> Paris
>
> Re
> I. 2. a) and b) - yes, this would have to be taken into an account
> I. 2. c) and IV. 2. - without those, end to end checkpoint time will probably be longer than it could be. It might affect external systems. For example Kafka, which automatically time outs lingering transactions, and for us, the transaction time is equal to the time between two checkpoints.
>
> II 1. - I’m confused. To make things straight. Flink is currently snapshotting once it receives all of the checkpoint barriers from all of the input channels and only then it broadcasts the checkpoint barrier down the stream. And this is correct from exactly-once perspective.
>
> As far as I understand, your proposal based on Chandy Lamport algorithm, is snapshotting the state of the operator on the first checkpoint barrier, which also looks correct to me.
>
> III. 1. As I responded to Zhu Zhu, let me think a bit more about this.
>
> V. Yes, we still need aligned checkpoints, as they are easier for state migration and upgrades.
>
> Piotrek
>
> > On 14 Aug 2019, at 11:22, Paris Carbone <[hidden email]> wrote:
> >
> > Now I see a little more clearly what you have in mind. Thanks for the explanation!
> > There are a few intermixed concepts here, some how to do with correctness some with performance.
> > Before delving deeper I will just enumerate a few things to make myself a little more helpful if I can.
> >
> > I. Initiation
> > -------------
> >
> > 1. RPC to sources only is a less intrusive way to initiate snapshots since you utilize better pipeline parallelism (only a small subset of tasks is running progressively the protocol at a time, if snapshotting is async the overall overhead might not even be observable).
> >
> > 2. If we really want an RPC to all initiation take notice of the following implications:
> >
> > a. (correctness) RPC calls are not guaranteed to arrive in every task before a marker from a preceding task.
> >
> > b. (correctness) Either the RPC call OR the first arriving marker should initiate the algorithm. Whichever comes first. If you only do it per RPC call then you capture a "late" state that includes side effects of already logged events.
> >
> > c. (performance) Lots of IO will be invoked at the same time on the backend store from all tasks. This might lead to high congestion in async snapshots.
> >
> > II. Capturing State First
> > -------------------------
> >
> > 1. (correctness) Capturing state at the last marker sounds incorrect to me (state contains side effects of already logged events based on the proposed scheme). This results into duplicate processing. No?
> >
> > III. Channel Blocking / "Alignment"
> > -----------------------------------
> >
> > 1. (performance?) What is the added benefit? We dont want a "complete" transactional snapshot, async snapshots are purely for failure-recovery. Thus, I dont see why this needs to be imposed at the expense of performance/throughput. With the proposed scheme the whole dataflow anyway enters snapshotting/logging mode so tasks more or less snapshot concurrently.
> >
> > IV Marker Bypassing
> > -------------------
> >
> > 1. (correctness) This leads to equivalent in-flight snapshots so with some quick thinking correct. I will try to model this later and get back to you in case I find something wrong.
> >
> > 2. (performance) It also sounds like a meaningful optimisation! I like thinking of this as a push-based snapshot. i.e., the producing task somehow triggers forward a consumer/channel to capture its state. By example consider T1 -> |marker t1| -> T2.
> >
> > V. Usage of "Async" Snapshots
> > ---------------------
> >
> > 1. Do you see this as a full replacement of "full" aligned snapshots/savepoints? In my view async shanpshots will be needed from time to time but not as frequently. Yet, it seems like a valid approach solely for failure-recovery on the same configuration. Here's why:
> >
> > a. With original snapshotting there is a strong duality between
> > a stream input (offsets) and committed side effects (internal states and external commits to transactional sinks). While in the async version, there are uncommitted operations (inflight records). Thus, you cannot use these snapshots for e.g., submitting sql queries with snapshot isolation. Also, the original snapshotting gives a lot of potential for flink to make proper transactional commits externally.
> >
> > b. Reconfiguration is very tricky, you probably know that better. Inflight channel state is no longer valid in a new configuration (i.e., new dataflow graph, new operators, updated operator logic, different channels, different parallelism)
> >
> > 2. Async snapshots can also be potentially useful for monitoring the general health of a dataflow since they can be analyzed by the task manager about the general performance of a job graph and spot bottlenecks for example.
> >
> >> On 14 Aug 2019, at 09:08, Piotr Nowojski <[hidden email]> wrote:
> >>
> >> Hi,
> >>
> >> Thomas:
> >> There are no Jira tickets yet (or maybe there is something very old somewhere). First we want to discuss it, next present FLIP and at last create tickets :)
> >>
> >>> if I understand correctly, then the proposal is to not block any
> >>> input channel at all, but only log data from the backpressured channel (and
> >>> make it part of the snapshot) until the barrier arrives
> >>
> >> I would guess that it would be better to block the reads, unless we can already process the records from the blocked channel…
> >>
> >> Paris:
> >>
> >> Thanks for the explanation Paris. I’m starting to understand this more and I like the idea of snapshotting the state of an operator before receiving all of the checkpoint barriers - this would allow more things to happen at the same time instead of sequentially. As Zhijiang has pointed out there are some things not considered in your proposal: overtaking output buffers, but maybe those things could be incorporated together.
> >>
> >> Another thing is that from the wiki description I understood that the initial checkpointing is not initialised by any checkpoint barrier, but by an independent call/message from the Observer. I haven’t played with this idea a lot, but I had some discussion with Nico and it seems that it might work:
> >>
> >> 1. JobManager sends and RPC “start checkpoint” to all tasks
> >> 2. Task (with two input channels l1 and l2) upon receiving RPC from 1., takes a snapshot of it's state and:
> >> a) broadcast checkpoint barrier down the stream to all channels (let’s ignore for a moment potential for this barrier to overtake the buffer output data)
> >> b) for any input channel for which it hasn’t yet received checkpoint barrier, the data are being added to the checkpoint
> >> c) once a channel (for example l1) receives a checkpoint barrier, the Task blocks reads from that channel (?)
> >> d) after all remaining channels (l2) receive checkpoint barriers, the Task first has to process the buffered data after that it can unblock the reads from the channels
> >>
> >> Checkpoint barriers do not cascade/flow through different tasks here. Checkpoint barrier emitted from Task1, reaches only the immediate downstream Tasks. Thanks to this setup, total checkpointing time is not sum of checkpointing times of all Tasks one by one, but more or less max of the slowest Tasks. Right?
> >>
> >> Couple of intriguing thoughts are:
> >> 3. checkpoint barriers overtaking the output buffers
> >> 4. can we keep processing some data (in order to not waste CPU cycles) after we have taking the snapshot of the Task. I think we could.
> >>
> >> Piotrek
> >>
> >>> On 14 Aug 2019, at 06:00, Thomas Weise <[hidden email]> wrote:
> >>>
> >>> Great discussion! I'm excited that this is already under consideration! Are
> >>> there any JIRAs or other traces of discussion to follow?
> >>>
> >>> Paris, if I understand correctly, then the proposal is to not block any
> >>> input channel at all, but only log data from the backpressured channel (and
> >>> make it part of the snapshot) until the barrier arrives? This is
> >>> intriguing. But probably there is also a benefit of to not continue reading
> >>> I1 since that could speed up retrieval from I2. Also, if the user code is
> >>> the cause of backpressure, this would avoid pumping more data into the
> >>> process function.
> >>>
> >>> Thanks,
> >>> Thomas
> >>>
> >>>
> >>> On Tue, Aug 13, 2019 at 8:02 AM zhijiang <[hidden email]>
> >>> wrote:
> >>>
> >>>> Hi Paris,
> >>>>
> >>>> Thanks for the detailed sharing. And I think it is very similar with the
> >>>> way of overtaking we proposed before.
> >>>>
> >>>> There are some tiny difference:
> >>>> The way of overtaking might need to snapshot all the input/output queues.
> >>>> Chandy Lamport seems only need to snaphost (n-1) input channels after the
> >>>> first barrier arrives, which might reduce the state sizea bit. But normally
> >>>> there should be less buffers for the first input channel with barrier.
> >>>> The output barrier still follows with regular data stream in Chandy
> >>>> Lamport, the same way as current flink. For overtaking way, we need to pay
> >>>> extra efforts to make barrier transport firstly before outque queue on
> >>>> upstream side, and change the way of barrier alignment based on receiving
> >>>> instead of current reading on downstream side.
> >>>> In the backpressure caused by data skew, the first barrier in almost empty
> >>>> input channel should arrive much eariler than the last heavy load input
> >>>> channel, so the Chandy Lamport could benefit well. But for the case of all
> >>>> balanced heavy load input channels, I mean the first arrived barrier might
> >>>> still take much time, then the overtaking way could still fit well to speed
> >>>> up checkpoint.
> >>>> Anyway, your proposed suggestion is helpful on my side, especially
> >>>> considering some implementation details .
> >>>>
> >>>> Best,
> >>>> Zhijiang
> >>>> ------------------------------------------------------------------
> >>>> From:Paris Carbone <[hidden email]>
> >>>> Send Time:2019年8月13日(星期二) 14:03
> >>>> To:dev <[hidden email]>
> >>>> Cc:zhijiang <[hidden email]>
> >>>> Subject:Re: Checkpointing under backpressure
> >>>>
> >>>> yes! It’s quite similar I think. Though mind that the devil is in the
> >>>> details, i.e., the temporal order actions are taken.
> >>>>
> >>>> To clarify, let us say you have a task T with two input channels I1 and I2.
> >>>> The Chandy Lamport execution flow is the following:
> >>>>
> >>>> 1) T receives barrier from I1 and...
> >>>> 2) ...the following three actions happen atomically
> >>>> I ) T snapshots its state T*
> >>>> II) T forwards marker to its outputs
> >>>> III) T starts logging all events of I2 (only) into a buffer M*
> >>>> - Also notice here that T does NOT block I1 as it does in aligned
> >>>> snapshots -
> >>>> 3) Eventually T receives barrier from I2 and stops recording events. Its
> >>>> asynchronously captured snapshot is now complete: {T*,M*}.
> >>>> Upon recovery all messages of M* should be replayed in FIFO order.
> >>>>
> >>>> With this approach alignment does not create a deadlock situation since
> >>>> anyway 2.II happens asynchronously and messages can be logged as well
> >>>> asynchronously during the process of the snapshot. If there is
> >>>> back-pressure in a pipeline the cause is most probably not this algorithm.
> >>>>
> >>>> Back to your observation, the answer : yes and no. In your network model,
> >>>> I can see the logic of “logging” and “committing” a final snapshot being
> >>>> provided by the channel implementation. However, do mind that the first
> >>>> barrier always needs to go “all the way” to initiate the Chandy Lamport
> >>>> algorithm logic.
> >>>>
> >>>> The above flow has been proven using temporal logic in my phd thesis in
> >>>> case you are interested about the proof.
> >>>> I hope this helps a little clarifying things. Let me know if there is any
> >>>> confusing point to disambiguate. I would be more than happy to help if I
> >>>> can.
> >>>>
> >>>> Paris
> >>>>
> >>>>> On 13 Aug 2019, at 13:28, Piotr Nowojski <[hidden email]> wrote:
> >>>>>
> >>>>> Thanks for the input. Regarding the Chandy-Lamport snapshots don’t you
> >>>> still have to wait for the “checkpoint barrier” to arrive in order to know
> >>>> when have you already received all possible messages from the upstream
> >>>> tasks/operators? So instead of processing the “in flight” messages (as the
> >>>> Flink is doing currently), you are sending them to an “observer”?
> >>>>>
> >>>>> In that case, that’s sounds similar to “checkpoint barriers overtaking
> >>>> in flight records” (aka unaligned checkpoints). Just for us, the observer
> >>>> is a snapshot state.
> >>>>>
> >>>>> Piotrek
> >>>>>
> >>>>>> On 13 Aug 2019, at 13:14, Paris Carbone <[hidden email]>
> >>>> wrote:
> >>>>>>
> >>>>>> Interesting problem! Thanks for bringing it up Thomas.
> >>>>>>
> >>>>>> Ignore/Correct me if I am wrong but I believe Chandy-Lamport snapshots
> >>>> [1] would help out solve this problem more elegantly without sacrificing
> >>>> correctness.
> >>>>>> - They do not need alignment, only (async) logging for in-flight
> >>>> records between the time the first barrier is processed until the last
> >>>> barrier arrives in a task.
> >>>>>> - They work fine for failure recovery as long as logged records are
> >>>> replayed on startup.
> >>>>>>
> >>>>>> Flink’s “alligned” savepoints would probably be still necessary for
> >>>> transactional sink commits + any sort of reconfiguration (e.g., rescaling,
> >>>> updating the logic of operators to evolve an application etc.).
> >>>>>>
> >>>>>> I don’t completely understand the “overtaking” approach but if you have
> >>>> a concrete definition I would be happy to check it out and help if I can!
> >>>>>> Mind that Chandy-Lamport essentially does this by logging things in
> >>>> pending channels in a task snapshot before the barrier arrives.
> >>>>>>
> >>>>>> -Paris
> >>>>>>
> >>>>>> [1] https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm <https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm> <
> >>>> https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm> <https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm%3E>
> >>>>>>
> >>>>>>> On 13 Aug 2019, at 10:27, Piotr Nowojski <[hidden email]> wrote:
> >>>>>>>
> >>>>>>> Hi Thomas,
> >>>>>>>
> >>>>>>> As Zhijiang has responded, we are now in the process of discussing how
> >>>> to address this issue and one of the solution that we are discussing is
> >>>> exactly what you are proposing: checkpoint barriers overtaking the in
> >>>> flight data and make the in flight data part of the checkpoint.
> >>>>>>>
> >>>>>>> If everything works well, we will be able to present result of our
> >>>> discussions on the dev mailing list soon.
> >>>>>>>
> >>>>>>> Piotrek
> >>>>>>>
> >>>>>>>> On 12 Aug 2019, at 23:23, zhijiang <[hidden email]>
> >>>> wrote:
> >>>>>>>>
> >>>>>>>> Hi Thomas,
> >>>>>>>>
> >>>>>>>> Thanks for proposing this concern. The barrier alignment takes long
> >>>> time in backpressure case which could cause several problems:
> >>>>>>>> 1. Checkpoint timeout as you mentioned.
> >>>>>>>> 2. The recovery cost is high once failover, because much data needs
> >>>> to be replayed.
> >>>>>>>> 3. The delay for commit-based sink is high in exactly-once.
> >>>>>>>>
> >>>>>>>> For credit-based flow control from release-1.5, the amount of
> >>>> in-flighting buffers before barrier alignment is reduced, so we could get a
> >>>> bit
> >>>>>>>> benefits from speeding checkpoint aspect.
> >>>>>>>>
> >>>>>>>> In release-1.8, I guess we did not suspend the channels which already
> >>>> received the barrier in practice. But actually we ever did the similar thing
> >>>>>>>> to speed barrier alighment before. I am not quite sure that
> >>>> release-1.8 covers this feature. There were some relevant discussions under
> >>>> jira [1].
> >>>>>>>>
> >>>>>>>> For release-1.10, the community is now discussing the feature of
> >>>> unaligned checkpoint which is mainly for resolving above concerns. The
> >>>> basic idea
> >>>>>>>> is to make barrier overtakes the output/input buffer queue to speed
> >>>> alignment, and snapshot the input/output buffers as part of checkpoint
> >>>> state. The
> >>>>>>>> details have not confirmed yet and is still under discussion. Wish we
> >>>> could make some improvments for the release-1.10.
> >>>>>>>>
> >>>>>>>> [1] https://issues.apache.org/jira/browse/FLINK-8523 <https://issues.apache.org/jira/browse/FLINK-8523>
> >>>>>>>>
> >>>>>>>> Best,
> >>>>>>>> Zhijiang
> >>>>>>>> ------------------------------------------------------------------
> >>>>>>>> From:Thomas Weise <[hidden email]>
> >>>>>>>> Send Time:2019年8月12日(星期一) 21:38
> >>>>>>>> To:dev <[hidden email]>
> >>>>>>>> Subject:Checkpointing under backpressure
> >>>>>>>>
> >>>>>>>> Hi,
> >>>>>>>>
> >>>>>>>> One of the major operational difficulties we observe with Flink are
> >>>>>>>> checkpoint timeouts under backpressure. I'm looking for both
> >>>> confirmation
> >>>>>>>> of my understanding of the current behavior as well as pointers for
> >>>> future
> >>>>>>>> improvement work:
> >>>>>>>>
> >>>>>>>> Prior to introduction of credit based flow control in the network
> >>>> stack [1]
> >>>>>>>> [2], checkpoint barriers would back up with the data for all logical
> >>>>>>>> channels due to TCP backpressure. Since Flink 1.5, the buffers are
> >>>>>>>> controlled per channel, and checkpoint barriers are only held back for
> >>>>>>>> channels that have backpressure, while others can continue processing
> >>>>>>>> normally. However, checkpoint barriers still cannot "overtake data",
> >>>>>>>> therefore checkpoint alignment remains affected for the channel with
> >>>>>>>> backpressure, with the potential for slow checkpointing and timeouts.
> >>>>>>>> Albeit the delay of barriers would be capped by the maximum in-transit
> >>>>>>>> buffers per channel, resulting in an improvement compared to previous
> >>>>>>>> versions of Flink. Also, the backpressure based checkpoint alignment
> >>>> can
> >>>>>>>> help the barrier advance faster on the receiver side (by suspending
> >>>>>>>> channels that have already delivered the barrier). Is that accurate
> >>>> as of
> >>>>>>>> Flink 1.8?
> >>>>>>>>
> >>>>>>>> What appears to be missing to completely unblock checkpointing is a
> >>>>>>>> mechanism for checkpoints to overtake the data. That would help in
> >>>>>>>> situations where the processing itself is the bottleneck and
> >>>> prioritization
> >>>>>>>> in the network stack alone cannot address the barrier delay. Was
> >>>> there any
> >>>>>>>> related discussion? One possible solution would be to drain incoming
> >>>> data
> >>>>>>>> till the barrier and make it part of the checkpoint instead of
> >>>> processing
> >>>>>>>> it. This is somewhat related to asynchronous processing, but I'm
> >>>> thinking
> >>>>>>>> more of a solution that is automated in the Flink runtime for the
> >>>>>>>> backpressure scenario only.
> >>>>>>>>
> >>>>>>>> Thanks,
> >>>>>>>> Thomas
> >>>>>>>>
> >>>>>>>> [1] https://flink.apache.org/2019/06/05/flink-network-stack.html <https://flink.apache.org/2019/06/05/flink-network-stack.html>
> >>>>>>>> [2]
> >>>>>>>>
> >>>> https://docs.google.com/document/d/1chTOuOqe0sBsjldA_r-wXYeSIhU2zRGpUaTaik7QZ84/edit#heading=h.pjh6mv7m2hjn <https://docs.google.com/document/d/1chTOuOqe0sBsjldA_r-wXYeSIhU2zRGpUaTaik7QZ84/edit#heading=h.pjh6mv7m2hjn>
> >>>>>>>>
> >>>>>>>
> >>>>>>
> >>>>>
> >>>>
> >>>>
> >>
> >
>
>

Paris Carbone-3

Re: Checkpointing under backpressure

In reply to this post by Stephan Ewen

+1 I think we are on the same page Stephan.

Rescaling on unaligned checkpoint sounds challenging and a bit unnecessary. No?
Why not sticking to aligned snapshots for live reconfiguration/rescaling? It’s a pretty rare operation and it would simplify things by a lot. Everything can be “staged” upon alignment including replacing channels and tasks.

-Paris

> On 14 Aug 2019, at 13:39, Stephan Ewen <[hidden email]> wrote:
>
> Hi all!
>
> Yes, the first proposal of "unaligend checkpoints" (probably two years back
> now) drew a major inspiration from Chandy Lamport, as did actually the
> original checkpointing algorithm.
>
> "Logging data between first and last barrier" versus "barrier jumping over
> buffer and storing those buffers" is pretty close same.
> However, there are a few nice benefits of the proposal of unaligned
> checkpoints over Chandy-Lamport.
>
> *## Benefits of Unaligned Checkpoints*
>
> (1) It is very similar to the original algorithm (can be seen an an
> optional feature purely in the network stack) and thus can share lot's of
> code paths.
>
> (2) Less data stored. If we make the "jump over buffers" part timeout based
> (for example barrier overtakes buffers if not flushed within 10ms) then
> checkpoints are in the common case of flowing pipelines aligned without
> in-flight data. Only back pressured cases store some in-flight data, which
> means we don't regress in the common case and only fix the back pressure
> case.
>
> (3) Faster checkpoints. Chandy Lamport still waits for all barriers to
> arrive naturally, logging on the way. If data processing is slow, this can
> still take quite a while.
>
> ==> I think both these points are strong reasons to not change the
> mechanism away from "trigger sources" and start with CL-style "trigger all".
>
>
> *## Possible ways to combine Chandy Lamport and Unaligned Checkpoints*
>
> We can think about something like "take state snapshot on first barrier"
> and then store buffers until the other barriers arrive. Inside the network
> stack, barriers could still overtake and persist buffers.
> The benefit would be less latency increase in the channels which already
> have received barriers.
> However, as mentioned before, not prioritizing the inputs from which
> barriers are still missing can also have an adverse effect.
>
>
> *## Concerning upgrades*
>
> I think it is a fair restriction to say that upgrades need to happen on
> aligned checkpoints. It is a rare enough operation.
>
>
> *## Concerning re-scaling (changing parallelism)*
>
> We need to support that on unaligned checkpoints as well. There are several
> feature proposals about automatic scaling, especially down scaling in case
> of missing resources. The last snapshot might be a regular checkpoint, so
> all checkpoints need to support rescaling.
>
>
> *## Concerning end-to-end checkpoint duration and "trigger sources" versus
> "trigger all"*
>
> I think for the end-to-end checkpoint duration, an "overtake buffers"
> approach yields faster checkpoints, as mentioned above (Chandy Lamport
> logging still needs to wait for barrier to flow).
>
> I don't see the benefit of a "trigger all tasks via RPC concurrently"
> approach. Bear in mind that it is still a globally coordinated approach and
> you need to wait for the global checkpoint to complete before committing
> any side effects.
> I believe that the checkpoint time is more determined by the state
> checkpoint writing, and the global coordination and metadata commit, than
> by the difference in alignment time between "trigger from source and jump
> over buffers" versus "trigger all tasks concurrently".
>
> Trying to optimize a few tens of milliseconds out of the network stack
> sends (and changing the overall checkpointing approach completely for that)
> while staying with a globally coordinated checkpoint will send us down a
> path to a dead end.
>
> To really bring task persistence latency down to 10s of milliseconds (so we
> can frequently commit in sinks), we need to take an approach without any
> global coordination. Tasks need to establish a persistent recovery point
> individually and at their own discretion, only then can it be frequent
> enough. To get there, they would need to decouple themselves from the
> predecessor and successor tasks (via something like persistent channels).
> This is a different discussion, though, somewhat orthogonal to this one
> here.
>
> Best,
> Stephan
>
>
> On Wed, Aug 14, 2019 at 12:37 PM Piotr Nowojski <[hidden email]> wrote:
>
>> Hi again,
>>
>> Zhu Zhu let me think about this more. Maybe as Paris is writing, we do not
>> need to block any channels at all, at least assuming credit base flow
>> control. Regarding what should happen with the following checkpoint is
>> another question. Also, should we support concurrent checkpoints and
>> subsuming checkpoints as we do now? Maybe not…
>>
>> Paris
>>
>> Re
>> I. 2. a) and b) - yes, this would have to be taken into an account
>> I. 2. c) and IV. 2. - without those, end to end checkpoint time will
>> probably be longer than it could be. It might affect external systems. For
>> example Kafka, which automatically time outs lingering transactions, and
>> for us, the transaction time is equal to the time between two checkpoints.
>>
>> II 1. - I’m confused. To make things straight. Flink is currently
>> snapshotting once it receives all of the checkpoint barriers from all of
>> the input channels and only then it broadcasts the checkpoint barrier down
>> the stream. And this is correct from exactly-once perspective.
>>
>> As far as I understand, your proposal based on Chandy Lamport algorithm,
>> is snapshotting the state of the operator on the first checkpoint barrier,
>> which also looks correct to me.
>>
>> III. 1. As I responded to Zhu Zhu, let me think a bit more about this.
>>
>> V. Yes, we still need aligned checkpoints, as they are easier for state
>> migration and upgrades.
>>
>> Piotrek
>>
>>> On 14 Aug 2019, at 11:22, Paris Carbone <[hidden email]> wrote:
>>>
>>> Now I see a little more clearly what you have in mind. Thanks for the
>> explanation!
>>> There are a few intermixed concepts here, some how to do with
>> correctness some with performance.
>>> Before delving deeper I will just enumerate a few things to make myself
>> a little more helpful if I can.
>>>
>>> I. Initiation
>>> -------------
>>>
>>> 1. RPC to sources only is a less intrusive way to initiate snapshots
>> since you utilize better pipeline parallelism (only a small subset of tasks
>> is running progressively the protocol at a time, if snapshotting is async
>> the overall overhead might not even be observable).
>>>
>>> 2. If we really want an RPC to all initiation take notice of the
>> following implications:
>>>
>>> a. (correctness) RPC calls are not guaranteed to arrive in every
>> task before a marker from a preceding task.
>>>
>>> b. (correctness) Either the RPC call OR the first arriving marker
>> should initiate the algorithm. Whichever comes first. If you only do it per
>> RPC call then you capture a "late" state that includes side effects of
>> already logged events.
>>>
>>> c. (performance) Lots of IO will be invoked at the same time on
>> the backend store from all tasks. This might lead to high congestion in
>> async snapshots.
>>>
>>> II. Capturing State First
>>> -------------------------
>>>
>>> 1. (correctness) Capturing state at the last marker sounds incorrect to
>> me (state contains side effects of already logged events based on the
>> proposed scheme). This results into duplicate processing. No?
>>>
>>> III. Channel Blocking / "Alignment"
>>> -----------------------------------
>>>
>>> 1. (performance?) What is the added benefit? We dont want a "complete"
>> transactional snapshot, async snapshots are purely for failure-recovery.
>> Thus, I dont see why this needs to be imposed at the expense of
>> performance/throughput. With the proposed scheme the whole dataflow anyway
>> enters snapshotting/logging mode so tasks more or less snapshot
>> concurrently.
>>>
>>> IV Marker Bypassing
>>> -------------------
>>>
>>> 1. (correctness) This leads to equivalent in-flight snapshots so with
>> some quick thinking correct. I will try to model this later and get back
>> to you in case I find something wrong.
>>>
>>> 2. (performance) It also sounds like a meaningful optimisation! I like
>> thinking of this as a push-based snapshot. i.e., the producing task somehow
>> triggers forward a consumer/channel to capture its state. By example
>> consider T1 -> |marker t1| -> T2.
>>>
>>> V. Usage of "Async" Snapshots
>>> ---------------------
>>>
>>> 1. Do you see this as a full replacement of "full" aligned
>> snapshots/savepoints? In my view async shanpshots will be needed from time
>> to time but not as frequently. Yet, it seems like a valid approach solely
>> for failure-recovery on the same configuration. Here's why:
>>>
>>> a. With original snapshotting there is a strong duality between
>>> a stream input (offsets) and committed side effects (internal
>> states and external commits to transactional sinks). While in the async
>> version, there are uncommitted operations (inflight records). Thus, you
>> cannot use these snapshots for e.g., submitting sql queries with snapshot
>> isolation. Also, the original snapshotting gives a lot of potential for
>> flink to make proper transactional commits externally.
>>>
>>> b. Reconfiguration is very tricky, you probably know that better.
>> Inflight channel state is no longer valid in a new configuration (i.e., new
>> dataflow graph, new operators, updated operator logic, different channels,
>> different parallelism)
>>>
>>> 2. Async snapshots can also be potentially useful for monitoring the
>> general health of a dataflow since they can be analyzed by the task manager
>> about the general performance of a job graph and spot bottlenecks for
>> example.
>>>
>>>> On 14 Aug 2019, at 09:08, Piotr Nowojski <[hidden email]> wrote:
>>>>
>>>> Hi,
>>>>
>>>> Thomas:
>>>> There are no Jira tickets yet (or maybe there is something very old
>> somewhere). First we want to discuss it, next present FLIP and at last
>> create tickets :)
>>>>
>>>>> if I understand correctly, then the proposal is to not block any
>>>>> input channel at all, but only log data from the backpressured channel
>> (and
>>>>> make it part of the snapshot) until the barrier arrives
>>>>
>>>> I would guess that it would be better to block the reads, unless we can
>> already process the records from the blocked channel…
>>>>
>>>> Paris:
>>>>
>>>> Thanks for the explanation Paris. I’m starting to understand this more
>> and I like the idea of snapshotting the state of an operator before
>> receiving all of the checkpoint barriers - this would allow more things to
>> happen at the same time instead of sequentially. As Zhijiang has pointed
>> out there are some things not considered in your proposal: overtaking
>> output buffers, but maybe those things could be incorporated together.
>>>>
>>>> Another thing is that from the wiki description I understood that the
>> initial checkpointing is not initialised by any checkpoint barrier, but by
>> an independent call/message from the Observer. I haven’t played with this
>> idea a lot, but I had some discussion with Nico and it seems that it might
>> work:
>>>>
>>>> 1. JobManager sends and RPC “start checkpoint” to all tasks
>>>> 2. Task (with two input channels l1 and l2) upon receiving RPC from 1.,
>> takes a snapshot of it's state and:
>>>> a) broadcast checkpoint barrier down the stream to all channels (let’s
>> ignore for a moment potential for this barrier to overtake the buffer
>> output data)
>>>> b) for any input channel for which it hasn’t yet received checkpoint
>> barrier, the data are being added to the checkpoint
>>>> c) once a channel (for example l1) receives a checkpoint barrier, the
>> Task blocks reads from that channel (?)
>>>> d) after all remaining channels (l2) receive checkpoint barriers, the
>> Task first has to process the buffered data after that it can unblock the
>> reads from the channels
>>>>
>>>> Checkpoint barriers do not cascade/flow through different tasks here.
>> Checkpoint barrier emitted from Task1, reaches only the immediate
>> downstream Tasks. Thanks to this setup, total checkpointing time is not sum
>> of checkpointing times of all Tasks one by one, but more or less max of the
>> slowest Tasks. Right?
>>>>
>>>> Couple of intriguing thoughts are:
>>>> 3. checkpoint barriers overtaking the output buffers
>>>> 4. can we keep processing some data (in order to not waste CPU cycles)
>> after we have taking the snapshot of the Task. I think we could.
>>>>
>>>> Piotrek
>>>>
>>>>> On 14 Aug 2019, at 06:00, Thomas Weise <[hidden email]> wrote:
>>>>>
>>>>> Great discussion! I'm excited that this is already under
>> consideration! Are
>>>>> there any JIRAs or other traces of discussion to follow?
>>>>>
>>>>> Paris, if I understand correctly, then the proposal is to not block any
>>>>> input channel at all, but only log data from the backpressured channel
>> (and
>>>>> make it part of the snapshot) until the barrier arrives? This is
>>>>> intriguing. But probably there is also a benefit of to not continue
>> reading
>>>>> I1 since that could speed up retrieval from I2. Also, if the user code
>> is
>>>>> the cause of backpressure, this would avoid pumping more data into the
>>>>> process function.
>>>>>
>>>>> Thanks,
>>>>> Thomas
>>>>>
>>>>>
>>>>> On Tue, Aug 13, 2019 at 8:02 AM zhijiang <[hidden email]
>> .invalid>
>>>>> wrote:
>>>>>
>>>>>> Hi Paris,
>>>>>>
>>>>>> Thanks for the detailed sharing. And I think it is very similar with
>> the
>>>>>> way of overtaking we proposed before.
>>>>>>
>>>>>> There are some tiny difference:
>>>>>> The way of overtaking might need to snapshot all the input/output
>> queues.
>>>>>> Chandy Lamport seems only need to snaphost (n-1) input channels after
>> the
>>>>>> first barrier arrives, which might reduce the state sizea bit. But
>> normally
>>>>>> there should be less buffers for the first input channel with barrier.
>>>>>> The output barrier still follows with regular data stream in Chandy
>>>>>> Lamport, the same way as current flink. For overtaking way, we need
>> to pay
>>>>>> extra efforts to make barrier transport firstly before outque queue on
>>>>>> upstream side, and change the way of barrier alignment based on
>> receiving
>>>>>> instead of current reading on downstream side.
>>>>>> In the backpressure caused by data skew, the first barrier in almost
>> empty
>>>>>> input channel should arrive much eariler than the last heavy load
>> input
>>>>>> channel, so the Chandy Lamport could benefit well. But for the case
>> of all
>>>>>> balanced heavy load input channels, I mean the first arrived barrier
>> might
>>>>>> still take much time, then the overtaking way could still fit well to
>> speed
>>>>>> up checkpoint.
>>>>>> Anyway, your proposed suggestion is helpful on my side, especially
>>>>>> considering some implementation details .
>>>>>>
>>>>>> Best,
>>>>>> Zhijiang
>>>>>> ------------------------------------------------------------------
>>>>>> From:Paris Carbone <[hidden email]>
>>>>>> Send Time:2019年8月13日(星期二) 14:03
>>>>>> To:dev <[hidden email]>
>>>>>> Cc:zhijiang <[hidden email]>
>>>>>> Subject:Re: Checkpointing under backpressure
>>>>>>
>>>>>> yes! It’s quite similar I think. Though mind that the devil is in the
>>>>>> details, i.e., the temporal order actions are taken.
>>>>>>
>>>>>> To clarify, let us say you have a task T with two input channels I1
>> and I2.
>>>>>> The Chandy Lamport execution flow is the following:
>>>>>>
>>>>>> 1) T receives barrier from I1 and...
>>>>>> 2) ...the following three actions happen atomically
>>>>>> I ) T snapshots its state T*
>>>>>> II) T forwards marker to its outputs
>>>>>> III) T starts logging all events of I2 (only) into a buffer M*
>>>>>> - Also notice here that T does NOT block I1 as it does in aligned
>>>>>> snapshots -
>>>>>> 3) Eventually T receives barrier from I2 and stops recording events.
>> Its
>>>>>> asynchronously captured snapshot is now complete: {T*,M*}.
>>>>>> Upon recovery all messages of M* should be replayed in FIFO order.
>>>>>>
>>>>>> With this approach alignment does not create a deadlock situation
>> since
>>>>>> anyway 2.II happens asynchronously and messages can be logged as well
>>>>>> asynchronously during the process of the snapshot. If there is
>>>>>> back-pressure in a pipeline the cause is most probably not this
>> algorithm.
>>>>>>
>>>>>> Back to your observation, the answer : yes and no. In your network
>> model,
>>>>>> I can see the logic of “logging” and “committing” a final snapshot
>> being
>>>>>> provided by the channel implementation. However, do mind that the
>> first
>>>>>> barrier always needs to go “all the way” to initiate the Chandy
>> Lamport
>>>>>> algorithm logic.
>>>>>>
>>>>>> The above flow has been proven using temporal logic in my phd thesis
>> in
>>>>>> case you are interested about the proof.
>>>>>> I hope this helps a little clarifying things. Let me know if there is
>> any
>>>>>> confusing point to disambiguate. I would be more than happy to help
>> if I
>>>>>> can.
>>>>>>
>>>>>> Paris
>>>>>>
>>>>>>> On 13 Aug 2019, at 13:28, Piotr Nowojski <[hidden email]>
>> wrote:
>>>>>>>
>>>>>>> Thanks for the input. Regarding the Chandy-Lamport snapshots don’t
>> you
>>>>>> still have to wait for the “checkpoint barrier” to arrive in order to
>> know
>>>>>> when have you already received all possible messages from the upstream
>>>>>> tasks/operators? So instead of processing the “in flight” messages
>> (as the
>>>>>> Flink is doing currently), you are sending them to an “observer”?
>>>>>>>
>>>>>>> In that case, that’s sounds similar to “checkpoint barriers
>> overtaking
>>>>>> in flight records” (aka unaligned checkpoints). Just for us, the
>> observer
>>>>>> is a snapshot state.
>>>>>>>
>>>>>>> Piotrek
>>>>>>>
>>>>>>>> On 13 Aug 2019, at 13:14, Paris Carbone <[hidden email]>
>>>>>> wrote:
>>>>>>>>
>>>>>>>> Interesting problem! Thanks for bringing it up Thomas.
>>>>>>>>
>>>>>>>> Ignore/Correct me if I am wrong but I believe Chandy-Lamport
>> snapshots
>>>>>> [1] would help out solve this problem more elegantly without
>> sacrificing
>>>>>> correctness.
>>>>>>>> - They do not need alignment, only (async) logging for in-flight
>>>>>> records between the time the first barrier is processed until the last
>>>>>> barrier arrives in a task.
>>>>>>>> - They work fine for failure recovery as long as logged records are
>>>>>> replayed on startup.
>>>>>>>>
>>>>>>>> Flink’s “alligned” savepoints would probably be still necessary for
>>>>>> transactional sink commits + any sort of reconfiguration (e.g.,
>> rescaling,
>>>>>> updating the logic of operators to evolve an application etc.).
>>>>>>>>
>>>>>>>> I don’t completely understand the “overtaking” approach but if you
>> have
>>>>>> a concrete definition I would be happy to check it out and help if I
>> can!
>>>>>>>> Mind that Chandy-Lamport essentially does this by logging things in
>>>>>> pending channels in a task snapshot before the barrier arrives.
>>>>>>>>
>>>>>>>> -Paris
>>>>>>>>
>>>>>>>> [1] https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm
>> <
>>>>>> https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm>
>>>>>>>>
>>>>>>>>> On 13 Aug 2019, at 10:27, Piotr Nowojski <[hidden email]>
>> wrote:
>>>>>>>>>
>>>>>>>>> Hi Thomas,
>>>>>>>>>
>>>>>>>>> As Zhijiang has responded, we are now in the process of discussing
>> how
>>>>>> to address this issue and one of the solution that we are discussing
>> is
>>>>>> exactly what you are proposing: checkpoint barriers overtaking the in
>>>>>> flight data and make the in flight data part of the checkpoint.
>>>>>>>>>
>>>>>>>>> If everything works well, we will be able to present result of our
>>>>>> discussions on the dev mailing list soon.
>>>>>>>>>
>>>>>>>>> Piotrek
>>>>>>>>>
>>>>>>>>>> On 12 Aug 2019, at 23:23, zhijiang <[hidden email]
>> .INVALID>
>>>>>> wrote:
>>>>>>>>>>
>>>>>>>>>> Hi Thomas,
>>>>>>>>>>
>>>>>>>>>> Thanks for proposing this concern. The barrier alignment takes
>> long
>>>>>> time in backpressure case which could cause several problems:
>>>>>>>>>> 1. Checkpoint timeout as you mentioned.
>>>>>>>>>> 2. The recovery cost is high once failover, because much data
>> needs
>>>>>> to be replayed.
>>>>>>>>>> 3. The delay for commit-based sink is high in exactly-once.
>>>>>>>>>>
>>>>>>>>>> For credit-based flow control from release-1.5, the amount of
>>>>>> in-flighting buffers before barrier alignment is reduced, so we could
>> get a
>>>>>> bit
>>>>>>>>>> benefits from speeding checkpoint aspect.
>>>>>>>>>>
>>>>>>>>>> In release-1.8, I guess we did not suspend the channels which
>> already
>>>>>> received the barrier in practice. But actually we ever did the
>> similar thing
>>>>>>>>>> to speed barrier alighment before. I am not quite sure that
>>>>>> release-1.8 covers this feature. There were some relevant discussions
>> under
>>>>>> jira [1].
>>>>>>>>>>
>>>>>>>>>> For release-1.10, the community is now discussing the feature of
>>>>>> unaligned checkpoint which is mainly for resolving above concerns. The
>>>>>> basic idea
>>>>>>>>>> is to make barrier overtakes the output/input buffer queue to
>> speed
>>>>>> alignment, and snapshot the input/output buffers as part of checkpoint
>>>>>> state. The
>>>>>>>>>> details have not confirmed yet and is still under discussion.
>> Wish we
>>>>>> could make some improvments for the release-1.10.
>>>>>>>>>>
>>>>>>>>>> [1] https://issues.apache.org/jira/browse/FLINK-8523
>>>>>>>>>>
>>>>>>>>>> Best,
>>>>>>>>>> Zhijiang
>>>>>>>>>> ------------------------------------------------------------------
>>>>>>>>>> From:Thomas Weise <[hidden email]>
>>>>>>>>>> Send Time:2019年8月12日(星期一) 21:38
>>>>>>>>>> To:dev <[hidden email]>
>>>>>>>>>> Subject:Checkpointing under backpressure
>>>>>>>>>>
>>>>>>>>>> Hi,
>>>>>>>>>>
>>>>>>>>>> One of the major operational difficulties we observe with Flink
>> are
>>>>>>>>>> checkpoint timeouts under backpressure. I'm looking for both
>>>>>> confirmation
>>>>>>>>>> of my understanding of the current behavior as well as pointers
>> for
>>>>>> future
>>>>>>>>>> improvement work:
>>>>>>>>>>
>>>>>>>>>> Prior to introduction of credit based flow control in the network
>>>>>> stack [1]
>>>>>>>>>> [2], checkpoint barriers would back up with the data for all
>> logical
>>>>>>>>>> channels due to TCP backpressure. Since Flink 1.5, the buffers are
>>>>>>>>>> controlled per channel, and checkpoint barriers are only held
>> back for
>>>>>>>>>> channels that have backpressure, while others can continue
>> processing
>>>>>>>>>> normally. However, checkpoint barriers still cannot "overtake
>> data",
>>>>>>>>>> therefore checkpoint alignment remains affected for the channel
>> with
>>>>>>>>>> backpressure, with the potential for slow checkpointing and
>> timeouts.
>>>>>>>>>> Albeit the delay of barriers would be capped by the maximum
>> in-transit
>>>>>>>>>> buffers per channel, resulting in an improvement compared to
>> previous
>>>>>>>>>> versions of Flink. Also, the backpressure based checkpoint
>> alignment
>>>>>> can
>>>>>>>>>> help the barrier advance faster on the receiver side (by
>> suspending
>>>>>>>>>> channels that have already delivered the barrier). Is that
>> accurate
>>>>>> as of
>>>>>>>>>> Flink 1.8?
>>>>>>>>>>
>>>>>>>>>> What appears to be missing to completely unblock checkpointing is
>> a
>>>>>>>>>> mechanism for checkpoints to overtake the data. That would help in
>>>>>>>>>> situations where the processing itself is the bottleneck and
>>>>>> prioritization
>>>>>>>>>> in the network stack alone cannot address the barrier delay. Was
>>>>>> there any
>>>>>>>>>> related discussion? One possible solution would be to drain
>> incoming
>>>>>> data
>>>>>>>>>> till the barrier and make it part of the checkpoint instead of
>>>>>> processing
>>>>>>>>>> it. This is somewhat related to asynchronous processing, but I'm
>>>>>> thinking
>>>>>>>>>> more of a solution that is automated in the Flink runtime for the
>>>>>>>>>> backpressure scenario only.
>>>>>>>>>>
>>>>>>>>>> Thanks,
>>>>>>>>>> Thomas
>>>>>>>>>>
>>>>>>>>>> [1] https://flink.apache.org/2019/06/05/flink-network-stack.html
>>>>>>>>>> [2]
>>>>>>>>>>
>>>>>>
>> https://docs.google.com/document/d/1chTOuOqe0sBsjldA_r-wXYeSIhU2zRGpUaTaik7QZ84/edit#heading=h.pjh6mv7m2hjn
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>>
>>>>
>>>
>>
>>

Stephan Ewen

Re: Checkpointing under backpressure

Scaling with unaligned checkpoints might be a necessity.

Let's assume the job failed due to a lost TaskManager, but no new
TaskManager becomes available.
In that case we need to scale down based on the latest complete checkpoint,
because we cannot produce a new checkpoint.

On Wed, Aug 14, 2019 at 2:05 PM Paris Carbone <[hidden email]>
wrote:

> +1 I think we are on the same page Stephan.
>
> Rescaling on unaligned checkpoint sounds challenging and a bit
> unnecessary. No?
> Why not sticking to aligned snapshots for live reconfiguration/rescaling?
> It’s a pretty rare operation and it would simplify things by a lot.
> Everything can be “staged” upon alignment including replacing channels and
> tasks.
>
> -Paris
>
> > On 14 Aug 2019, at 13:39, Stephan Ewen <[hidden email]> wrote:
> >
> > Hi all!
> >
> > Yes, the first proposal of "unaligend checkpoints" (probably two years
> back
> > now) drew a major inspiration from Chandy Lamport, as did actually the
> > original checkpointing algorithm.
> >
> > "Logging data between first and last barrier" versus "barrier jumping
> over
> > buffer and storing those buffers" is pretty close same.
> > However, there are a few nice benefits of the proposal of unaligned
> > checkpoints over Chandy-Lamport.
> >
> > *## Benefits of Unaligned Checkpoints*
> >
> > (1) It is very similar to the original algorithm (can be seen an an
> > optional feature purely in the network stack) and thus can share lot's of
> > code paths.
> >
> > (2) Less data stored. If we make the "jump over buffers" part timeout
> based
> > (for example barrier overtakes buffers if not flushed within 10ms) then
> > checkpoints are in the common case of flowing pipelines aligned without
> > in-flight data. Only back pressured cases store some in-flight data,
> which
> > means we don't regress in the common case and only fix the back pressure
> > case.
> >
> > (3) Faster checkpoints. Chandy Lamport still waits for all barriers to
> > arrive naturally, logging on the way. If data processing is slow, this
> can
> > still take quite a while.
> >
> > ==> I think both these points are strong reasons to not change the
> > mechanism away from "trigger sources" and start with CL-style "trigger
> all".
> >
> >
> > *## Possible ways to combine Chandy Lamport and Unaligned Checkpoints*
> >
> > We can think about something like "take state snapshot on first barrier"
> > and then store buffers until the other barriers arrive. Inside the
> network
> > stack, barriers could still overtake and persist buffers.
> > The benefit would be less latency increase in the channels which already
> > have received barriers.
> > However, as mentioned before, not prioritizing the inputs from which
> > barriers are still missing can also have an adverse effect.
> >
> >
> > *## Concerning upgrades*
> >
> > I think it is a fair restriction to say that upgrades need to happen on
> > aligned checkpoints. It is a rare enough operation.
> >
> >
> > *## Concerning re-scaling (changing parallelism)*
> >
> > We need to support that on unaligned checkpoints as well. There are
> several
> > feature proposals about automatic scaling, especially down scaling in
> case
> > of missing resources. The last snapshot might be a regular checkpoint, so
> > all checkpoints need to support rescaling.
> >
> >
> > *## Concerning end-to-end checkpoint duration and "trigger sources"
> versus
> > "trigger all"*
> >
> > I think for the end-to-end checkpoint duration, an "overtake buffers"
> > approach yields faster checkpoints, as mentioned above (Chandy Lamport
> > logging still needs to wait for barrier to flow).
> >
> > I don't see the benefit of a "trigger all tasks via RPC concurrently"
> > approach. Bear in mind that it is still a globally coordinated approach
> and
> > you need to wait for the global checkpoint to complete before committing
> > any side effects.
> > I believe that the checkpoint time is more determined by the state
> > checkpoint writing, and the global coordination and metadata commit, than
> > by the difference in alignment time between "trigger from source and jump
> > over buffers" versus "trigger all tasks concurrently".
> >
> > Trying to optimize a few tens of milliseconds out of the network stack
> > sends (and changing the overall checkpointing approach completely for
> that)
> > while staying with a globally coordinated checkpoint will send us down a
> > path to a dead end.
> >
> > To really bring task persistence latency down to 10s of milliseconds (so
> we
> > can frequently commit in sinks), we need to take an approach without any
> > global coordination. Tasks need to establish a persistent recovery point
> > individually and at their own discretion, only then can it be frequent
> > enough. To get there, they would need to decouple themselves from the
> > predecessor and successor tasks (via something like persistent channels).
> > This is a different discussion, though, somewhat orthogonal to this one
> > here.
> >
> > Best,
> > Stephan
> >
> >
> > On Wed, Aug 14, 2019 at 12:37 PM Piotr Nowojski <[hidden email]>
> wrote:
> >
> >> Hi again,
> >>
> >> Zhu Zhu let me think about this more. Maybe as Paris is writing, we do
> not
> >> need to block any channels at all, at least assuming credit base flow
> >> control. Regarding what should happen with the following checkpoint is
> >> another question. Also, should we support concurrent checkpoints and
> >> subsuming checkpoints as we do now? Maybe not…
> >>
> >> Paris
> >>
> >> Re
> >> I. 2. a) and b) - yes, this would have to be taken into an account
> >> I. 2. c) and IV. 2. - without those, end to end checkpoint time will
> >> probably be longer than it could be. It might affect external systems.
> For
> >> example Kafka, which automatically time outs lingering transactions, and
> >> for us, the transaction time is equal to the time between two
> checkpoints.
> >>
> >> II 1. - I’m confused. To make things straight. Flink is currently
> >> snapshotting once it receives all of the checkpoint barriers from all of
> >> the input channels and only then it broadcasts the checkpoint barrier
> down
> >> the stream. And this is correct from exactly-once perspective.
> >>
> >> As far as I understand, your proposal based on Chandy Lamport algorithm,
> >> is snapshotting the state of the operator on the first checkpoint
> barrier,
> >> which also looks correct to me.
> >>
> >> III. 1. As I responded to Zhu Zhu, let me think a bit more about this.
> >>
> >> V. Yes, we still need aligned checkpoints, as they are easier for state
> >> migration and upgrades.
> >>
> >> Piotrek
> >>
> >>> On 14 Aug 2019, at 11:22, Paris Carbone <[hidden email]>
> wrote:
> >>>
> >>> Now I see a little more clearly what you have in mind. Thanks for the
> >> explanation!
> >>> There are a few intermixed concepts here, some how to do with
> >> correctness some with performance.
> >>> Before delving deeper I will just enumerate a few things to make myself
> >> a little more helpful if I can.
> >>>
> >>> I. Initiation
> >>> -------------
> >>>
> >>> 1. RPC to sources only is a less intrusive way to initiate snapshots
> >> since you utilize better pipeline parallelism (only a small subset of
> tasks
> >> is running progressively the protocol at a time, if snapshotting is
> async
> >> the overall overhead might not even be observable).
> >>>
> >>> 2. If we really want an RPC to all initiation take notice of the
> >> following implications:
> >>>
> >>> a. (correctness) RPC calls are not guaranteed to arrive in every
> >> task before a marker from a preceding task.
> >>>
> >>> b. (correctness) Either the RPC call OR the first arriving marker
> >> should initiate the algorithm. Whichever comes first. If you only do it
> per
> >> RPC call then you capture a "late" state that includes side effects of
> >> already logged events.
> >>>
> >>> c. (performance) Lots of IO will be invoked at the same time on
> >> the backend store from all tasks. This might lead to high congestion in
> >> async snapshots.
> >>>
> >>> II. Capturing State First
> >>> -------------------------
> >>>
> >>> 1. (correctness) Capturing state at the last marker sounds incorrect to
> >> me (state contains side effects of already logged events based on the
> >> proposed scheme). This results into duplicate processing. No?
> >>>
> >>> III. Channel Blocking / "Alignment"
> >>> -----------------------------------
> >>>
> >>> 1. (performance?) What is the added benefit? We dont want a "complete"
> >> transactional snapshot, async snapshots are purely for failure-recovery.
> >> Thus, I dont see why this needs to be imposed at the expense of
> >> performance/throughput. With the proposed scheme the whole dataflow
> anyway
> >> enters snapshotting/logging mode so tasks more or less snapshot
> >> concurrently.
> >>>
> >>> IV Marker Bypassing
> >>> -------------------
> >>>
> >>> 1. (correctness) This leads to equivalent in-flight snapshots so with
> >> some quick thinking correct. I will try to model this later and get
> back
> >> to you in case I find something wrong.
> >>>
> >>> 2. (performance) It also sounds like a meaningful optimisation! I like
> >> thinking of this as a push-based snapshot. i.e., the producing task
> somehow
> >> triggers forward a consumer/channel to capture its state. By example
> >> consider T1 -> |marker t1| -> T2.
> >>>
> >>> V. Usage of "Async" Snapshots
> >>> ---------------------
> >>>
> >>> 1. Do you see this as a full replacement of "full" aligned
> >> snapshots/savepoints? In my view async shanpshots will be needed from
> time
> >> to time but not as frequently. Yet, it seems like a valid approach
> solely
> >> for failure-recovery on the same configuration. Here's why:
> >>>
> >>> a. With original snapshotting there is a strong duality between
> >>> a stream input (offsets) and committed side effects (internal
> >> states and external commits to transactional sinks). While in the async
> >> version, there are uncommitted operations (inflight records). Thus, you
> >> cannot use these snapshots for e.g., submitting sql queries with
> snapshot
> >> isolation. Also, the original snapshotting gives a lot of potential for
> >> flink to make proper transactional commits externally.
> >>>
> >>> b. Reconfiguration is very tricky, you probably know that better.
> >> Inflight channel state is no longer valid in a new configuration (i.e.,
> new
> >> dataflow graph, new operators, updated operator logic, different
> channels,
> >> different parallelism)
> >>>
> >>> 2. Async snapshots can also be potentially useful for monitoring the
> >> general health of a dataflow since they can be analyzed by the task
> manager
> >> about the general performance of a job graph and spot bottlenecks for
> >> example.
> >>>
> >>>> On 14 Aug 2019, at 09:08, Piotr Nowojski <[hidden email]> wrote:
> >>>>
> >>>> Hi,
> >>>>
> >>>> Thomas:
> >>>> There are no Jira tickets yet (or maybe there is something very old
> >> somewhere). First we want to discuss it, next present FLIP and at last
> >> create tickets :)
> >>>>
> >>>>> if I understand correctly, then the proposal is to not block any
> >>>>> input channel at all, but only log data from the backpressured
> channel
> >> (and
> >>>>> make it part of the snapshot) until the barrier arrives
> >>>>
> >>>> I would guess that it would be better to block the reads, unless we
> can
> >> already process the records from the blocked channel…
> >>>>
> >>>> Paris:
> >>>>
> >>>> Thanks for the explanation Paris. I’m starting to understand this more
> >> and I like the idea of snapshotting the state of an operator before
> >> receiving all of the checkpoint barriers - this would allow more things
> to
> >> happen at the same time instead of sequentially. As Zhijiang has pointed
> >> out there are some things not considered in your proposal: overtaking
> >> output buffers, but maybe those things could be incorporated together.
> >>>>
> >>>> Another thing is that from the wiki description I understood that the
> >> initial checkpointing is not initialised by any checkpoint barrier, but
> by
> >> an independent call/message from the Observer. I haven’t played with
> this
> >> idea a lot, but I had some discussion with Nico and it seems that it
> might
> >> work:
> >>>>
> >>>> 1. JobManager sends and RPC “start checkpoint” to all tasks
> >>>> 2. Task (with two input channels l1 and l2) upon receiving RPC from
> 1.,
> >> takes a snapshot of it's state and:
> >>>> a) broadcast checkpoint barrier down the stream to all channels (let’s
> >> ignore for a moment potential for this barrier to overtake the buffer
> >> output data)
> >>>> b) for any input channel for which it hasn’t yet received checkpoint
> >> barrier, the data are being added to the checkpoint
> >>>> c) once a channel (for example l1) receives a checkpoint barrier, the
> >> Task blocks reads from that channel (?)
> >>>> d) after all remaining channels (l2) receive checkpoint barriers, the
> >> Task first has to process the buffered data after that it can unblock
> the
> >> reads from the channels
> >>>>
> >>>> Checkpoint barriers do not cascade/flow through different tasks here.
> >> Checkpoint barrier emitted from Task1, reaches only the immediate
> >> downstream Tasks. Thanks to this setup, total checkpointing time is not
> sum
> >> of checkpointing times of all Tasks one by one, but more or less max of
> the
> >> slowest Tasks. Right?
> >>>>
> >>>> Couple of intriguing thoughts are:
> >>>> 3. checkpoint barriers overtaking the output buffers
> >>>> 4. can we keep processing some data (in order to not waste CPU cycles)
> >> after we have taking the snapshot of the Task. I think we could.
> >>>>
> >>>> Piotrek
> >>>>
> >>>>> On 14 Aug 2019, at 06:00, Thomas Weise <[hidden email]> wrote:
> >>>>>
> >>>>> Great discussion! I'm excited that this is already under
> >> consideration! Are
> >>>>> there any JIRAs or other traces of discussion to follow?
> >>>>>
> >>>>> Paris, if I understand correctly, then the proposal is to not block
> any
> >>>>> input channel at all, but only log data from the backpressured
> channel
> >> (and
> >>>>> make it part of the snapshot) until the barrier arrives? This is
> >>>>> intriguing. But probably there is also a benefit of to not continue
> >> reading
> >>>>> I1 since that could speed up retrieval from I2. Also, if the user
> code
> >> is
> >>>>> the cause of backpressure, this would avoid pumping more data into
> the
> >>>>> process function.
> >>>>>
> >>>>> Thanks,
> >>>>> Thomas
> >>>>>
> >>>>>
> >>>>> On Tue, Aug 13, 2019 at 8:02 AM zhijiang <[hidden email]
> >> .invalid>
> >>>>> wrote:
> >>>>>
> >>>>>> Hi Paris,
> >>>>>>
> >>>>>> Thanks for the detailed sharing. And I think it is very similar with
> >> the
> >>>>>> way of overtaking we proposed before.
> >>>>>>
> >>>>>> There are some tiny difference:
> >>>>>> The way of overtaking might need to snapshot all the input/output
> >> queues.
> >>>>>> Chandy Lamport seems only need to snaphost (n-1) input channels
> after
> >> the
> >>>>>> first barrier arrives, which might reduce the state sizea bit. But
> >> normally
> >>>>>> there should be less buffers for the first input channel with
> barrier.
> >>>>>> The output barrier still follows with regular data stream in Chandy
> >>>>>> Lamport, the same way as current flink. For overtaking way, we need
> >> to pay
> >>>>>> extra efforts to make barrier transport firstly before outque queue
> on
> >>>>>> upstream side, and change the way of barrier alignment based on
> >> receiving
> >>>>>> instead of current reading on downstream side.
> >>>>>> In the backpressure caused by data skew, the first barrier in almost
> >> empty
> >>>>>> input channel should arrive much eariler than the last heavy load
> >> input
> >>>>>> channel, so the Chandy Lamport could benefit well. But for the case
> >> of all
> >>>>>> balanced heavy load input channels, I mean the first arrived barrier
> >> might
> >>>>>> still take much time, then the overtaking way could still fit well
> to
> >> speed
> >>>>>> up checkpoint.
> >>>>>> Anyway, your proposed suggestion is helpful on my side, especially
> >>>>>> considering some implementation details .
> >>>>>>
> >>>>>> Best,
> >>>>>> Zhijiang
> >>>>>> ------------------------------------------------------------------
> >>>>>> From:Paris Carbone <[hidden email]>
> >>>>>> Send Time:2019年8月13日(星期二) 14:03
> >>>>>> To:dev <[hidden email]>
> >>>>>> Cc:zhijiang <[hidden email]>
> >>>>>> Subject:Re: Checkpointing under backpressure
> >>>>>>
> >>>>>> yes! It’s quite similar I think. Though mind that the devil is in
> the
> >>>>>> details, i.e., the temporal order actions are taken.
> >>>>>>
> >>>>>> To clarify, let us say you have a task T with two input channels I1
> >> and I2.
> >>>>>> The Chandy Lamport execution flow is the following:
> >>>>>>
> >>>>>> 1) T receives barrier from I1 and...
> >>>>>> 2) ...the following three actions happen atomically
> >>>>>> I ) T snapshots its state T*
> >>>>>> II) T forwards marker to its outputs
> >>>>>> III) T starts logging all events of I2 (only) into a buffer M*
> >>>>>> - Also notice here that T does NOT block I1 as it does in aligned
> >>>>>> snapshots -
> >>>>>> 3) Eventually T receives barrier from I2 and stops recording events.
> >> Its
> >>>>>> asynchronously captured snapshot is now complete: {T*,M*}.
> >>>>>> Upon recovery all messages of M* should be replayed in FIFO order.
> >>>>>>
> >>>>>> With this approach alignment does not create a deadlock situation
> >> since
> >>>>>> anyway 2.II happens asynchronously and messages can be logged as
> well
> >>>>>> asynchronously during the process of the snapshot. If there is
> >>>>>> back-pressure in a pipeline the cause is most probably not this
> >> algorithm.
> >>>>>>
> >>>>>> Back to your observation, the answer : yes and no. In your network
> >> model,
> >>>>>> I can see the logic of “logging” and “committing” a final snapshot
> >> being
> >>>>>> provided by the channel implementation. However, do mind that the
> >> first
> >>>>>> barrier always needs to go “all the way” to initiate the Chandy
> >> Lamport
> >>>>>> algorithm logic.
> >>>>>>
> >>>>>> The above flow has been proven using temporal logic in my phd thesis
> >> in
> >>>>>> case you are interested about the proof.
> >>>>>> I hope this helps a little clarifying things. Let me know if there
> is
> >> any
> >>>>>> confusing point to disambiguate. I would be more than happy to help
> >> if I
> >>>>>> can.
> >>>>>>
> >>>>>> Paris
> >>>>>>
> >>>>>>> On 13 Aug 2019, at 13:28, Piotr Nowojski <[hidden email]>
> >> wrote:
> >>>>>>>
> >>>>>>> Thanks for the input. Regarding the Chandy-Lamport snapshots don’t
> >> you
> >>>>>> still have to wait for the “checkpoint barrier” to arrive in order
> to
> >> know
> >>>>>> when have you already received all possible messages from the
> upstream
> >>>>>> tasks/operators? So instead of processing the “in flight” messages
> >> (as the
> >>>>>> Flink is doing currently), you are sending them to an “observer”?
> >>>>>>>
> >>>>>>> In that case, that’s sounds similar to “checkpoint barriers
> >> overtaking
> >>>>>> in flight records” (aka unaligned checkpoints). Just for us, the
> >> observer
> >>>>>> is a snapshot state.
> >>>>>>>
> >>>>>>> Piotrek
> >>>>>>>
> >>>>>>>> On 13 Aug 2019, at 13:14, Paris Carbone <[hidden email]>
> >>>>>> wrote:
> >>>>>>>>
> >>>>>>>> Interesting problem! Thanks for bringing it up Thomas.
> >>>>>>>>
> >>>>>>>> Ignore/Correct me if I am wrong but I believe Chandy-Lamport
> >> snapshots
> >>>>>> [1] would help out solve this problem more elegantly without
> >> sacrificing
> >>>>>> correctness.
> >>>>>>>> - They do not need alignment, only (async) logging for in-flight
> >>>>>> records between the time the first barrier is processed until the
> last
> >>>>>> barrier arrives in a task.
> >>>>>>>> - They work fine for failure recovery as long as logged records
> are
> >>>>>> replayed on startup.
> >>>>>>>>
> >>>>>>>> Flink’s “alligned” savepoints would probably be still necessary
> for
> >>>>>> transactional sink commits + any sort of reconfiguration (e.g.,
> >> rescaling,
> >>>>>> updating the logic of operators to evolve an application etc.).
> >>>>>>>>
> >>>>>>>> I don’t completely understand the “overtaking” approach but if you
> >> have
> >>>>>> a concrete definition I would be happy to check it out and help if I
> >> can!
> >>>>>>>> Mind that Chandy-Lamport essentially does this by logging things
> in
> >>>>>> pending channels in a task snapshot before the barrier arrives.
> >>>>>>>>
> >>>>>>>> -Paris
> >>>>>>>>
> >>>>>>>> [1]
> https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm
> >> <
> >>>>>> https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm>
> >>>>>>>>
> >>>>>>>>> On 13 Aug 2019, at 10:27, Piotr Nowojski <[hidden email]>
> >> wrote:
> >>>>>>>>>
> >>>>>>>>> Hi Thomas,
> >>>>>>>>>
> >>>>>>>>> As Zhijiang has responded, we are now in the process of
> discussing
> >> how
> >>>>>> to address this issue and one of the solution that we are discussing
> >> is
> >>>>>> exactly what you are proposing: checkpoint barriers overtaking the
> in
> >>>>>> flight data and make the in flight data part of the checkpoint.
> >>>>>>>>>
> >>>>>>>>> If everything works well, we will be able to present result of
> our
> >>>>>> discussions on the dev mailing list soon.
> >>>>>>>>>
> >>>>>>>>> Piotrek
> >>>>>>>>>
> >>>>>>>>>> On 12 Aug 2019, at 23:23, zhijiang <[hidden email]
> >> .INVALID>
> >>>>>> wrote:
> >>>>>>>>>>
> >>>>>>>>>> Hi Thomas,
> >>>>>>>>>>
> >>>>>>>>>> Thanks for proposing this concern. The barrier alignment takes
> >> long
> >>>>>> time in backpressure case which could cause several problems:
> >>>>>>>>>> 1. Checkpoint timeout as you mentioned.
> >>>>>>>>>> 2. The recovery cost is high once failover, because much data
> >> needs
> >>>>>> to be replayed.
> >>>>>>>>>> 3. The delay for commit-based sink is high in exactly-once.
> >>>>>>>>>>
> >>>>>>>>>> For credit-based flow control from release-1.5, the amount of
> >>>>>> in-flighting buffers before barrier alignment is reduced, so we
> could
> >> get a
> >>>>>> bit
> >>>>>>>>>> benefits from speeding checkpoint aspect.
> >>>>>>>>>>
> >>>>>>>>>> In release-1.8, I guess we did not suspend the channels which
> >> already
> >>>>>> received the barrier in practice. But actually we ever did the
> >> similar thing
> >>>>>>>>>> to speed barrier alighment before. I am not quite sure that
> >>>>>> release-1.8 covers this feature. There were some relevant
> discussions
> >> under
> >>>>>> jira [1].
> >>>>>>>>>>
> >>>>>>>>>> For release-1.10, the community is now discussing the feature of
> >>>>>> unaligned checkpoint which is mainly for resolving above concerns.
> The
> >>>>>> basic idea
> >>>>>>>>>> is to make barrier overtakes the output/input buffer queue to
> >> speed
> >>>>>> alignment, and snapshot the input/output buffers as part of
> checkpoint
> >>>>>> state. The
> >>>>>>>>>> details have not confirmed yet and is still under discussion.
> >> Wish we
> >>>>>> could make some improvments for the release-1.10.
> >>>>>>>>>>
> >>>>>>>>>> [1] https://issues.apache.org/jira/browse/FLINK-8523
> >>>>>>>>>>
> >>>>>>>>>> Best,
> >>>>>>>>>> Zhijiang
> >>>>>>>>>>
> ------------------------------------------------------------------
> >>>>>>>>>> From:Thomas Weise <[hidden email]>
> >>>>>>>>>> Send Time:2019年8月12日(星期一) 21:38
> >>>>>>>>>> To:dev <[hidden email]>
> >>>>>>>>>> Subject:Checkpointing under backpressure
> >>>>>>>>>>
> >>>>>>>>>> Hi,
> >>>>>>>>>>
> >>>>>>>>>> One of the major operational difficulties we observe with Flink
> >> are
> >>>>>>>>>> checkpoint timeouts under backpressure. I'm looking for both
> >>>>>> confirmation
> >>>>>>>>>> of my understanding of the current behavior as well as pointers
> >> for
> >>>>>> future
> >>>>>>>>>> improvement work:
> >>>>>>>>>>
> >>>>>>>>>> Prior to introduction of credit based flow control in the
> network
> >>>>>> stack [1]
> >>>>>>>>>> [2], checkpoint barriers would back up with the data for all
> >> logical
> >>>>>>>>>> channels due to TCP backpressure. Since Flink 1.5, the buffers
> are
> >>>>>>>>>> controlled per channel, and checkpoint barriers are only held
> >> back for
> >>>>>>>>>> channels that have backpressure, while others can continue
> >> processing
> >>>>>>>>>> normally. However, checkpoint barriers still cannot "overtake
> >> data",
> >>>>>>>>>> therefore checkpoint alignment remains affected for the channel
> >> with
> >>>>>>>>>> backpressure, with the potential for slow checkpointing and
> >> timeouts.
> >>>>>>>>>> Albeit the delay of barriers would be capped by the maximum
> >> in-transit
> >>>>>>>>>> buffers per channel, resulting in an improvement compared to
> >> previous
> >>>>>>>>>> versions of Flink. Also, the backpressure based checkpoint
> >> alignment
> >>>>>> can
> >>>>>>>>>> help the barrier advance faster on the receiver side (by
> >> suspending
> >>>>>>>>>> channels that have already delivered the barrier). Is that
> >> accurate
> >>>>>> as of
> >>>>>>>>>> Flink 1.8?
> >>>>>>>>>>
> >>>>>>>>>> What appears to be missing to completely unblock checkpointing
> is
> >> a
> >>>>>>>>>> mechanism for checkpoints to overtake the data. That would help
> in
> >>>>>>>>>> situations where the processing itself is the bottleneck and
> >>>>>> prioritization
> >>>>>>>>>> in the network stack alone cannot address the barrier delay. Was
> >>>>>> there any
> >>>>>>>>>> related discussion? One possible solution would be to drain
> >> incoming
> >>>>>> data
> >>>>>>>>>> till the barrier and make it part of the checkpoint instead of
> >>>>>> processing
> >>>>>>>>>> it. This is somewhat related to asynchronous processing, but I'm
> >>>>>> thinking
> >>>>>>>>>> more of a solution that is automated in the Flink runtime for
> the
> >>>>>>>>>> backpressure scenario only.
> >>>>>>>>>>
> >>>>>>>>>> Thanks,
> >>>>>>>>>> Thomas
> >>>>>>>>>>
> >>>>>>>>>> [1]
> https://flink.apache.org/2019/06/05/flink-network-stack.html
> >>>>>>>>>> [2]
> >>>>>>>>>>
> >>>>>>
> >>
> https://docs.google.com/document/d/1chTOuOqe0sBsjldA_r-wXYeSIhU2zRGpUaTaik7QZ84/edit#heading=h.pjh6mv7m2hjn
> >>>>>>>>>>
> >>>>>>>>>
> >>>>>>>>
> >>>>>>>
> >>>>>>
> >>>>>>
> >>>>
> >>>
> >>
> >>
>
>

Paris Carbone-3

Re: Checkpointing under backpressure

Sure I see. In cases when no periodic aligned snapshots are employed this is the only option.

Two things that were not highlighted enough so far on the proposed protocol (included my mails):
- The Recovery/Reconfiguration strategy should strictly prioritise processing logged events before entering normal task input operation. Otherwise causality can be violated. This also means dataflow recovery will be expected to be slower to the one employed on an aligned snapshot.
- Same as with state capture, markers should be forwarded upon first marker received on input. No later than that. Otherwise we have duplicate side effects.

Thanks for the great ideas so far.

Paris

> On 14 Aug 2019, at 14:33, Stephan Ewen <[hidden email]> wrote:
>
> Scaling with unaligned checkpoints might be a necessity.
>
> Let's assume the job failed due to a lost TaskManager, but no new
> TaskManager becomes available.
> In that case we need to scale down based on the latest complete checkpoint,
> because we cannot produce a new checkpoint.
>
>
> On Wed, Aug 14, 2019 at 2:05 PM Paris Carbone <[hidden email]>
> wrote:
>
>> +1 I think we are on the same page Stephan.
>>
>> Rescaling on unaligned checkpoint sounds challenging and a bit
>> unnecessary. No?
>> Why not sticking to aligned snapshots for live reconfiguration/rescaling?
>> It’s a pretty rare operation and it would simplify things by a lot.
>> Everything can be “staged” upon alignment including replacing channels and
>> tasks.
>>
>> -Paris
>>
>>> On 14 Aug 2019, at 13:39, Stephan Ewen <[hidden email]> wrote:
>>>
>>> Hi all!
>>>
>>> Yes, the first proposal of "unaligend checkpoints" (probably two years
>> back
>>> now) drew a major inspiration from Chandy Lamport, as did actually the
>>> original checkpointing algorithm.
>>>
>>> "Logging data between first and last barrier" versus "barrier jumping
>> over
>>> buffer and storing those buffers" is pretty close same.
>>> However, there are a few nice benefits of the proposal of unaligned
>>> checkpoints over Chandy-Lamport.
>>>
>>> *## Benefits of Unaligned Checkpoints*
>>>
>>> (1) It is very similar to the original algorithm (can be seen an an
>>> optional feature purely in the network stack) and thus can share lot's of
>>> code paths.
>>>
>>> (2) Less data stored. If we make the "jump over buffers" part timeout
>> based
>>> (for example barrier overtakes buffers if not flushed within 10ms) then
>>> checkpoints are in the common case of flowing pipelines aligned without
>>> in-flight data. Only back pressured cases store some in-flight data,
>> which
>>> means we don't regress in the common case and only fix the back pressure
>>> case.
>>>
>>> (3) Faster checkpoints. Chandy Lamport still waits for all barriers to
>>> arrive naturally, logging on the way. If data processing is slow, this
>> can
>>> still take quite a while.
>>>
>>> ==> I think both these points are strong reasons to not change the
>>> mechanism away from "trigger sources" and start with CL-style "trigger
>> all".
>>>
>>>
>>> *## Possible ways to combine Chandy Lamport and Unaligned Checkpoints*
>>>
>>> We can think about something like "take state snapshot on first barrier"
>>> and then store buffers until the other barriers arrive. Inside the
>> network
>>> stack, barriers could still overtake and persist buffers.
>>> The benefit would be less latency increase in the channels which already
>>> have received barriers.
>>> However, as mentioned before, not prioritizing the inputs from which
>>> barriers are still missing can also have an adverse effect.
>>>
>>>
>>> *## Concerning upgrades*
>>>
>>> I think it is a fair restriction to say that upgrades need to happen on
>>> aligned checkpoints. It is a rare enough operation.
>>>
>>>
>>> *## Concerning re-scaling (changing parallelism)*
>>>
>>> We need to support that on unaligned checkpoints as well. There are
>> several
>>> feature proposals about automatic scaling, especially down scaling in
>> case
>>> of missing resources. The last snapshot might be a regular checkpoint, so
>>> all checkpoints need to support rescaling.
>>>
>>>
>>> *## Concerning end-to-end checkpoint duration and "trigger sources"
>> versus
>>> "trigger all"*
>>>
>>> I think for the end-to-end checkpoint duration, an "overtake buffers"
>>> approach yields faster checkpoints, as mentioned above (Chandy Lamport
>>> logging still needs to wait for barrier to flow).
>>>
>>> I don't see the benefit of a "trigger all tasks via RPC concurrently"
>>> approach. Bear in mind that it is still a globally coordinated approach
>> and
>>> you need to wait for the global checkpoint to complete before committing
>>> any side effects.
>>> I believe that the checkpoint time is more determined by the state
>>> checkpoint writing, and the global coordination and metadata commit, than
>>> by the difference in alignment time between "trigger from source and jump
>>> over buffers" versus "trigger all tasks concurrently".
>>>
>>> Trying to optimize a few tens of milliseconds out of the network stack
>>> sends (and changing the overall checkpointing approach completely for
>> that)
>>> while staying with a globally coordinated checkpoint will send us down a
>>> path to a dead end.
>>>
>>> To really bring task persistence latency down to 10s of milliseconds (so
>> we
>>> can frequently commit in sinks), we need to take an approach without any
>>> global coordination. Tasks need to establish a persistent recovery point
>>> individually and at their own discretion, only then can it be frequent
>>> enough. To get there, they would need to decouple themselves from the
>>> predecessor and successor tasks (via something like persistent channels).
>>> This is a different discussion, though, somewhat orthogonal to this one
>>> here.
>>>
>>> Best,
>>> Stephan
>>>
>>>
>>> On Wed, Aug 14, 2019 at 12:37 PM Piotr Nowojski <[hidden email]>
>> wrote:
>>>
>>>> Hi again,
>>>>
>>>> Zhu Zhu let me think about this more. Maybe as Paris is writing, we do
>> not
>>>> need to block any channels at all, at least assuming credit base flow
>>>> control. Regarding what should happen with the following checkpoint is
>>>> another question. Also, should we support concurrent checkpoints and
>>>> subsuming checkpoints as we do now? Maybe not…
>>>>
>>>> Paris
>>>>
>>>> Re
>>>> I. 2. a) and b) - yes, this would have to be taken into an account
>>>> I. 2. c) and IV. 2. - without those, end to end checkpoint time will
>>>> probably be longer than it could be. It might affect external systems.
>> For
>>>> example Kafka, which automatically time outs lingering transactions, and
>>>> for us, the transaction time is equal to the time between two
>> checkpoints.
>>>>
>>>> II 1. - I’m confused. To make things straight. Flink is currently
>>>> snapshotting once it receives all of the checkpoint barriers from all of
>>>> the input channels and only then it broadcasts the checkpoint barrier
>> down
>>>> the stream. And this is correct from exactly-once perspective.
>>>>
>>>> As far as I understand, your proposal based on Chandy Lamport algorithm,
>>>> is snapshotting the state of the operator on the first checkpoint
>> barrier,
>>>> which also looks correct to me.
>>>>
>>>> III. 1. As I responded to Zhu Zhu, let me think a bit more about this.
>>>>
>>>> V. Yes, we still need aligned checkpoints, as they are easier for state
>>>> migration and upgrades.
>>>>
>>>> Piotrek
>>>>
>>>>> On 14 Aug 2019, at 11:22, Paris Carbone <[hidden email]>
>> wrote:
>>>>>
>>>>> Now I see a little more clearly what you have in mind. Thanks for the
>>>> explanation!
>>>>> There are a few intermixed concepts here, some how to do with
>>>> correctness some with performance.
>>>>> Before delving deeper I will just enumerate a few things to make myself
>>>> a little more helpful if I can.
>>>>>
>>>>> I. Initiation
>>>>> -------------
>>>>>
>>>>> 1. RPC to sources only is a less intrusive way to initiate snapshots
>>>> since you utilize better pipeline parallelism (only a small subset of
>> tasks
>>>> is running progressively the protocol at a time, if snapshotting is
>> async
>>>> the overall overhead might not even be observable).
>>>>>
>>>>> 2. If we really want an RPC to all initiation take notice of the
>>>> following implications:
>>>>>
>>>>> a. (correctness) RPC calls are not guaranteed to arrive in every
>>>> task before a marker from a preceding task.
>>>>>
>>>>> b. (correctness) Either the RPC call OR the first arriving marker
>>>> should initiate the algorithm. Whichever comes first. If you only do it
>> per
>>>> RPC call then you capture a "late" state that includes side effects of
>>>> already logged events.
>>>>>
>>>>> c. (performance) Lots of IO will be invoked at the same time on
>>>> the backend store from all tasks. This might lead to high congestion in
>>>> async snapshots.
>>>>>
>>>>> II. Capturing State First
>>>>> -------------------------
>>>>>
>>>>> 1. (correctness) Capturing state at the last marker sounds incorrect to
>>>> me (state contains side effects of already logged events based on the
>>>> proposed scheme). This results into duplicate processing. No?
>>>>>
>>>>> III. Channel Blocking / "Alignment"
>>>>> -----------------------------------
>>>>>
>>>>> 1. (performance?) What is the added benefit? We dont want a "complete"
>>>> transactional snapshot, async snapshots are purely for failure-recovery.
>>>> Thus, I dont see why this needs to be imposed at the expense of
>>>> performance/throughput. With the proposed scheme the whole dataflow
>> anyway
>>>> enters snapshotting/logging mode so tasks more or less snapshot
>>>> concurrently.
>>>>>
>>>>> IV Marker Bypassing
>>>>> -------------------
>>>>>
>>>>> 1. (correctness) This leads to equivalent in-flight snapshots so with
>>>> some quick thinking correct. I will try to model this later and get
>> back
>>>> to you in case I find something wrong.
>>>>>
>>>>> 2. (performance) It also sounds like a meaningful optimisation! I like
>>>> thinking of this as a push-based snapshot. i.e., the producing task
>> somehow
>>>> triggers forward a consumer/channel to capture its state. By example
>>>> consider T1 -> |marker t1| -> T2.
>>>>>
>>>>> V. Usage of "Async" Snapshots
>>>>> ---------------------
>>>>>
>>>>> 1. Do you see this as a full replacement of "full" aligned
>>>> snapshots/savepoints? In my view async shanpshots will be needed from
>> time
>>>> to time but not as frequently. Yet, it seems like a valid approach
>> solely
>>>> for failure-recovery on the same configuration. Here's why:
>>>>>
>>>>> a. With original snapshotting there is a strong duality between
>>>>> a stream input (offsets) and committed side effects (internal
>>>> states and external commits to transactional sinks). While in the async
>>>> version, there are uncommitted operations (inflight records). Thus, you
>>>> cannot use these snapshots for e.g., submitting sql queries with
>> snapshot
>>>> isolation. Also, the original snapshotting gives a lot of potential for
>>>> flink to make proper transactional commits externally.
>>>>>
>>>>> b. Reconfiguration is very tricky, you probably know that better.
>>>> Inflight channel state is no longer valid in a new configuration (i.e.,
>> new
>>>> dataflow graph, new operators, updated operator logic, different
>> channels,
>>>> different parallelism)
>>>>>
>>>>> 2. Async snapshots can also be potentially useful for monitoring the
>>>> general health of a dataflow since they can be analyzed by the task
>> manager
>>>> about the general performance of a job graph and spot bottlenecks for
>>>> example.
>>>>>
>>>>>> On 14 Aug 2019, at 09:08, Piotr Nowojski <[hidden email]> wrote:
>>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> Thomas:
>>>>>> There are no Jira tickets yet (or maybe there is something very old
>>>> somewhere). First we want to discuss it, next present FLIP and at last
>>>> create tickets :)
>>>>>>
>>>>>>> if I understand correctly, then the proposal is to not block any
>>>>>>> input channel at all, but only log data from the backpressured
>> channel
>>>> (and
>>>>>>> make it part of the snapshot) until the barrier arrives
>>>>>>
>>>>>> I would guess that it would be better to block the reads, unless we
>> can
>>>> already process the records from the blocked channel…
>>>>>>
>>>>>> Paris:
>>>>>>
>>>>>> Thanks for the explanation Paris. I’m starting to understand this more
>>>> and I like the idea of snapshotting the state of an operator before
>>>> receiving all of the checkpoint barriers - this would allow more things
>> to
>>>> happen at the same time instead of sequentially. As Zhijiang has pointed
>>>> out there are some things not considered in your proposal: overtaking
>>>> output buffers, but maybe those things could be incorporated together.
>>>>>>
>>>>>> Another thing is that from the wiki description I understood that the
>>>> initial checkpointing is not initialised by any checkpoint barrier, but
>> by
>>>> an independent call/message from the Observer. I haven’t played with
>> this
>>>> idea a lot, but I had some discussion with Nico and it seems that it
>> might
>>>> work:
>>>>>>
>>>>>> 1. JobManager sends and RPC “start checkpoint” to all tasks
>>>>>> 2. Task (with two input channels l1 and l2) upon receiving RPC from
>> 1.,
>>>> takes a snapshot of it's state and:
>>>>>> a) broadcast checkpoint barrier down the stream to all channels (let’s
>>>> ignore for a moment potential for this barrier to overtake the buffer
>>>> output data)
>>>>>> b) for any input channel for which it hasn’t yet received checkpoint
>>>> barrier, the data are being added to the checkpoint
>>>>>> c) once a channel (for example l1) receives a checkpoint barrier, the
>>>> Task blocks reads from that channel (?)
>>>>>> d) after all remaining channels (l2) receive checkpoint barriers, the
>>>> Task first has to process the buffered data after that it can unblock
>> the
>>>> reads from the channels
>>>>>>
>>>>>> Checkpoint barriers do not cascade/flow through different tasks here.
>>>> Checkpoint barrier emitted from Task1, reaches only the immediate
>>>> downstream Tasks. Thanks to this setup, total checkpointing time is not
>> sum
>>>> of checkpointing times of all Tasks one by one, but more or less max of
>> the
>>>> slowest Tasks. Right?
>>>>>>
>>>>>> Couple of intriguing thoughts are:
>>>>>> 3. checkpoint barriers overtaking the output buffers
>>>>>> 4. can we keep processing some data (in order to not waste CPU cycles)
>>>> after we have taking the snapshot of the Task. I think we could.
>>>>>>
>>>>>> Piotrek
>>>>>>
>>>>>>> On 14 Aug 2019, at 06:00, Thomas Weise <[hidden email]> wrote:
>>>>>>>
>>>>>>> Great discussion! I'm excited that this is already under
>>>> consideration! Are
>>>>>>> there any JIRAs or other traces of discussion to follow?
>>>>>>>
>>>>>>> Paris, if I understand correctly, then the proposal is to not block
>> any
>>>>>>> input channel at all, but only log data from the backpressured
>> channel
>>>> (and
>>>>>>> make it part of the snapshot) until the barrier arrives? This is
>>>>>>> intriguing. But probably there is also a benefit of to not continue
>>>> reading
>>>>>>> I1 since that could speed up retrieval from I2. Also, if the user
>> code
>>>> is
>>>>>>> the cause of backpressure, this would avoid pumping more data into
>> the
>>>>>>> process function.
>>>>>>>
>>>>>>> Thanks,
>>>>>>> Thomas
>>>>>>>
>>>>>>>
>>>>>>> On Tue, Aug 13, 2019 at 8:02 AM zhijiang <[hidden email]
>>>> .invalid>
>>>>>>> wrote:
>>>>>>>
>>>>>>>> Hi Paris,
>>>>>>>>
>>>>>>>> Thanks for the detailed sharing. And I think it is very similar with
>>>> the
>>>>>>>> way of overtaking we proposed before.
>>>>>>>>
>>>>>>>> There are some tiny difference:
>>>>>>>> The way of overtaking might need to snapshot all the input/output
>>>> queues.
>>>>>>>> Chandy Lamport seems only need to snaphost (n-1) input channels
>> after
>>>> the
>>>>>>>> first barrier arrives, which might reduce the state sizea bit. But
>>>> normally
>>>>>>>> there should be less buffers for the first input channel with
>> barrier.
>>>>>>>> The output barrier still follows with regular data stream in Chandy
>>>>>>>> Lamport, the same way as current flink. For overtaking way, we need
>>>> to pay
>>>>>>>> extra efforts to make barrier transport firstly before outque queue
>> on
>>>>>>>> upstream side, and change the way of barrier alignment based on
>>>> receiving
>>>>>>>> instead of current reading on downstream side.
>>>>>>>> In the backpressure caused by data skew, the first barrier in almost
>>>> empty
>>>>>>>> input channel should arrive much eariler than the last heavy load
>>>> input
>>>>>>>> channel, so the Chandy Lamport could benefit well. But for the case
>>>> of all
>>>>>>>> balanced heavy load input channels, I mean the first arrived barrier
>>>> might
>>>>>>>> still take much time, then the overtaking way could still fit well
>> to
>>>> speed
>>>>>>>> up checkpoint.
>>>>>>>> Anyway, your proposed suggestion is helpful on my side, especially
>>>>>>>> considering some implementation details .
>>>>>>>>
>>>>>>>> Best,
>>>>>>>> Zhijiang
>>>>>>>> ------------------------------------------------------------------
>>>>>>>> From:Paris Carbone <[hidden email]>
>>>>>>>> Send Time:2019年8月13日(星期二) 14:03
>>>>>>>> To:dev <[hidden email]>
>>>>>>>> Cc:zhijiang <[hidden email]>
>>>>>>>> Subject:Re: Checkpointing under backpressure
>>>>>>>>
>>>>>>>> yes! It’s quite similar I think. Though mind that the devil is in
>> the
>>>>>>>> details, i.e., the temporal order actions are taken.
>>>>>>>>
>>>>>>>> To clarify, let us say you have a task T with two input channels I1
>>>> and I2.
>>>>>>>> The Chandy Lamport execution flow is the following:
>>>>>>>>
>>>>>>>> 1) T receives barrier from I1 and...
>>>>>>>> 2) ...the following three actions happen atomically
>>>>>>>> I ) T snapshots its state T*
>>>>>>>> II) T forwards marker to its outputs
>>>>>>>> III) T starts logging all events of I2 (only) into a buffer M*
>>>>>>>> - Also notice here that T does NOT block I1 as it does in aligned
>>>>>>>> snapshots -
>>>>>>>> 3) Eventually T receives barrier from I2 and stops recording events.
>>>> Its
>>>>>>>> asynchronously captured snapshot is now complete: {T*,M*}.
>>>>>>>> Upon recovery all messages of M* should be replayed in FIFO order.
>>>>>>>>
>>>>>>>> With this approach alignment does not create a deadlock situation
>>>> since
>>>>>>>> anyway 2.II happens asynchronously and messages can be logged as
>> well
>>>>>>>> asynchronously during the process of the snapshot. If there is
>>>>>>>> back-pressure in a pipeline the cause is most probably not this
>>>> algorithm.
>>>>>>>>
>>>>>>>> Back to your observation, the answer : yes and no. In your network
>>>> model,
>>>>>>>> I can see the logic of “logging” and “committing” a final snapshot
>>>> being
>>>>>>>> provided by the channel implementation. However, do mind that the
>>>> first
>>>>>>>> barrier always needs to go “all the way” to initiate the Chandy
>>>> Lamport
>>>>>>>> algorithm logic.
>>>>>>>>
>>>>>>>> The above flow has been proven using temporal logic in my phd thesis
>>>> in
>>>>>>>> case you are interested about the proof.
>>>>>>>> I hope this helps a little clarifying things. Let me know if there
>> is
>>>> any
>>>>>>>> confusing point to disambiguate. I would be more than happy to help
>>>> if I
>>>>>>>> can.
>>>>>>>>
>>>>>>>> Paris
>>>>>>>>
>>>>>>>>> On 13 Aug 2019, at 13:28, Piotr Nowojski <[hidden email]>
>>>> wrote:
>>>>>>>>>
>>>>>>>>> Thanks for the input. Regarding the Chandy-Lamport snapshots don’t
>>>> you
>>>>>>>> still have to wait for the “checkpoint barrier” to arrive in order
>> to
>>>> know
>>>>>>>> when have you already received all possible messages from the
>> upstream
>>>>>>>> tasks/operators? So instead of processing the “in flight” messages
>>>> (as the
>>>>>>>> Flink is doing currently), you are sending them to an “observer”?
>>>>>>>>>
>>>>>>>>> In that case, that’s sounds similar to “checkpoint barriers
>>>> overtaking
>>>>>>>> in flight records” (aka unaligned checkpoints). Just for us, the
>>>> observer
>>>>>>>> is a snapshot state.
>>>>>>>>>
>>>>>>>>> Piotrek
>>>>>>>>>
>>>>>>>>>> On 13 Aug 2019, at 13:14, Paris Carbone <[hidden email]>
>>>>>>>> wrote:
>>>>>>>>>>
>>>>>>>>>> Interesting problem! Thanks for bringing it up Thomas.
>>>>>>>>>>
>>>>>>>>>> Ignore/Correct me if I am wrong but I believe Chandy-Lamport
>>>> snapshots
>>>>>>>> [1] would help out solve this problem more elegantly without
>>>> sacrificing
>>>>>>>> correctness.
>>>>>>>>>> - They do not need alignment, only (async) logging for in-flight
>>>>>>>> records between the time the first barrier is processed until the
>> last
>>>>>>>> barrier arrives in a task.
>>>>>>>>>> - They work fine for failure recovery as long as logged records
>> are
>>>>>>>> replayed on startup.
>>>>>>>>>>
>>>>>>>>>> Flink’s “alligned” savepoints would probably be still necessary
>> for
>>>>>>>> transactional sink commits + any sort of reconfiguration (e.g.,
>>>> rescaling,
>>>>>>>> updating the logic of operators to evolve an application etc.).
>>>>>>>>>>
>>>>>>>>>> I don’t completely understand the “overtaking” approach but if you
>>>> have
>>>>>>>> a concrete definition I would be happy to check it out and help if I
>>>> can!
>>>>>>>>>> Mind that Chandy-Lamport essentially does this by logging things
>> in
>>>>>>>> pending channels in a task snapshot before the barrier arrives.
>>>>>>>>>>
>>>>>>>>>> -Paris
>>>>>>>>>>
>>>>>>>>>> [1]
>> https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm
>>>> <
>>>>>>>> https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm>
>>>>>>>>>>
>>>>>>>>>>> On 13 Aug 2019, at 10:27, Piotr Nowojski <[hidden email]>
>>>> wrote:
>>>>>>>>>>>
>>>>>>>>>>> Hi Thomas,
>>>>>>>>>>>
>>>>>>>>>>> As Zhijiang has responded, we are now in the process of
>> discussing
>>>> how
>>>>>>>> to address this issue and one of the solution that we are discussing
>>>> is
>>>>>>>> exactly what you are proposing: checkpoint barriers overtaking the
>> in
>>>>>>>> flight data and make the in flight data part of the checkpoint.
>>>>>>>>>>>
>>>>>>>>>>> If everything works well, we will be able to present result of
>> our
>>>>>>>> discussions on the dev mailing list soon.
>>>>>>>>>>>
>>>>>>>>>>> Piotrek
>>>>>>>>>>>
>>>>>>>>>>>> On 12 Aug 2019, at 23:23, zhijiang <[hidden email]
>>>> .INVALID>
>>>>>>>> wrote:
>>>>>>>>>>>>
>>>>>>>>>>>> Hi Thomas,
>>>>>>>>>>>>
>>>>>>>>>>>> Thanks for proposing this concern. The barrier alignment takes
>>>> long
>>>>>>>> time in backpressure case which could cause several problems:
>>>>>>>>>>>> 1. Checkpoint timeout as you mentioned.
>>>>>>>>>>>> 2. The recovery cost is high once failover, because much data
>>>> needs
>>>>>>>> to be replayed.
>>>>>>>>>>>> 3. The delay for commit-based sink is high in exactly-once.
>>>>>>>>>>>>
>>>>>>>>>>>> For credit-based flow control from release-1.5, the amount of
>>>>>>>> in-flighting buffers before barrier alignment is reduced, so we
>> could
>>>> get a
>>>>>>>> bit
>>>>>>>>>>>> benefits from speeding checkpoint aspect.
>>>>>>>>>>>>
>>>>>>>>>>>> In release-1.8, I guess we did not suspend the channels which
>>>> already
>>>>>>>> received the barrier in practice. But actually we ever did the
>>>> similar thing
>>>>>>>>>>>> to speed barrier alighment before. I am not quite sure that
>>>>>>>> release-1.8 covers this feature. There were some relevant
>> discussions
>>>> under
>>>>>>>> jira [1].
>>>>>>>>>>>>
>>>>>>>>>>>> For release-1.10, the community is now discussing the feature of
>>>>>>>> unaligned checkpoint which is mainly for resolving above concerns.
>> The
>>>>>>>> basic idea
>>>>>>>>>>>> is to make barrier overtakes the output/input buffer queue to
>>>> speed
>>>>>>>> alignment, and snapshot the input/output buffers as part of
>> checkpoint
>>>>>>>> state. The
>>>>>>>>>>>> details have not confirmed yet and is still under discussion.
>>>> Wish we
>>>>>>>> could make some improvments for the release-1.10.
>>>>>>>>>>>>
>>>>>>>>>>>> [1] https://issues.apache.org/jira/browse/FLINK-8523
>>>>>>>>>>>>
>>>>>>>>>>>> Best,
>>>>>>>>>>>> Zhijiang
>>>>>>>>>>>>
>> ------------------------------------------------------------------
>>>>>>>>>>>> From:Thomas Weise <[hidden email]>
>>>>>>>>>>>> Send Time:2019年8月12日(星期一) 21:38
>>>>>>>>>>>> To:dev <[hidden email]>
>>>>>>>>>>>> Subject:Checkpointing under backpressure
>>>>>>>>>>>>
>>>>>>>>>>>> Hi,
>>>>>>>>>>>>
>>>>>>>>>>>> One of the major operational difficulties we observe with Flink
>>>> are
>>>>>>>>>>>> checkpoint timeouts under backpressure. I'm looking for both
>>>>>>>> confirmation
>>>>>>>>>>>> of my understanding of the current behavior as well as pointers
>>>> for
>>>>>>>> future
>>>>>>>>>>>> improvement work:
>>>>>>>>>>>>
>>>>>>>>>>>> Prior to introduction of credit based flow control in the
>> network
>>>>>>>> stack [1]
>>>>>>>>>>>> [2], checkpoint barriers would back up with the data for all
>>>> logical
>>>>>>>>>>>> channels due to TCP backpressure. Since Flink 1.5, the buffers
>> are
>>>>>>>>>>>> controlled per channel, and checkpoint barriers are only held
>>>> back for
>>>>>>>>>>>> channels that have backpressure, while others can continue
>>>> processing
>>>>>>>>>>>> normally. However, checkpoint barriers still cannot "overtake
>>>> data",
>>>>>>>>>>>> therefore checkpoint alignment remains affected for the channel
>>>> with
>>>>>>>>>>>> backpressure, with the potential for slow checkpointing and
>>>> timeouts.
>>>>>>>>>>>> Albeit the delay of barriers would be capped by the maximum
>>>> in-transit
>>>>>>>>>>>> buffers per channel, resulting in an improvement compared to
>>>> previous
>>>>>>>>>>>> versions of Flink. Also, the backpressure based checkpoint
>>>> alignment
>>>>>>>> can
>>>>>>>>>>>> help the barrier advance faster on the receiver side (by
>>>> suspending
>>>>>>>>>>>> channels that have already delivered the barrier). Is that
>>>> accurate
>>>>>>>> as of
>>>>>>>>>>>> Flink 1.8?
>>>>>>>>>>>>
>>>>>>>>>>>> What appears to be missing to completely unblock checkpointing
>> is
>>>> a
>>>>>>>>>>>> mechanism for checkpoints to overtake the data. That would help
>> in
>>>>>>>>>>>> situations where the processing itself is the bottleneck and
>>>>>>>> prioritization
>>>>>>>>>>>> in the network stack alone cannot address the barrier delay. Was
>>>>>>>> there any
>>>>>>>>>>>> related discussion? One possible solution would be to drain
>>>> incoming
>>>>>>>> data
>>>>>>>>>>>> till the barrier and make it part of the checkpoint instead of
>>>>>>>> processing
>>>>>>>>>>>> it. This is somewhat related to asynchronous processing, but I'm
>>>>>>>> thinking
>>>>>>>>>>>> more of a solution that is automated in the Flink runtime for
>> the
>>>>>>>>>>>> backpressure scenario only.
>>>>>>>>>>>>
>>>>>>>>>>>> Thanks,
>>>>>>>>>>>> Thomas
>>>>>>>>>>>>
>>>>>>>>>>>> [1]
>> https://flink.apache.org/2019/06/05/flink-network-stack.html
>>>>>>>>>>>> [2]
>>>>>>>>>>>>
>>>>>>>>
>>>>
>> https://docs.google.com/document/d/1chTOuOqe0sBsjldA_r-wXYeSIhU2zRGpUaTaik7QZ84/edit#heading=h.pjh6mv7m2hjn
>>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>
>>>>>
>>>>
>>>>
>>
>>

Piotr Nowojski-3

Re: Checkpointing under backpressure

> Thanks for the great ideas so far.

+1

Regarding other things raised, I mostly agree with Stephan.

I like the idea of simultaneously starting the checkpoint everywhere via RPC call (especially in cases where Tasks are busy doing some synchronous operations for example for tens of milliseconds. In that case every network exchange adds tens of milliseconds of delay in propagating the checkpoint). However I agree that this might be a premature optimisation assuming the current state of our code (we already have checkpoint barriers).

However I like the idea of switching from:

1. A -> S -> F (Align -> snapshot -> forward markers)

To

2. S -> F -> L (Snapshot -> forward markers -> log pending channels)

Or even to

6. F -> S -> L (Forward markers -> snapshot -> log pending channels)

It feels to me like this would decouple propagation of checkpoints from costs of synchronous snapshots and waiting for all of the checkpoint barriers to arrive (even if they will overtake in-flight records, this might take some time).

> What I like about the Chandy Lamport approach (2.) initiated from sources is that:
> - Snapshotting imposes no modification to normal processing.

Yes, I agree that would be nice. Currently, during the alignment and blocking of the input channels, we might be wasting CPU cycles of up stream tasks. If we succeed in designing new checkpointing mechanism to not disrupt/block regular data processing (% the extra IO cost for logging the in-flight records), that would be a huge improvement.

Piotrek

> On 14 Aug 2019, at 14:56, Paris Carbone <[hidden email]> wrote:
>
> Sure I see. In cases when no periodic aligned snapshots are employed this is the only option.
>
> Two things that were not highlighted enough so far on the proposed protocol (included my mails):
> - The Recovery/Reconfiguration strategy should strictly prioritise processing logged events before entering normal task input operation. Otherwise causality can be violated. This also means dataflow recovery will be expected to be slower to the one employed on an aligned snapshot.
> - Same as with state capture, markers should be forwarded upon first marker received on input. No later than that. Otherwise we have duplicate side effects.
>
> Thanks for the great ideas so far.
>
> Paris
>
>> On 14 Aug 2019, at 14:33, Stephan Ewen <[hidden email]> wrote:
>>
>> Scaling with unaligned checkpoints might be a necessity.
>>
>> Let's assume the job failed due to a lost TaskManager, but no new
>> TaskManager becomes available.
>> In that case we need to scale down based on the latest complete checkpoint,
>> because we cannot produce a new checkpoint.
>>
>>
>> On Wed, Aug 14, 2019 at 2:05 PM Paris Carbone <[hidden email]>
>> wrote:
>>
>>> +1 I think we are on the same page Stephan.
>>>
>>> Rescaling on unaligned checkpoint sounds challenging and a bit
>>> unnecessary. No?
>>> Why not sticking to aligned snapshots for live reconfiguration/rescaling?
>>> It’s a pretty rare operation and it would simplify things by a lot.
>>> Everything can be “staged” upon alignment including replacing channels and
>>> tasks.
>>>
>>> -Paris
>>>
>>>> On 14 Aug 2019, at 13:39, Stephan Ewen <[hidden email]> wrote:
>>>>
>>>> Hi all!
>>>>
>>>> Yes, the first proposal of "unaligend checkpoints" (probably two years
>>> back
>>>> now) drew a major inspiration from Chandy Lamport, as did actually the
>>>> original checkpointing algorithm.
>>>>
>>>> "Logging data between first and last barrier" versus "barrier jumping
>>> over
>>>> buffer and storing those buffers" is pretty close same.
>>>> However, there are a few nice benefits of the proposal of unaligned
>>>> checkpoints over Chandy-Lamport.
>>>>
>>>> *## Benefits of Unaligned Checkpoints*
>>>>
>>>> (1) It is very similar to the original algorithm (can be seen an an
>>>> optional feature purely in the network stack) and thus can share lot's of
>>>> code paths.
>>>>
>>>> (2) Less data stored. If we make the "jump over buffers" part timeout
>>> based
>>>> (for example barrier overtakes buffers if not flushed within 10ms) then
>>>> checkpoints are in the common case of flowing pipelines aligned without
>>>> in-flight data. Only back pressured cases store some in-flight data,
>>> which
>>>> means we don't regress in the common case and only fix the back pressure
>>>> case.
>>>>
>>>> (3) Faster checkpoints. Chandy Lamport still waits for all barriers to
>>>> arrive naturally, logging on the way. If data processing is slow, this
>>> can
>>>> still take quite a while.
>>>>
>>>> ==> I think both these points are strong reasons to not change the
>>>> mechanism away from "trigger sources" and start with CL-style "trigger
>>> all".
>>>>
>>>>
>>>> *## Possible ways to combine Chandy Lamport and Unaligned Checkpoints*
>>>>
>>>> We can think about something like "take state snapshot on first barrier"
>>>> and then store buffers until the other barriers arrive. Inside the
>>> network
>>>> stack, barriers could still overtake and persist buffers.
>>>> The benefit would be less latency increase in the channels which already
>>>> have received barriers.
>>>> However, as mentioned before, not prioritizing the inputs from which
>>>> barriers are still missing can also have an adverse effect.
>>>>
>>>>
>>>> *## Concerning upgrades*
>>>>
>>>> I think it is a fair restriction to say that upgrades need to happen on
>>>> aligned checkpoints. It is a rare enough operation.
>>>>
>>>>
>>>> *## Concerning re-scaling (changing parallelism)*
>>>>
>>>> We need to support that on unaligned checkpoints as well. There are
>>> several
>>>> feature proposals about automatic scaling, especially down scaling in
>>> case
>>>> of missing resources. The last snapshot might be a regular checkpoint, so
>>>> all checkpoints need to support rescaling.
>>>>
>>>>
>>>> *## Concerning end-to-end checkpoint duration and "trigger sources"
>>> versus
>>>> "trigger all"*
>>>>
>>>> I think for the end-to-end checkpoint duration, an "overtake buffers"
>>>> approach yields faster checkpoints, as mentioned above (Chandy Lamport
>>>> logging still needs to wait for barrier to flow).
>>>>
>>>> I don't see the benefit of a "trigger all tasks via RPC concurrently"
>>>> approach. Bear in mind that it is still a globally coordinated approach
>>> and
>>>> you need to wait for the global checkpoint to complete before committing
>>>> any side effects.
>>>> I believe that the checkpoint time is more determined by the state
>>>> checkpoint writing, and the global coordination and metadata commit, than
>>>> by the difference in alignment time between "trigger from source and jump
>>>> over buffers" versus "trigger all tasks concurrently".
>>>>
>>>> Trying to optimize a few tens of milliseconds out of the network stack
>>>> sends (and changing the overall checkpointing approach completely for
>>> that)
>>>> while staying with a globally coordinated checkpoint will send us down a
>>>> path to a dead end.
>>>>
>>>> To really bring task persistence latency down to 10s of milliseconds (so
>>> we
>>>> can frequently commit in sinks), we need to take an approach without any
>>>> global coordination. Tasks need to establish a persistent recovery point
>>>> individually and at their own discretion, only then can it be frequent
>>>> enough. To get there, they would need to decouple themselves from the
>>>> predecessor and successor tasks (via something like persistent channels).
>>>> This is a different discussion, though, somewhat orthogonal to this one
>>>> here.
>>>>
>>>> Best,
>>>> Stephan
>>>>
>>>>
>>>> On Wed, Aug 14, 2019 at 12:37 PM Piotr Nowojski <[hidden email]>
>>> wrote:
>>>>
>>>>> Hi again,
>>>>>
>>>>> Zhu Zhu let me think about this more. Maybe as Paris is writing, we do
>>> not
>>>>> need to block any channels at all, at least assuming credit base flow
>>>>> control. Regarding what should happen with the following checkpoint is
>>>>> another question. Also, should we support concurrent checkpoints and
>>>>> subsuming checkpoints as we do now? Maybe not…
>>>>>
>>>>> Paris
>>>>>
>>>>> Re
>>>>> I. 2. a) and b) - yes, this would have to be taken into an account
>>>>> I. 2. c) and IV. 2. - without those, end to end checkpoint time will
>>>>> probably be longer than it could be. It might affect external systems.
>>> For
>>>>> example Kafka, which automatically time outs lingering transactions, and
>>>>> for us, the transaction time is equal to the time between two
>>> checkpoints.
>>>>>
>>>>> II 1. - I’m confused. To make things straight. Flink is currently
>>>>> snapshotting once it receives all of the checkpoint barriers from all of
>>>>> the input channels and only then it broadcasts the checkpoint barrier
>>> down
>>>>> the stream. And this is correct from exactly-once perspective.
>>>>>
>>>>> As far as I understand, your proposal based on Chandy Lamport algorithm,
>>>>> is snapshotting the state of the operator on the first checkpoint
>>> barrier,
>>>>> which also looks correct to me.
>>>>>
>>>>> III. 1. As I responded to Zhu Zhu, let me think a bit more about this.
>>>>>
>>>>> V. Yes, we still need aligned checkpoints, as they are easier for state
>>>>> migration and upgrades.
>>>>>
>>>>> Piotrek
>>>>>
>>>>>> On 14 Aug 2019, at 11:22, Paris Carbone <[hidden email]>
>>> wrote:
>>>>>>
>>>>>> Now I see a little more clearly what you have in mind. Thanks for the
>>>>> explanation!
>>>>>> There are a few intermixed concepts here, some how to do with
>>>>> correctness some with performance.
>>>>>> Before delving deeper I will just enumerate a few things to make myself
>>>>> a little more helpful if I can.
>>>>>>
>>>>>> I. Initiation
>>>>>> -------------
>>>>>>
>>>>>> 1. RPC to sources only is a less intrusive way to initiate snapshots
>>>>> since you utilize better pipeline parallelism (only a small subset of
>>> tasks
>>>>> is running progressively the protocol at a time, if snapshotting is
>>> async
>>>>> the overall overhead might not even be observable).
>>>>>>
>>>>>> 2. If we really want an RPC to all initiation take notice of the
>>>>> following implications:
>>>>>>
>>>>>> a. (correctness) RPC calls are not guaranteed to arrive in every
>>>>> task before a marker from a preceding task.
>>>>>>
>>>>>> b. (correctness) Either the RPC call OR the first arriving marker
>>>>> should initiate the algorithm. Whichever comes first. If you only do it
>>> per
>>>>> RPC call then you capture a "late" state that includes side effects of
>>>>> already logged events.
>>>>>>
>>>>>> c. (performance) Lots of IO will be invoked at the same time on
>>>>> the backend store from all tasks. This might lead to high congestion in
>>>>> async snapshots.
>>>>>>
>>>>>> II. Capturing State First
>>>>>> -------------------------
>>>>>>
>>>>>> 1. (correctness) Capturing state at the last marker sounds incorrect to
>>>>> me (state contains side effects of already logged events based on the
>>>>> proposed scheme). This results into duplicate processing. No?
>>>>>>
>>>>>> III. Channel Blocking / "Alignment"
>>>>>> -----------------------------------
>>>>>>
>>>>>> 1. (performance?) What is the added benefit? We dont want a "complete"
>>>>> transactional snapshot, async snapshots are purely for failure-recovery.
>>>>> Thus, I dont see why this needs to be imposed at the expense of
>>>>> performance/throughput. With the proposed scheme the whole dataflow
>>> anyway
>>>>> enters snapshotting/logging mode so tasks more or less snapshot
>>>>> concurrently.
>>>>>>
>>>>>> IV Marker Bypassing
>>>>>> -------------------
>>>>>>
>>>>>> 1. (correctness) This leads to equivalent in-flight snapshots so with
>>>>> some quick thinking correct. I will try to model this later and get
>>> back
>>>>> to you in case I find something wrong.
>>>>>>
>>>>>> 2. (performance) It also sounds like a meaningful optimisation! I like
>>>>> thinking of this as a push-based snapshot. i.e., the producing task
>>> somehow
>>>>> triggers forward a consumer/channel to capture its state. By example
>>>>> consider T1 -> |marker t1| -> T2.
>>>>>>
>>>>>> V. Usage of "Async" Snapshots
>>>>>> ---------------------
>>>>>>
>>>>>> 1. Do you see this as a full replacement of "full" aligned
>>>>> snapshots/savepoints? In my view async shanpshots will be needed from
>>> time
>>>>> to time but not as frequently. Yet, it seems like a valid approach
>>> solely
>>>>> for failure-recovery on the same configuration. Here's why:
>>>>>>
>>>>>> a. With original snapshotting there is a strong duality between
>>>>>> a stream input (offsets) and committed side effects (internal
>>>>> states and external commits to transactional sinks). While in the async
>>>>> version, there are uncommitted operations (inflight records). Thus, you
>>>>> cannot use these snapshots for e.g., submitting sql queries with
>>> snapshot
>>>>> isolation. Also, the original snapshotting gives a lot of potential for
>>>>> flink to make proper transactional commits externally.
>>>>>>
>>>>>> b. Reconfiguration is very tricky, you probably know that better.
>>>>> Inflight channel state is no longer valid in a new configuration (i.e.,
>>> new
>>>>> dataflow graph, new operators, updated operator logic, different
>>> channels,
>>>>> different parallelism)
>>>>>>
>>>>>> 2. Async snapshots can also be potentially useful for monitoring the
>>>>> general health of a dataflow since they can be analyzed by the task
>>> manager
>>>>> about the general performance of a job graph and spot bottlenecks for
>>>>> example.
>>>>>>
>>>>>>> On 14 Aug 2019, at 09:08, Piotr Nowojski <[hidden email]> wrote:
>>>>>>>
>>>>>>> Hi,
>>>>>>>
>>>>>>> Thomas:
>>>>>>> There are no Jira tickets yet (or maybe there is something very old
>>>>> somewhere). First we want to discuss it, next present FLIP and at last
>>>>> create tickets :)
>>>>>>>
>>>>>>>> if I understand correctly, then the proposal is to not block any
>>>>>>>> input channel at all, but only log data from the backpressured
>>> channel
>>>>> (and
>>>>>>>> make it part of the snapshot) until the barrier arrives
>>>>>>>
>>>>>>> I would guess that it would be better to block the reads, unless we
>>> can
>>>>> already process the records from the blocked channel…
>>>>>>>
>>>>>>> Paris:
>>>>>>>
>>>>>>> Thanks for the explanation Paris. I’m starting to understand this more
>>>>> and I like the idea of snapshotting the state of an operator before
>>>>> receiving all of the checkpoint barriers - this would allow more things
>>> to
>>>>> happen at the same time instead of sequentially. As Zhijiang has pointed
>>>>> out there are some things not considered in your proposal: overtaking
>>>>> output buffers, but maybe those things could be incorporated together.
>>>>>>>
>>>>>>> Another thing is that from the wiki description I understood that the
>>>>> initial checkpointing is not initialised by any checkpoint barrier, but
>>> by
>>>>> an independent call/message from the Observer. I haven’t played with
>>> this
>>>>> idea a lot, but I had some discussion with Nico and it seems that it
>>> might
>>>>> work:
>>>>>>>
>>>>>>> 1. JobManager sends and RPC “start checkpoint” to all tasks
>>>>>>> 2. Task (with two input channels l1 and l2) upon receiving RPC from
>>> 1.,
>>>>> takes a snapshot of it's state and:
>>>>>>> a) broadcast checkpoint barrier down the stream to all channels (let’s
>>>>> ignore for a moment potential for this barrier to overtake the buffer
>>>>> output data)
>>>>>>> b) for any input channel for which it hasn’t yet received checkpoint
>>>>> barrier, the data are being added to the checkpoint
>>>>>>> c) once a channel (for example l1) receives a checkpoint barrier, the
>>>>> Task blocks reads from that channel (?)
>>>>>>> d) after all remaining channels (l2) receive checkpoint barriers, the
>>>>> Task first has to process the buffered data after that it can unblock
>>> the
>>>>> reads from the channels
>>>>>>>
>>>>>>> Checkpoint barriers do not cascade/flow through different tasks here.
>>>>> Checkpoint barrier emitted from Task1, reaches only the immediate
>>>>> downstream Tasks. Thanks to this setup, total checkpointing time is not
>>> sum
>>>>> of checkpointing times of all Tasks one by one, but more or less max of
>>> the
>>>>> slowest Tasks. Right?
>>>>>>>
>>>>>>> Couple of intriguing thoughts are:
>>>>>>> 3. checkpoint barriers overtaking the output buffers
>>>>>>> 4. can we keep processing some data (in order to not waste CPU cycles)
>>>>> after we have taking the snapshot of the Task. I think we could.
>>>>>>>
>>>>>>> Piotrek
>>>>>>>
>>>>>>>> On 14 Aug 2019, at 06:00, Thomas Weise <[hidden email]> wrote:
>>>>>>>>
>>>>>>>> Great discussion! I'm excited that this is already under
>>>>> consideration! Are
>>>>>>>> there any JIRAs or other traces of discussion to follow?
>>>>>>>>
>>>>>>>> Paris, if I understand correctly, then the proposal is to not block
>>> any
>>>>>>>> input channel at all, but only log data from the backpressured
>>> channel
>>>>> (and
>>>>>>>> make it part of the snapshot) until the barrier arrives? This is
>>>>>>>> intriguing. But probably there is also a benefit of to not continue
>>>>> reading
>>>>>>>> I1 since that could speed up retrieval from I2. Also, if the user
>>> code
>>>>> is
>>>>>>>> the cause of backpressure, this would avoid pumping more data into
>>> the
>>>>>>>> process function.
>>>>>>>>
>>>>>>>> Thanks,
>>>>>>>> Thomas
>>>>>>>>
>>>>>>>>
>>>>>>>> On Tue, Aug 13, 2019 at 8:02 AM zhijiang <[hidden email]
>>>>> .invalid>
>>>>>>>> wrote:
>>>>>>>>
>>>>>>>>> Hi Paris,
>>>>>>>>>
>>>>>>>>> Thanks for the detailed sharing. And I think it is very similar with
>>>>> the
>>>>>>>>> way of overtaking we proposed before.
>>>>>>>>>
>>>>>>>>> There are some tiny difference:
>>>>>>>>> The way of overtaking might need to snapshot all the input/output
>>>>> queues.
>>>>>>>>> Chandy Lamport seems only need to snaphost (n-1) input channels
>>> after
>>>>> the
>>>>>>>>> first barrier arrives, which might reduce the state sizea bit. But
>>>>> normally
>>>>>>>>> there should be less buffers for the first input channel with
>>> barrier.
>>>>>>>>> The output barrier still follows with regular data stream in Chandy
>>>>>>>>> Lamport, the same way as current flink. For overtaking way, we need
>>>>> to pay
>>>>>>>>> extra efforts to make barrier transport firstly before outque queue
>>> on
>>>>>>>>> upstream side, and change the way of barrier alignment based on
>>>>> receiving
>>>>>>>>> instead of current reading on downstream side.
>>>>>>>>> In the backpressure caused by data skew, the first barrier in almost
>>>>> empty
>>>>>>>>> input channel should arrive much eariler than the last heavy load
>>>>> input
>>>>>>>>> channel, so the Chandy Lamport could benefit well. But for the case
>>>>> of all
>>>>>>>>> balanced heavy load input channels, I mean the first arrived barrier
>>>>> might
>>>>>>>>> still take much time, then the overtaking way could still fit well
>>> to
>>>>> speed
>>>>>>>>> up checkpoint.
>>>>>>>>> Anyway, your proposed suggestion is helpful on my side, especially
>>>>>>>>> considering some implementation details .
>>>>>>>>>
>>>>>>>>> Best,
>>>>>>>>> Zhijiang
>>>>>>>>> ------------------------------------------------------------------
>>>>>>>>> From:Paris Carbone <[hidden email]>
>>>>>>>>> Send Time:2019年8月13日(星期二) 14:03
>>>>>>>>> To:dev <[hidden email]>
>>>>>>>>> Cc:zhijiang <[hidden email]>
>>>>>>>>> Subject:Re: Checkpointing under backpressure
>>>>>>>>>
>>>>>>>>> yes! It’s quite similar I think. Though mind that the devil is in
>>> the
>>>>>>>>> details, i.e., the temporal order actions are taken.
>>>>>>>>>
>>>>>>>>> To clarify, let us say you have a task T with two input channels I1
>>>>> and I2.
>>>>>>>>> The Chandy Lamport execution flow is the following:
>>>>>>>>>
>>>>>>>>> 1) T receives barrier from I1 and...
>>>>>>>>> 2) ...the following three actions happen atomically
>>>>>>>>> I ) T snapshots its state T*
>>>>>>>>> II) T forwards marker to its outputs
>>>>>>>>> III) T starts logging all events of I2 (only) into a buffer M*
>>>>>>>>> - Also notice here that T does NOT block I1 as it does in aligned
>>>>>>>>> snapshots -
>>>>>>>>> 3) Eventually T receives barrier from I2 and stops recording events.
>>>>> Its
>>>>>>>>> asynchronously captured snapshot is now complete: {T*,M*}.
>>>>>>>>> Upon recovery all messages of M* should be replayed in FIFO order.
>>>>>>>>>
>>>>>>>>> With this approach alignment does not create a deadlock situation
>>>>> since
>>>>>>>>> anyway 2.II happens asynchronously and messages can be logged as
>>> well
>>>>>>>>> asynchronously during the process of the snapshot. If there is
>>>>>>>>> back-pressure in a pipeline the cause is most probably not this
>>>>> algorithm.
>>>>>>>>>
>>>>>>>>> Back to your observation, the answer : yes and no. In your network
>>>>> model,
>>>>>>>>> I can see the logic of “logging” and “committing” a final snapshot
>>>>> being
>>>>>>>>> provided by the channel implementation. However, do mind that the
>>>>> first
>>>>>>>>> barrier always needs to go “all the way” to initiate the Chandy
>>>>> Lamport
>>>>>>>>> algorithm logic.
>>>>>>>>>
>>>>>>>>> The above flow has been proven using temporal logic in my phd thesis
>>>>> in
>>>>>>>>> case you are interested about the proof.
>>>>>>>>> I hope this helps a little clarifying things. Let me know if there
>>> is
>>>>> any
>>>>>>>>> confusing point to disambiguate. I would be more than happy to help
>>>>> if I
>>>>>>>>> can.
>>>>>>>>>
>>>>>>>>> Paris
>>>>>>>>>
>>>>>>>>>> On 13 Aug 2019, at 13:28, Piotr Nowojski <[hidden email]>
>>>>> wrote:
>>>>>>>>>>
>>>>>>>>>> Thanks for the input. Regarding the Chandy-Lamport snapshots don’t
>>>>> you
>>>>>>>>> still have to wait for the “checkpoint barrier” to arrive in order
>>> to
>>>>> know
>>>>>>>>> when have you already received all possible messages from the
>>> upstream
>>>>>>>>> tasks/operators? So instead of processing the “in flight” messages
>>>>> (as the
>>>>>>>>> Flink is doing currently), you are sending them to an “observer”?
>>>>>>>>>>
>>>>>>>>>> In that case, that’s sounds similar to “checkpoint barriers
>>>>> overtaking
>>>>>>>>> in flight records” (aka unaligned checkpoints). Just for us, the
>>>>> observer
>>>>>>>>> is a snapshot state.
>>>>>>>>>>
>>>>>>>>>> Piotrek
>>>>>>>>>>
>>>>>>>>>>> On 13 Aug 2019, at 13:14, Paris Carbone <[hidden email]>
>>>>>>>>> wrote:
>>>>>>>>>>>
>>>>>>>>>>> Interesting problem! Thanks for bringing it up Thomas.
>>>>>>>>>>>
>>>>>>>>>>> Ignore/Correct me if I am wrong but I believe Chandy-Lamport
>>>>> snapshots
>>>>>>>>> [1] would help out solve this problem more elegantly without
>>>>> sacrificing
>>>>>>>>> correctness.
>>>>>>>>>>> - They do not need alignment, only (async) logging for in-flight
>>>>>>>>> records between the time the first barrier is processed until the
>>> last
>>>>>>>>> barrier arrives in a task.
>>>>>>>>>>> - They work fine for failure recovery as long as logged records
>>> are
>>>>>>>>> replayed on startup.
>>>>>>>>>>>
>>>>>>>>>>> Flink’s “alligned” savepoints would probably be still necessary
>>> for
>>>>>>>>> transactional sink commits + any sort of reconfiguration (e.g.,
>>>>> rescaling,
>>>>>>>>> updating the logic of operators to evolve an application etc.).
>>>>>>>>>>>
>>>>>>>>>>> I don’t completely understand the “overtaking” approach but if you
>>>>> have
>>>>>>>>> a concrete definition I would be happy to check it out and help if I
>>>>> can!
>>>>>>>>>>> Mind that Chandy-Lamport essentially does this by logging things
>>> in
>>>>>>>>> pending channels in a task snapshot before the barrier arrives.
>>>>>>>>>>>
>>>>>>>>>>> -Paris
>>>>>>>>>>>
>>>>>>>>>>> [1]
>>> https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm
>>>>> <
>>>>>>>>> https://en.wikipedia.org/wiki/Chandy%E2%80%93Lamport_algorithm>
>>>>>>>>>>>
>>>>>>>>>>>> On 13 Aug 2019, at 10:27, Piotr Nowojski <[hidden email]>
>>>>> wrote:
>>>>>>>>>>>>
>>>>>>>>>>>> Hi Thomas,
>>>>>>>>>>>>
>>>>>>>>>>>> As Zhijiang has responded, we are now in the process of
>>> discussing
>>>>> how
>>>>>>>>> to address this issue and one of the solution that we are discussing
>>>>> is
>>>>>>>>> exactly what you are proposing: checkpoint barriers overtaking the
>>> in
>>>>>>>>> flight data and make the in flight data part of the checkpoint.
>>>>>>>>>>>>
>>>>>>>>>>>> If everything works well, we will be able to present result of
>>> our
>>>>>>>>> discussions on the dev mailing list soon.
>>>>>>>>>>>>
>>>>>>>>>>>> Piotrek
>>>>>>>>>>>>
>>>>>>>>>>>>> On 12 Aug 2019, at 23:23, zhijiang <[hidden email]
>>>>> .INVALID>
>>>>>>>>> wrote:
>>>>>>>>>>>>>
>>>>>>>>>>>>> Hi Thomas,
>>>>>>>>>>>>>
>>>>>>>>>>>>> Thanks for proposing this concern. The barrier alignment takes
>>>>> long
>>>>>>>>> time in backpressure case which could cause several problems:
>>>>>>>>>>>>> 1. Checkpoint timeout as you mentioned.
>>>>>>>>>>>>> 2. The recovery cost is high once failover, because much data
>>>>> needs
>>>>>>>>> to be replayed.
>>>>>>>>>>>>> 3. The delay for commit-based sink is high in exactly-once.
>>>>>>>>>>>>>
>>>>>>>>>>>>> For credit-based flow control from release-1.5, the amount of
>>>>>>>>> in-flighting buffers before barrier alignment is reduced, so we
>>> could
>>>>> get a
>>>>>>>>> bit
>>>>>>>>>>>>> benefits from speeding checkpoint aspect.
>>>>>>>>>>>>>
>>>>>>>>>>>>> In release-1.8, I guess we did not suspend the channels which
>>>>> already
>>>>>>>>> received the barrier in practice. But actually we ever did the
>>>>> similar thing
>>>>>>>>>>>>> to speed barrier alighment before. I am not quite sure that
>>>>>>>>> release-1.8 covers this feature. There were some relevant
>>> discussions
>>>>> under
>>>>>>>>> jira [1].
>>>>>>>>>>>>>
>>>>>>>>>>>>> For release-1.10, the community is now discussing the feature of
>>>>>>>>> unaligned checkpoint which is mainly for resolving above concerns.
>>> The
>>>>>>>>> basic idea
>>>>>>>>>>>>> is to make barrier overtakes the output/input buffer queue to
>>>>> speed
>>>>>>>>> alignment, and snapshot the input/output buffers as part of
>>> checkpoint
>>>>>>>>> state. The
>>>>>>>>>>>>> details have not confirmed yet and is still under discussion.
>>>>> Wish we
>>>>>>>>> could make some improvments for the release-1.10.
>>>>>>>>>>>>>
>>>>>>>>>>>>> [1] https://issues.apache.org/jira/browse/FLINK-8523
>>>>>>>>>>>>>
>>>>>>>>>>>>> Best,
>>>>>>>>>>>>> Zhijiang
>>>>>>>>>>>>>
>>> ------------------------------------------------------------------
>>>>>>>>>>>>> From:Thomas Weise <[hidden email]>
>>>>>>>>>>>>> Send Time:2019年8月12日(星期一) 21:38
>>>>>>>>>>>>> To:dev <[hidden email]>
>>>>>>>>>>>>> Subject:Checkpointing under backpressure
>>>>>>>>>>>>>
>>>>>>>>>>>>> Hi,
>>>>>>>>>>>>>
>>>>>>>>>>>>> One of the major operational difficulties we observe with Flink
>>>>> are
>>>>>>>>>>>>> checkpoint timeouts under backpressure. I'm looking for both
>>>>>>>>> confirmation
>>>>>>>>>>>>> of my understanding of the current behavior as well as pointers
>>>>> for
>>>>>>>>> future
>>>>>>>>>>>>> improvement work:
>>>>>>>>>>>>>
>>>>>>>>>>>>> Prior to introduction of credit based flow control in the
>>> network
>>>>>>>>> stack [1]
>>>>>>>>>>>>> [2], checkpoint barriers would back up with the data for all
>>>>> logical
>>>>>>>>>>>>> channels due to TCP backpressure. Since Flink 1.5, the buffers
>>> are
>>>>>>>>>>>>> controlled per channel, and checkpoint barriers are only held
>>>>> back for
>>>>>>>>>>>>> channels that have backpressure, while others can continue
>>>>> processing
>>>>>>>>>>>>> normally. However, checkpoint barriers still cannot "overtake
>>>>> data",
>>>>>>>>>>>>> therefore checkpoint alignment remains affected for the channel
>>>>> with
>>>>>>>>>>>>> backpressure, with the potential for slow checkpointing and
>>>>> timeouts.
>>>>>>>>>>>>> Albeit the delay of barriers would be capped by the maximum
>>>>> in-transit
>>>>>>>>>>>>> buffers per channel, resulting in an improvement compared to
>>>>> previous
>>>>>>>>>>>>> versions of Flink. Also, the backpressure based checkpoint
>>>>> alignment
>>>>>>>>> can
>>>>>>>>>>>>> help the barrier advance faster on the receiver side (by
>>>>> suspending
>>>>>>>>>>>>> channels that have already delivered the barrier). Is that
>>>>> accurate
>>>>>>>>> as of
>>>>>>>>>>>>> Flink 1.8?
>>>>>>>>>>>>>
>>>>>>>>>>>>> What appears to be missing to completely unblock checkpointing
>>> is
>>>>> a
>>>>>>>>>>>>> mechanism for checkpoints to overtake the data. That would help
>>> in
>>>>>>>>>>>>> situations where the processing itself is the bottleneck and
>>>>>>>>> prioritization
>>>>>>>>>>>>> in the network stack alone cannot address the barrier delay. Was
>>>>>>>>> there any
>>>>>>>>>>>>> related discussion? One possible solution would be to drain
>>>>> incoming
>>>>>>>>> data
>>>>>>>>>>>>> till the barrier and make it part of the checkpoint instead of
>>>>>>>>> processing
>>>>>>>>>>>>> it. This is somewhat related to asynchronous processing, but I'm
>>>>>>>>> thinking
>>>>>>>>>>>>> more of a solution that is automated in the Flink runtime for
>>> the
>>>>>>>>>>>>> backpressure scenario only.
>>>>>>>>>>>>>
>>>>>>>>>>>>> Thanks,
>>>>>>>>>>>>> Thomas
>>>>>>>>>>>>>
>>>>>>>>>>>>> [1]
>>> https://flink.apache.org/2019/06/05/flink-network-stack.html
>>>>>>>>>>>>> [2]
>>>>>>>>>>>>>
>>>>>>>>>
>>>>>
>>> https://docs.google.com/document/d/1chTOuOqe0sBsjldA_r-wXYeSIhU2zRGpUaTaik7QZ84/edit#heading=h.pjh6mv7m2hjn
>>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>>
>>>
>>>
>