[GitHub] [flink] Myasuka commented on pull request #8693: [FLINK-8871] Support to cancel checkpoing via notification

2020-05-19 Thread GitBox


Myasuka commented on pull request #8693:
URL: https://github.com/apache/flink/pull/8693#issuecomment-630659431


   @pnowojski thanks for your remind.
   I think I have addressed all comments to update the commits.
   
   Since current CI is still pending, my local pipeline based on the same 
commit 
[d023c50](https://github.com/apache/flink/commit/d023c507a423b31f38203581c8eed27d28cb3a3c)
 get a [green 
result](https://myasuka.visualstudio.com/flink/_build/results?buildId=60=results).



This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org




[GitHub] [flink] Myasuka commented on pull request #8693: [FLINK-8871] Support to cancel checkpoing via notification

2020-05-15 Thread GitBox


Myasuka commented on pull request #8693:
URL: https://github.com/apache/flink/pull/8693#issuecomment-629180857


   I think current IT cases do not target to verify the cancel notification, 
I'll create one to target to veriy this.



This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org




[GitHub] [flink] Myasuka commented on pull request #8693: [FLINK-8871] Support to cancel checkpoing via notification

2020-05-14 Thread GitBox


Myasuka commented on pull request #8693:
URL: https://github.com/apache/flink/pull/8693#issuecomment-629017420


   For the 2nd question @romanovacca , after discussed with @zhijiangW offline, 
we believe this PR with cancellation via RPC would not take part in the data 
transmission in channel, thus the upstream task would be unblocked once the 
channel is unblocked which is the same logic as before.



This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org




[GitHub] [flink] Myasuka commented on pull request #8693: [FLINK-8871] Support to cancel checkpoing via notification

2020-05-14 Thread GitBox


Myasuka commented on pull request #8693:
URL: https://github.com/apache/flink/pull/8693#issuecomment-628828516


   @rkhachatryan thanks for your review. It's quite late for me at my time now 
and plan to update at my day time.
   Before go to sleep, I could answer the 1st question: barrier alignment would 
still happen as usual, current PR would not affect barrier align part as 
discussed in the beginning, and we could improve this in the future. 



This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org




[GitHub] [flink] Myasuka commented on pull request #8693: [FLINK-8871] Support to cancel checkpoing via notification

2020-05-14 Thread GitBox


Myasuka commented on pull request #8693:
URL: https://github.com/apache/flink/pull/8693#issuecomment-628754467


   @flinkbot run azure



This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org




[GitHub] [flink] Myasuka commented on pull request #8693: [FLINK-8871] Support to cancel checkpoing via notification

2020-05-11 Thread GitBox


Myasuka commented on pull request #8693:
URL: https://github.com/apache/flink/pull/8693#issuecomment-626524792


   @rkhachatryan Would you please review this PR again?
   I have split into two commits now and rebase with latest code.



This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org




[GitHub] [flink] Myasuka commented on pull request #8693: [FLINK-8871] Support to cancel checkpoing via notification

2020-04-30 Thread GitBox


Myasuka commented on pull request #8693:
URL: https://github.com/apache/flink/pull/8693#issuecomment-621674425


   @rkhachatryan Thanks for your review!
   I'll rebase master code on top of https://github.com/apache/flink/pull/11721 
recently.
   
   1. > does it make sense to split the commit (or even PR) into several pieces 
addressing JM and TM separately?
   
   I also previous single commit is a bit large, I'll split into several 
smaller ones.
   
   2. > can you add a component name to the commit message?
   
   Sure, no problem. It was my fault to missing component information.
   
   3. > AsyncCheckpointRunnable.cancelled seems not to be used. And why not to 
add a new AsyncCheckpointState instead (or reuse DISCARDED)?
   
   Previous `AsyncCheckpointRunnable.cancelled` is used for testing to verify 
whether this async runnable is cancelled. The reason why I not add a new 
`AsyncCheckpointState` is due to I reuse `AsyncCheckpointRunnable#close` to 
cancel the running thread which would set the state as `DISCARDED`. I think 
this is an open question to discuss to search for a better solution.
   
   4. > I don't see how AsyncCheckpointRunnable is canceled or its thread is 
stopped; which is the original goal of this PR I believe
   
   As I said above, I reuse `AsyncCheckpointRunnable#close` to cancel the 
running async checkpoint task.
   
   
   
   



This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org