You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by tillrohrmann <gi...@git.apache.org> on 2017/01/25 13:20:21 UTC

[GitHub] flink pull request #3210: [FLINK-5638] [asyncIO] Fix deadlock when closing t...

GitHub user tillrohrmann opened a pull request:

    https://github.com/apache/flink/pull/3210

    [FLINK-5638] [asyncIO] Fix deadlock when closing two chained AsyncWaitOperators

    This PR is a backport of #3209 onto `release-1.2.`
    
    This PR addresses the problem by changing the Emitter's behaviour to first output the
    element before removing it from the StreamElementQueue. That way the close method waits
    until also the Emitter has outputted the last completed element. Additionally, the
    stopResources method now frees the checkpoint lock in order to let the emitter thread
    react to the interrupt signal.

You can merge this pull request into a Git repository by running:

    $ git pull https://github.com/tillrohrmann/flink asyncIOFixDeadlockBackport

Alternatively you can review and apply these changes as the patch at:

    https://github.com/apache/flink/pull/3210.patch

To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:

    This closes #3210
    
----
commit 770589dd286b1bc30d9bf79d813bc30a371e1995
Author: Till Rohrmann <tr...@apache.org>
Date:   2017-01-25T13:11:48Z

    [FLINK-5638] [asyncIO] Fix deadlock when closing two chained AsyncWaitOperators
    
    This PR addresses the problem by changing the Emitter's behaviour to first output the
    element before removing it from the StreamElementQueue. That way the close method waits
    until also the Emitter has outputted the last completed element. Additionally, the
    stopResources method now frees the checkpoint lock in order to let the emitter thread
    react to the interrupt signal.

----


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---

[GitHub] flink pull request #3210: [backport] [FLINK-5638] [asyncIO] Fix deadlock whe...

Posted by tillrohrmann <gi...@git.apache.org>.
Github user tillrohrmann closed the pull request at:

    https://github.com/apache/flink/pull/3210


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---

[GitHub] flink issue #3210: [backport] [FLINK-5638] [asyncIO] Fix deadlock when closi...

Posted by tillrohrmann <gi...@git.apache.org>.
Github user tillrohrmann commented on the issue:

    https://github.com/apache/flink/pull/3210
  
    Merged


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---

[GitHub] flink issue #3210: [FLINK-5638] [asyncIO] Fix deadlock when closing two chai...

Posted by StephanEwen <gi...@git.apache.org>.
Github user StephanEwen commented on the issue:

    https://github.com/apache/flink/pull/3210
  
    It's a bit hacky to release the lock in scope via `wait()` but seems to work.
    
    One more reason to go for a post-box model...
    
    +1


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---

[GitHub] flink issue #3210: [backport] [FLINK-5638] [asyncIO] Fix deadlock when closi...

Posted by tillrohrmann <gi...@git.apache.org>.
Github user tillrohrmann commented on the issue:

    https://github.com/apache/flink/pull/3210
  
    Agreed, the problem is a little bit that you don't control from where the `close` method is called... Thanks for the review @StephanEwen.
    
    Merging once Travis has passed.


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---