You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by uce <gi...@git.apache.org> on 2016/02/29 16:04:56 UTC
[GitHub] flink pull request: [FLINK-3534] [runtime] Prevent canceling Execu...
GitHub user uce opened a pull request:
https://github.com/apache/flink/pull/1735
[FLINK-3534] [runtime] Prevent canceling Execution from failing
@StephanEwen, can you have a look at this?
This should handle cases where a cancelled execution is failed (as reported by @rmetzger), but still allow the `ExecutionGraph` to be failed (and not suffer from the point you raised in #1726).
Robert would like this in the next RC, but in general I think that it's better to have more exposure for a change to the `ExecutionGraph`... I'm undecided right now. What do you think?
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/uce/flink 3534-prevent_restart
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/flink/pull/1735.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 #1735
----
commit 54c1b12a67ac928b8461865adef6794b82d16325
Author: Ufuk Celebi <uc...@apache.org>
Date: 2016-02-29T14:37:45Z
[FLINK-3534] [runtime] Prevent canceling Execution from failing
----
---
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: [FLINK-3534] [runtime] Prevent canceling Execu...
Posted by StephanEwen <gi...@git.apache.org>.
Github user StephanEwen commented on the pull request:
https://github.com/apache/flink/pull/1735#issuecomment-190261584
Looks good to me!
+1 to merge this
---
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: [FLINK-3534] [runtime] Prevent canceling Execu...
Posted by StephanEwen <gi...@git.apache.org>.
Github user StephanEwen commented on the pull request:
https://github.com/apache/flink/pull/1735#issuecomment-190293717
True, that is rather ugly behavior. I guess the important thing is that at least one of the Executions results in state `failed`.
---
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: [FLINK-3534] [runtime] Prevent canceling Execu...
Posted by uce <gi...@git.apache.org>.
Github user uce commented on the pull request:
https://github.com/apache/flink/pull/1735#issuecomment-190327892
My local [Travis build has passed](https://travis-ci.org/uce/flink/builds/112634854). I will merge this to `master` and `release-1.0`.
---
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: [FLINK-3534] [runtime] Prevent canceling Execu...
Posted by uce <gi...@git.apache.org>.
Github user uce commented on the pull request:
https://github.com/apache/flink/pull/1735#issuecomment-190291277
I've fixed the test failures.
@StephanEwen, the changes can result in the following behaviour: if a slot is released, some of its executions can end up in cancelled rather than failed state. Is that OK? The following happens: The slot's first execution is failed. This triggers an execution graph fail, after which all executions are cancelled. When the slot fails further executions, these are already cancelling and go to state cancelled. Before they would still go state `FAILED`.
---
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: [FLINK-3534] [runtime] Prevent canceling Execu...
Posted by asfgit <gi...@git.apache.org>.
Github user asfgit closed the pull request at:
https://github.com/apache/flink/pull/1735
---
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.
---