Make job fail when partial tasks' pre-dependent tasks finished and exceeds the waiting timeout #621
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Refer to #610 |
|
@zuston thanks.. can you update READ.me as well? also provide some examples and context in description? |
…ceeds the waiting timeout Signed-off-by: zhangjunfan <junfan.zhang@outlook.com>
Signed-off-by: zhangjunfan <junfan.zhang@outlook.com>
Signed-off-by: zhangjunfan <junfan.zhang@outlook.com>
|
@oliverhu Updated. |
oliverhu
approved these changes
Nov 30, 2021
| * tony.application.group.a = worker,chief | ||
| * tony.application.group.b = evaluator | ||
| * | ||
| * tony.application.dependency.b.timeout.after.a = 3600 |
There was a problem hiding this comment.
does b have to be a group? I think worker.timeout.after.a also works from conversations. Can you clarify here?
There was a problem hiding this comment.
Mistake for code comment, i will fix it.
And tony.application.dependency.evaluator.timeout.after.a is best practise.
Comment on lines
+345
to
+347
| public static final String GROUP_REGEX = TONY_APPLICATION_PREFIX + "group\\.([A-Za-z]+)$"; | ||
| public static final String GROUP_DEPEND_TIMEOUT_REGEX = | ||
| TONY_APPLICATION_PREFIX + "dependency\\.([A-Za-z]+)\\.timeout\\.after\\.([A-Za-z]+)$"; |
There was a problem hiding this comment.
also add some unit tests for these regex..
There was a problem hiding this comment.
It has been tested in Utils.getAllGroupJobTypes and Utils.getJobTypeDependentGrps
|
Updating configurations in wiki will be done after this PR merged @oliverhu |
Merged
zuston
added a commit
to zuston/TonY
that referenced
this pull request
Feb 9, 2022
…hed and exceeds the waiting timeout (tony-framework#621) PR: tony-framework#621 * Make job fail when partial tasks' pre-dependent tasks finished and exceeds the waiting timeout * Modify the config * Fix compile error * Correct the code comment Signed-off-by: zhangjunfan <junfan.zhang@outlook.com>
zuston
pushed a commit
to zuston/TonY
that referenced
this pull request
Feb 9, 2022
Backport: Make job fail when partial tasks' pre-dependent tasks finished and exceeds the waiting timeout (tony-framework#621) See merge request !72
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR is to solve the problems of tf workers hang when chief has been finished and other bugs.
For example, tensorflow estimator training job will include some roles of ps/worker/evaluator/chief. Actually, due to the bug of tensorflow or misusing the estimator api, sometimes evaluator will hang. So if we use the configuration as follows, when evaluator is still running after timeout and chief and workers are all finished, the mechanism of dependency group timeout will make job failed.
Conf as follows, the evaluator will be alive 3600(sec) after workers and chief are all finished.
So, this PR introduce the conf of
tony.application.group.{GROUP_NAME}={JOB_TYPES}andtony.application.dependency.{JOB_TYPE}.timeout.after.{GROUP_NAME} = 3600Besides, due to tensorflow bug, sometimes the workers will hang after the chief finished. We could use above conf to solve it. The worker will be alive 3600(sec) after chief finished.