Fix the bug of unregistered workers in worker pool by jovany-wang · Pull Request #7343 · ray-project/ray

jovany-wang · 2020-02-27T01:37:39Z

What's the issue

The reason for Java CI issues is figured out:

In CheckpointableTest and ReconstrcutionTest, we will kill a worker process to trigger the failover of actor. There're multiple worker threads in a worker process, once we kill a worker process which has some worker threads not registered to the raylet, the worker threads will be bimzie workers. Then StartWorkerProcess will return early at https://github.com/ray-project/ray/blob/master/src/ray/raylet/worker_pool.cc#L137

This also fix another case testDeleteObject in direct call, otherwise the ci couldn't pass.

How to Fix

Add a timer for worker process to check if the worker is timeout to register to raylet.

AmplabJenkins · 2020-02-27T01:45:42Z

Can one of the admins verify this patch?

src/ray/raylet/worker_pool.cc

src/ray/raylet/worker_pool.h

src/ray/raylet/worker_pool.cc

AmplabJenkins · 2020-02-27T04:14:02Z

Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/22482/
Test FAILed.

AmplabJenkins · 2020-02-27T04:51:26Z

Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/22475/
Test FAILed.

AmplabJenkins · 2020-02-27T06:06:57Z

Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/22483/
Test FAILed.

AmplabJenkins · 2020-02-27T06:21:01Z

Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/22484/
Test FAILed.

AmplabJenkins · 2020-02-27T08:41:04Z

Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/22486/
Test PASSed.

AmplabJenkins · 2020-02-28T02:20:56Z

Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/22534/
Test FAILed.

AmplabJenkins · 2020-02-28T02:57:13Z

Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/22530/
Test PASSed.

AmplabJenkins · 2020-02-28T03:59:55Z

Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/22541/
Test FAILed.

AmplabJenkins · 2020-02-28T04:59:39Z

Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/22540/
Test FAILed.

AmplabJenkins · 2020-02-28T06:17:52Z

Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/22546/
Test PASSed.

AmplabJenkins · 2020-02-28T07:39:01Z

Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/22551/
Test FAILed.

kfstorm

Thanks for the fix! LGTM. Please fix the checkstyle warnings though.

kfstorm

LGTM. Thanks!

AmplabJenkins · 2020-02-28T12:15:27Z

Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/22557/
Test PASSed.

src/ray/raylet/worker_pool_test.cc

src/ray/raylet/worker_pool.h

src/ray/raylet/worker_pool.cc

raulchen · 2020-02-28T14:05:07Z

Also, consider changing the title to something like Fix the bug of unregistered workers in worker pool. Because this is not a Java-specific bug.

raulchen

Looks good to me. Please fix the small issues before merging.

Co-Authored-By: Hao Chen <chenh1024@gmail.com>

AmplabJenkins · 2020-02-28T15:35:02Z

Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/22560/
Test FAILed.

AmplabJenkins · 2020-02-28T18:19:08Z

Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/22562/
Test PASSed.

AmplabJenkins · 2020-02-29T05:38:54Z

Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/22580/
Test FAILed.

AmplabJenkins · 2020-03-02T01:57:15Z

Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/Ray-PRB/22600/
Test FAILed.

* Fix * Fix * Fix complie * Fix lint * Fix linting * Fix testDeleteObject * Fix linting * Update src/ray/raylet/worker_pool.cc Co-Authored-By: Hao Chen <chenh1024@gmail.com> * Update src/ray/raylet/worker_pool.cc Co-Authored-By: Hao Chen <chenh1024@gmail.com> * Update src/ray/raylet/worker_pool.h Co-Authored-By: Hao Chen <chenh1024@gmail.com> * Update src/ray/raylet/worker_pool.cc Co-Authored-By: Hao Chen <chenh1024@gmail.com> * Address comments. * FIx linting Co-authored-by: Hao Chen <chenh1024@gmail.com>

jovany-wang mentioned this pull request Feb 27, 2020

[Java] Fix java CI hang #7042

Closed

3 tasks

raulchen requested review from kfstorm and raulchen February 27, 2020 01:53

kfstorm reviewed Feb 27, 2020

View reviewed changes

src/ray/raylet/worker_pool.cc Outdated Show resolved Hide resolved

raulchen reviewed Feb 27, 2020

View reviewed changes

src/ray/raylet/worker_pool.cc Outdated Show resolved Hide resolved

src/ray/raylet/worker_pool.h Outdated Show resolved Hide resolved

src/ray/raylet/worker_pool.cc Outdated Show resolved Hide resolved

jovany-wang force-pushed the ci_fixing branch from b8ffa9e to 26dedc2 Compare February 27, 2020 05:18

Fix

fcd1a06

jovany-wang force-pushed the ci_fixing branch from 418dfc6 to fcd1a06 Compare February 27, 2020 06:49

jovany-wang added 2 commits February 28, 2020 09:17

Fix

2e1b4c3

Fix complie

b14f1a6

Fix lint

abb77fe

jovany-wang force-pushed the ci_fixing branch from 0181a65 to abb77fe Compare February 28, 2020 03:14

Fix linting

8012609

Fix testDeleteObject

4f658fe

kfstorm approved these changes Feb 28, 2020

View reviewed changes

Fix linting

d22b5ef

kfstorm reviewed Feb 28, 2020

View reviewed changes

raulchen reviewed Feb 28, 2020

View reviewed changes

raulchen approved these changes Feb 28, 2020

View reviewed changes

jovany-wang and others added 4 commits February 28, 2020 22:52

Update src/ray/raylet/worker_pool.cc

05d18a4

Co-Authored-By: Hao Chen <chenh1024@gmail.com>

Update src/ray/raylet/worker_pool.cc

8745e14

Co-Authored-By: Hao Chen <chenh1024@gmail.com>

Update src/ray/raylet/worker_pool.h

828116e

Co-Authored-By: Hao Chen <chenh1024@gmail.com>

Update src/ray/raylet/worker_pool.cc

16cccbe

Co-Authored-By: Hao Chen <chenh1024@gmail.com>

jovany-wang changed the title ~~[Fix Java CI] Add a timer for zombie worker process~~ Fix the bug of unregistered workers in worker pool Feb 28, 2020

Address comments.

e544af7

FIx linting

1740dde

Merge branch 'master' into ci_fixing

ab0fe31

jovany-wang merged commit 2771af1 into ray-project:master Mar 2, 2020

jovany-wang deleted the ci_fixing branch March 2, 2020 08:30

Conversation

jovany-wang commented Feb 27, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What's the issue

How to Fix

Uh oh!

AmplabJenkins commented Feb 27, 2020

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

AmplabJenkins commented Feb 27, 2020

Uh oh!

AmplabJenkins commented Feb 27, 2020

Uh oh!

AmplabJenkins commented Feb 27, 2020

Uh oh!

AmplabJenkins commented Feb 27, 2020

Uh oh!

AmplabJenkins commented Feb 27, 2020

Uh oh!

AmplabJenkins commented Feb 28, 2020

Uh oh!

AmplabJenkins commented Feb 28, 2020

Uh oh!

AmplabJenkins commented Feb 28, 2020

Uh oh!

AmplabJenkins commented Feb 28, 2020

Uh oh!

AmplabJenkins commented Feb 28, 2020

Uh oh!

AmplabJenkins commented Feb 28, 2020

Uh oh!

kfstorm left a comment

Choose a reason for hiding this comment

Uh oh!

kfstorm left a comment

Choose a reason for hiding this comment

Uh oh!

AmplabJenkins commented Feb 28, 2020

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

raulchen commented Feb 28, 2020

Uh oh!

raulchen left a comment

Choose a reason for hiding this comment

Uh oh!

AmplabJenkins commented Feb 28, 2020

Uh oh!

AmplabJenkins commented Feb 28, 2020

Uh oh!

AmplabJenkins commented Feb 29, 2020

Uh oh!

AmplabJenkins commented Mar 2, 2020

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

jovany-wang commented Feb 27, 2020 •

edited

Loading