You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by "huanliwang-db (via GitHub)" <gi...@apache.org> on 2023/02/24 19:34:41 UTC
[GitHub] [spark] huanliwang-db opened a new pull request, #40162: [SPARK-42566][SS] RocksDB StateStore lock acquisition should happen after getting input iterator from inputRDD
huanliwang-db opened a new pull request, #40162:
URL: https://github.com/apache/spark/pull/40162
The current behavior of the `compute` method in both `StateStoreRDD` and `ReadStateStoreRDD` is: we first get the state store instance and then get the input iterator for the inputRDD.
For RocksDB state store, the running task will acquire and hold the lock for this instance. The retried task or speculative task will fail to acquire the lock and eventually abort the job if there are some network issues. For example, When we shrink the executors, the alive one will try to fetch data from the killed ones because it doesn't know the target location (prefetched from the driver) is dead until it tries to fetch data. The query might be hanging for a long time as the executor will retry `spark.shuffle.io.maxRetries=3` times and for each retry wait for `spark.shuffle.io.connectionTimeout` (default value is 120s) before timeout. In total, the task could be hanging for about 6 minutes. And the retried or speculative tasks won't be able to acquire the lock in this period.
Making lock acquisition happen after retrieving the input iterator should be able to avoid this situation.
### What changes were proposed in this pull request?
<!--
Please clarify what changes you are proposing. The purpose of this section is to outline the changes and how this PR fixes the issue.
If possible, please consider writing useful notes for better and faster reviews in your PR. See the examples below.
1. If you refactor some codes with changing classes, showing the class hierarchy will help reviewers.
2. If you fix some SQL features, you can provide some references of other DBMSes.
3. If there is design documentation, please add the link.
4. If there is a discussion in the mailing list, please add the link.
-->
Making lock acquisition happen after retrieving the input iterator.
### Why are the changes needed?
<!--
Please clarify why the changes are needed. For instance,
1. If you propose a new API, clarify the use case for a new API.
2. If you fix a bug, you can clarify why it is a bug.
-->
Avoid the failure like the following when there is a network issue
```
java.lang.IllegalStateException: StateStoreId(opId=0,partId=3,name=default): RocksDB instance could not be acquired by
[ThreadId: Some(47), task: 3.1 in stage 57, TID 793] as it was not released by [ThreadId: Some(51), task: 3.1 in stage 57,
TID 342] after 60003 ms.
```
### Does this PR introduce _any_ user-facing change?
<!--
Note that it means *any* user-facing change including all aspects such as the documentation fix.
If yes, please clarify the previous behavior and the change this PR proposes - provide the console output, description and/or an example to show the behavior difference if possible.
If possible, please also clarify if this is a user-facing change compared to the released Spark versions or within the unreleased branches such as master.
If no, write 'No'.
-->
No
### How was this patch tested?
<!--
If tests were added, say they were added here. Please make sure to add some test cases that check the changes thoroughly including negative and positive cases if possible.
If it was tested in a way different from regular unit tests, please clarify how you tested step by step, ideally copy and paste-able, so that other reviewers can test and check, and descendants can verify in the future.
If tests were not added, please describe why they were not added and/or why it was difficult to add.
If benchmark tests were added, please run the benchmarks in GitHub Actions for the consistent environment, and the instructions could accord to: https://spark.apache.org/developer-tools.html#github-workflow-benchmarks.
-->
existing UT should be good enough
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] HeartSaVioR closed pull request #40162: [SPARK-42566][SS] RocksDB StateStore lock acquisition should happen after getting input iterator from inputRDD
Posted by "HeartSaVioR (via GitHub)" <gi...@apache.org>.
HeartSaVioR closed pull request #40162: [SPARK-42566][SS] RocksDB StateStore lock acquisition should happen after getting input iterator from inputRDD
URL: https://github.com/apache/spark/pull/40162
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] HeartSaVioR commented on pull request #40162: [SPARK-42566][SS] RocksDB StateStore lock acquisition should happen after getting input iterator from inputRDD
Posted by "HeartSaVioR (via GitHub)" <gi...@apache.org>.
HeartSaVioR commented on PR #40162:
URL: https://github.com/apache/spark/pull/40162#issuecomment-1445041550
Thanks! Merging to master.
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org