Hive: Add timeout for TestHiveIcebergStorageHandlerWithEngine tests #2448
Conversation
This file contains hidden or 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
Collaborator
|
Can we use as a global timeout rule in the class, and only override it for that single test method where more time is needed? |
Contributor
Author
I have not known about this before. Added timeout by a global rule |
marton-bod
approved these changes
Apr 9, 2021
Contributor
Author
|
Based on the logs the current issue is that the HMS thread pool is exhausted: This might, or might not be the root cause. |
Contributor
Author
|
We need the pool size change. Failed with the original size a second time |
Contributor
Author
|
@RussellSpitzer: Could you please review? Thanks, |
RussellSpitzer
approved these changes
Apr 13, 2021
Member
RussellSpitzer
left a comment
RussellSpitzer
left a comment
There was a problem hiding this comment.
Looks reasonable to me :)
aokolnychyi
reviewed
Apr 13, 2021
| public void start() { | ||
| // Create a copy of the HiveConf for the metastore | ||
| metastore.start(new HiveConf(hs2Conf)); | ||
| metastore.start(new HiveConf(hs2Conf), 10); |
Contributor
There was a problem hiding this comment.
Do we know why the pool is exhausted? In the past, we had a few leaks in the Spark catalog code which led to this. It can be also a valid use case too if we simply need a larger pool.
Contributor
There was a problem hiding this comment.
Maybe, worth creating an issue if we want to take a look later.
Contributor
There was a problem hiding this comment.
Or is it related to the recent change around how we manage the pool?
Contributor
Author
There was a problem hiding this comment.
I have 2 ideas:
- Maybe this was caused by the Hive: Configure catalog type on table level. #2129 where we start to handle / cache the catalogs differently
- Or we added some more queries and Hive creates / caches HMSClients and we might ended up using a different codepath.
Created #2474
Contributor
Author
|
Thanks @marton-bod, @massdosage and @RussellSpitzer for the review! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
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.
In #2437 we seem to hit an issue where the tests are failing because of resource problems.
Github infra handles the timeouts in this case, but when there is a timeout the logs are not collected.
We should handle the timeouts on the tests itself. This is a good practice and in this case we might be able to squeeze out more information about the problems from the logs.
Used a 40s timeout in most tests (usually they run around 5-10 s on my local machine), and used 100s for longer tests (they run around 20-30 s on my local machine). Let's see how they behave on the CI infra.