[jira] [Comment Edited] (SOLR-15639) S3 tests are needlessly slow.

["Mark Robert Miller (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SOLR-15639?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17420393#comment-17420393 ] 

Mark Robert Miller edited comment on SOLR-15639 at 9/26/21, 9:24 PM:
---------------------------------------------------------------------

Got super sick for a bit there, but I'll be back to this shortly - I'll try and break this issue out first. They are a bit intertwined, because I turned off various close/shutdown buffers. That's where the "needlessly" adjective refers to - issues that are not part of the actual testing.

With these changes, S3IncrementalBackupTest and related has very little tear down overhead and in isolation will run on my machine in just under 30 seconds. With a lot of test jvms that can drop to about 50 seconds, but like I said, that's out of the long tail I'm trying to address in my runs. The Hdfs version is about 10 seconds off that, but again, the tear down improvements don't show up in what junit pushes out for times. Some of these other tests also gain in their behavior, but given their current times, I'm not so worried about them, but I did look at making sure they have quick tear downs as well, as they do add up all together.

Before I try and break this work out it, I may take a quick look at this other mock and it's tear down behavior.


was (Author: markrmiller):
Got super sick for a bit there, but I'll be back to this shortly - I'll try and break this issue out first. They are a bit intertwined, because I turned off various close/shutdown buffers. That's where the "needlessly" adjective refers to - issues that are not part of the actual testing.

With these changes, S3IncrementalBackupTest and related has very little tear down overhead and in isolation will run on my machine in just under 30 seconds. With a lot of test jvms that can drop to about 50 seconds, but like I said, that's out of the long tail I'm trying to address in my runs. The Hdfs version is about 10 seconds off that, but again, the tear down improvements don't show up in what junit pushes out for times. Some of these other tests other gain in their behavior, but given their current times, I'm not so worried about them, but I did look at making sure they have quick tear downs as well, as they do add up all together.

Before I try and breaking this work out it, I may take a quick look at this other mock and it's tear down behavior.

> S3 tests are needlessly slow.
> -----------------------------
>
>                 Key: SOLR-15639
>                 URL: https://issues.apache.org/jira/browse/SOLR-15639
>             Project: Solr
>          Issue Type: Test
>      Security Level: Public(Default Security Level. Issues are Public) 
>          Components: contrib - S3 Repository, Tests
>            Reporter: Mark Robert Miller
>            Assignee: Mark Robert Miller
>            Priority: Major
>




--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@solr.apache.org
For additional commands, e-mail: issues-help@solr.apache.org