[
https://issues.apache.org/jira/browse/HDDS-6083?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17485534#comment-17485534
]
István Fajth commented on HDDS-6083:
------------------------------------
To capture the root of the problem:
The problem is with the selection of failing nodes in
TestOzoneECClient#testNodeFailuresWhileWriting.
If we write two chunks only after a failure, and If we select for example node3
in the block group to fail, then the write will not fail, as we don't actually
write to node3 in this case.
Similarly, if we write one chunk after a failure, and if we select node 2 or
node 3 to fail, then as we don't actually will write to these nodes the write
won't fail.
The problem was solved already with the new allocation logic in the server side
mock via HDDS-6036, and with the test changes in the second pull request we are
good here, and even if we specify in a new test a node to fail which we do not
write later on, the assertions will have to be wrong to pass, while the
behaviour remains consistent there as well.
Thank you [~umamaheswararao] for adjusting the tests!
> EC: Fix flakyness of tests around nodefailures
> ----------------------------------------------
>
> Key: HDDS-6083
> URL: https://issues.apache.org/jira/browse/HDDS-6083
> Project: Apache Ozone
> Issue Type: Sub-task
> Reporter: István Fajth
> Assignee: István Fajth
> Priority: Major
> Labels: pull-request-available
>
> We haven't seen much occurance, but what we have seen a couple of times
> already is this failure:
> {code}
> Error:
> testWriteShouldSuccessIfLessThanParityNodesFail(org.apache.hadoop.ozone.client.TestOzoneECClient)
> Time elapsed: 0.116 s <<< FAILURE!
> java.lang.AssertionError: expected:<2> but was:<1>
> {code}
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@ozone.apache.org
For additional commands, e-mail: issues-h...@ozone.apache.org
