[
https://issues.apache.org/jira/browse/HBASE-22895?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sakthi updated HBASE-22895:
---------------------------
Description:
branch-2.2 shows this tests frequent failure in one of the precommit jobs
(808,809). Also master branch has this test on it's flakey list. Looks like our
TestSpaceQuotas is pretty overloaded because of all the sleeps(of 5 seconds) in
the write-until-policy-violated attempts. Also to verify the violation, our
number of retries is 10 and each time we sleep 5 secs before checking again. So
we could often cross the 13 mins limit and get timeout eventually leading to
appearance of TestSpaceQuotas in the flakey list.
An e.g. here:
[https://builds.apache.org/view/H-L/view/HBase/job/HBase-adhoc-run-tests/27/]
Let's refactor & break TestSpaceQuotas into more tests so that none of the
tests cross the 13 min limit as we cannot make this into an IntegrationTest.
was:
branch-2.2 shows this tests frequent failure in one of the rest precommit jobs
(808,809). Also master branch has this test on it's flakey list. Looks like our
TestSpaceQuotas is pretty overloaded because of all the sleeps(of 5 seconds) in
the write-until-policy-violated attempts. Also to verify the violation, our
number of retries is 10 and each time we sleep 5 secs before checking again. So
we could often cross the 13 mins limit and get timeout eventually leading to
appearance of TestSpaceQuotas in the flakey list.
An e.g. here:
[https://builds.apache.org/view/H-L/view/HBase/job/HBase-adhoc-run-tests/27/]
Let's refactor & break TestSpaceQuotas into more tests so that none of the
tests cross the 13 min limit as we cannot make this into an IntegrationTest.
> Fix the flakey TestSpaceQuotas
> ------------------------------
>
> Key: HBASE-22895
> URL: https://issues.apache.org/jira/browse/HBASE-22895
> Project: HBase
> Issue Type: Task
> Components: Quotas, test
> Reporter: Sakthi
> Assignee: Sakthi
> Priority: Major
> Labels: Quota, Space
>
> branch-2.2 shows this tests frequent failure in one of the precommit jobs
> (808,809). Also master branch has this test on it's flakey list. Looks like
> our TestSpaceQuotas is pretty overloaded because of all the sleeps(of 5
> seconds) in the write-until-policy-violated attempts. Also to verify the
> violation, our number of retries is 10 and each time we sleep 5 secs before
> checking again. So we could often cross the 13 mins limit and get timeout
> eventually leading to appearance of TestSpaceQuotas in the flakey list.
> An e.g. here:
> [https://builds.apache.org/view/H-L/view/HBase/job/HBase-adhoc-run-tests/27/]
>
> Let's refactor & break TestSpaceQuotas into more tests so that none of the
> tests cross the 13 min limit as we cannot make this into an IntegrationTest.
>
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
