You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "Josh Elser (JIRA)" <ji...@apache.org> on 2017/03/24 20:04:41 UTC

[jira] [Updated] (HBASE-17833) Flakey TestSpaceQuotasWithSnapshots

     [ https://issues.apache.org/jira/browse/HBASE-17833?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Josh Elser updated HBASE-17833:
-------------------------------
    Attachment: HBASE-17833.001.patch

.001 Stabilizes these tests by waiting for a "non-changing" value. There was a case where the test would proceed because a value was within our expected "ballpark", but there system actually would sent another small region size report. Thus, the final expectation of the test would be wrong and the test would never complete because we calculated the wrong result.

> Flakey TestSpaceQuotasWithSnapshots
> -----------------------------------
>
>                 Key: HBASE-17833
>                 URL: https://issues.apache.org/jira/browse/HBASE-17833
>             Project: HBase
>          Issue Type: Sub-task
>            Reporter: Josh Elser
>            Assignee: Josh Elser
>         Attachments: HBASE-17833.001.patch
>
>
> I've started seeing some intermittent unit test failures in this class.
> After digging in, it seems like the expected usage size is ~10KB different than what we want. The tests wait to see the region size reports (physical size of the table) to reach a certain range, but it seems like sometimes there is a small change after the test already decided what the final "size" was.
> Should be able to stabilize the test by waiting for the region size report to be constant over a few iterations.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)