[jira] [Updated] (HBASE-13143) TestCacheOnWrite is flaky and needs a diet

["Andrew Purtell (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/HBASE-13143?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Andrew Purtell updated HBASE-13143:
-----------------------------------
    Attachment: HBASE-13143.patch

What seems critical here is the different cache on write options and whether or not to cache compressed data or not. Removing the block encoder variation drops the number of tests from > 200 to 72.

{noformat}
Running org.apache.hadoop.hbase.io.hfile.TestCacheOnWrite
Tests run: 72, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 144.41 sec - in org.apache.hadoop.hbase.io.hfile.TestCacheOnWrite

Results :

Tests run: 72, Failures: 0, Errors: 0, Skipped: 0
{noformat}


> TestCacheOnWrite is flaky and needs a diet
> ------------------------------------------
>
>                 Key: HBASE-13143
>                 URL: https://issues.apache.org/jira/browse/HBASE-13143
>             Project: HBase
>          Issue Type: Bug
>    Affects Versions: 0.98.11
>            Reporter: Andrew Purtell
>            Assignee: Esteban Gutierrez
>            Priority: Critical
>             Fix For: 2.0.0, 1.2.0, 1.3.0, 0.98.15, 1.0.3, 1.1.3
>
>         Attachments: HBASE-13143.patch
>
>
> TestCacheOnWrite passes locally but has been flaking in 0.98 builds on Jenkins, most recently https://builds.apache.org/job/HBase-0.98/878/
> The test takes a long time to execute (338.492 sec) and is resource intensive (216 tests). Neither of these characteristics endear it to Jenkins.
> When I ran this unit test on a macbook after a minute the fan was running so fast I thought it would take flight. 



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)