[ 
https://issues.apache.org/jira/browse/HBASE-16775?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15965877#comment-15965877
 ] 

Jonathan Hsieh commented on HBASE-16775:
----------------------------------------

Hm.. this is pretty subtle.  Looking at the test case and if I understand 
correctly, we have different export to directories and different table names in 
each case and on the surface the test looks to be doing reasonable things to 
try isolate concurrent runs of versions of the test. (and failing because mr 
tmp dirs are trampling each other).  

This makes most important change in the patch the HTU#getDataTestDir call which 
calls HTU#setupDataTestDir which setup MR localizing properties[1].  

If that is the case I'm +1 for the patch if you add more comments about this 
specific trampling scenario in each place where you currently have this comment 
"// This will setup separate directory for use in MR cluster."

As a follow up, is there any reason why this should't always be run whenever a 
miniMR cluster is setup?

[1] 
https://github.com/apache/hbase/blob/master/hbase-common/src/test/java/org/apache/hadoop/hbase/HBaseCommonTestingUtility.java#L102

> Flakey test with TestExportSnapshot#testExportRetry and 
> TestMobExportSnapshot#testExportRetry 
> ----------------------------------------------------------------------------------------------
>
>                 Key: HBASE-16775
>                 URL: https://issues.apache.org/jira/browse/HBASE-16775
>             Project: HBase
>          Issue Type: Bug
>    Affects Versions: 2.0.0
>            Reporter: huaxiang sun
>            Assignee: Appy
>         Attachments: disable.patch, HBASE-16775.master.001.patch, 
> HBASE-16775.master.002.patch, HBASE-16775.master.003.patch, 
> HBASE-16775.master.004.patch, HBASE-16775.master.005.patch, 
> HBASE-16775.master.006.patch, HBASE-16775.master.007.patch
>
>
> The root cause is that conf.setInt("mapreduce.map.maxattempts", 10) is not 
> taken by the mapper job, so the retry is actually 0. Debugging to see why 
> this is the case.



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

Reply via email to