[ 
https://issues.apache.org/jira/browse/HDFS-8506?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14570868#comment-14570868
 ] 

Tony Reix commented on HDFS-8506:
---------------------------------

With 18 runs, there are now 39 unstable tests.

New unstable tests
 hdfs.server.namenode.TestCacheDirectives
 ipc.TestIPC
 mapreduce.v2.app.TestRuntimeEstimators
 security.ssl.TestReloadingX509TrustManager
 yarn.client.TestApplicationClientProtocolOnHA

hdfs.server.namenode.TestCacheDirectives                 12    0-1      0      
0 |   1/18
ipc.TestIPC                                                                    
32    0-1      0      0 |   1/18
mapreduce.v2.app.TestRuntimeEstimators                    2    0-1      0      
0 |   1/18
security.ssl.TestReloadingX509TrustManager                5    0-1      0      
0 |   1/18
yarn.client.TestApplicationClientProtocolOnHA             17      0    0-1      
0 |   1/18


> List of 33 Unstable tests on branch-2.7
> ---------------------------------------
>
>                 Key: HDFS-8506
>                 URL: https://issues.apache.org/jira/browse/HDFS-8506
>             Project: Hadoop HDFS
>          Issue Type: Bug
>          Components: test
>    Affects Versions: 2.7.1
>         Environment: Ubuntu / x86_64
> Hadoop branch-2.7 (source code of Monday May 26th)
>            Reporter: Tony Reix
>         Attachments: Unstable-branch-2.7-May26th, 
> UnstableDetails-branch-2.7-May26th
>
>
> On my Ubuntu / x86_64 machine, configured for Hadoop since months, I've run 
> Hadoop tests of branch branch-2.7 (source code of Monday May 26th) during 
> days. It produced 14 runs in the EXACT same environment. And it shows that 
> several tests sometimes fail, randomly.
> 12 runs gave the exact same number of tests done and tests skipped:
>     - 10977 tests
>     -     254 skipped
> 1 test gave only 10972 tests. Another gave only 9760 tests.
> I've used the 12 runs with 10977 tests for building the attached result file, 
> which shows that 33 tests sometimes fail.
> T: Tests
> F: Failures
> E: Errors
> S: Skipped
> NN/n : Number of times the issue appeared out of the 12 runs
> m-M: minimum number of failure up to Maximum number of failures.
> Example:
>                                                                   T      F    
>   E      S |   NN/n
> --------------------------------------------------------------------------------------
> cli.TestHDFSCLI                                        1    0-1      0      0 
> |  11/12
> hdfs.TestAppendSnapshotTruncate          1      0    0-1      0 |   1/12
> ...



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

Reply via email to