You are viewing a plain text version of this content. The canonical link for it is here.

Posted to notifications@yetus.apache.org by "Sean Busbey (JIRA)" <ji...@apache.org> on 2015/10/27 22:20:27 UTC

[jira] [Created] (YETUS-139) plugins for project specific infra-error handling

Sean Busbey created YETUS-139:
---------------------------------

             Summary: plugins for project specific infra-error handling
                 Key: YETUS-139
                 URL: https://issues.apache.org/jira/browse/YETUS-139
             Project: Yetus
          Issue Type: New Feature
          Components: Test Patch
            Reporter: Sean Busbey


Hadoop and HBase both have some project specific infra failures they check for. It would be nice ot have a generalized approach so others can leverage the same logic before realizing they need it.

* Hadoop Bad Perms - some hadoop tests change permissions in the target/ dir to simulate disk failures. These changes then wedge Jenkins because neither git nor rm can delete the files. Essentially, we'd need to walk the filesystem after test to ensure all directories are executable.
* HBase Zombies - HBase has some tests that historically crashed in a way that left a process around, eventually bringing down Jenkins worker hosts. So they have some code for finding them to report and clean up.



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