You are viewing a plain text version of this content. The canonical link for it is here.
Posted to notifications@accumulo.apache.org by "Josh Elser (JIRA)" <ji...@apache.org> on 2016/07/08 18:40:11 UTC

[jira] [Commented] (ACCUMULO-4361) Test race conditions in ShellServerIT

    [ https://issues.apache.org/jira/browse/ACCUMULO-4361?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15368159#comment-15368159 ] 

Josh Elser commented on ACCUMULO-4361:
--------------------------------------

I know we have an open issue somewhere to fix the "zookeeper config propagation" race condition issue once and for all, but my search terms are failing me.

> Test race conditions in ShellServerIT
> -------------------------------------
>
>                 Key: ACCUMULO-4361
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-4361
>             Project: Accumulo
>          Issue Type: Bug
>          Components: test
>            Reporter: Josh Elser
>            Assignee: Josh Elser
>            Priority: Minor
>             Fix For: 1.7.3, 1.8.0
>
>
> [~romil.choksi] found a test issue where ShellServerIT.getAuths() would intermittently fail.
> This is because of the use of ZooCache inside of ZKAuthorizor. The minicluster has two tabletservers. If the shell is so unlucky to update the authorizations on one tserver but then ask for the authorizations on the other tserver, the test is subject to the propagation of the ZK watcher update to fire.
> We should just switch over ShellServerIT to run a minicluster with only one tserver than try to hack around this in the test itself with retries.



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