You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@mesos.apache.org by "Jie Yu (JIRA)" <ji...@apache.org> on 2015/12/04 20:42:11 UTC

[jira] [Commented] (MESOS-4068) Multiple log coordinators can be elected concurrently

    [ https://issues.apache.org/jira/browse/MESOS-4068?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15042078#comment-15042078 ] 

Jie Yu commented on MESOS-4068:
-------------------------------

Looked at the test case. It's possible that coord1 was elected first and then demoted by coord3 who got elected later. In that case, num_elected will be 2.

> Multiple log coordinators can be elected concurrently
> -----------------------------------------------------
>
>                 Key: MESOS-4068
>                 URL: https://issues.apache.org/jira/browse/MESOS-4068
>             Project: Mesos
>          Issue Type: Bug
>          Components: replicated log
>            Reporter: Neil Conway
>            Assignee: Neil Conway
>              Labels: mesosphere
>         Attachments: rep_log_multiple_coord_test_case-1.patch
>
>
> Attached is a test case (reduced from a test case submitted by Michael Maged at IBM).
> Running {{mesos-tests --gtest_filter="CoordinatorTest.SimulationDriver" --gtest_repeat=1000}} consistently fails in < 100 iterations on a CentOS 7 VM. However, it only failed ~once in 12,000 iterations on the Mac OS X 10.10 host.



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