accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "John Vines (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (ACCUMULO-4513) Master hangs if shutdown requested while in state "HAS_LOCK"
Date Fri, 04 Nov 2016 17:31:58 GMT

    [ https://issues.apache.org/jira/browse/ACCUMULO-4513?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15637086#comment-15637086
] 

John Vines commented on ACCUMULO-4513:
--------------------------------------

I believe this more prominent on systems with multiple masters in the event of the main master
going down when stopAll has been issued.

> Master hangs if shutdown requested while in state "HAS_LOCK"
> ------------------------------------------------------------
>
>                 Key: ACCUMULO-4513
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-4513
>             Project: Accumulo
>          Issue Type: Bug
>    Affects Versions: 1.6.6, 1.7.2
>            Reporter: John Vines
>            Priority: Critical
>
> The StatusThread.run() method does not have any means to change the masters state out
of HAS_LOCK when CLEAN_STOP is requested, which can cause the master to get perpetually stuck
> Additionally, because the master never hit the NORMAL state, the main Master.run() was
jammed at waitForMetadataUpgrade.await() so the thrift server was never started and shutdown
couldn't be issued directly (not entirely confident that would have actually fixed anything
though this is contrary to the the state transition chart which says we can go from HAVE_LOCK->STOP)



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

Mime
View raw message