ignite-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Semen Boikov (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (IGNITE-1911) Discovery MessageWorker thread moves node to "zombie" state when fails
Date Thu, 19 Nov 2015 07:35:11 GMT

    [ https://issues.apache.org/jira/browse/IGNITE-1911?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15013073#comment-15013073
] 

Semen Boikov commented on IGNITE-1911:
--------------------------------------

Denis,

I is not safe to stop node from the RingMessageWorker since stop method will interrupt and
join RingMessageWorker thread, let's stop node from another thread.
Also please improve test to check that not only node left event was received, but Ignite instance
was really stopped.

> Discovery MessageWorker thread moves node to "zombie" state when fails
> ----------------------------------------------------------------------
>
>                 Key: IGNITE-1911
>                 URL: https://issues.apache.org/jira/browse/IGNITE-1911
>             Project: Ignite
>          Issue Type: Bug
>    Affects Versions: ignite-1.4
>            Reporter: Denis Magda
>            Assignee: Denis Magda
>            Priority: Critical
>             Fix For: 1.5
>
>         Attachments: ignite-1911.patch
>
>
> If MessageWorker thread fails by some reason (i.e. because of uncaught exception) it
will move a node to "zombie" state. The node will still accept messages through SocketReader
but won't process them. 
> At some point of time the node will fail because the message queue will be overflowed.

> Such a node must be stopped as soon as MessageWorker terminates.



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

Mime
View raw message