cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Yuki Morishita (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-7909) Do not exit nodetool repair when receiving JMX NOTIF_LOST
Date Tue, 07 Oct 2014 16:22:34 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-7909?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14162061#comment-14162061
] 

Yuki Morishita commented on CASSANDRA-7909:
-------------------------------------------

Hi [~philip.doctor],

If we lost the last message, then user has to terminate manually, or (I think) wait until
JMX times out.
However, NOTIF_LOST happens more likely during repair, since JMX notifications can be emitted
in parallel for several ranges, compare to the last JMX message to indicate repair completion
which is just sent once after all ranges finish.

> Do not exit nodetool repair when receiving JMX NOTIF_LOST
> ---------------------------------------------------------
>
>                 Key: CASSANDRA-7909
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-7909
>             Project: Cassandra
>          Issue Type: Improvement
>            Reporter: Yuki Morishita
>            Assignee: Yuki Morishita
>            Priority: Trivial
>             Fix For: 2.0.11, 2.1.1
>
>         Attachments: 0001-Do-not-quit-nodetool-when-JMX-notif_lost-received.patch
>
>
> {{nodetool repair}} prints out 'Lost notification...' and exits when JMX NOTIF_LOST message
is received. But we should not exit right away since that message just indicates some messages
are lost because "they arrive so fast that they cannot be delivered to the remote client quickly
enough" according to https://weblogs.java.net/blog/emcmanus/archive/2007/08/when_can_jmx_no.html.
So we should just continue to listen to events until repair finishes or connection is really
closed.



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

Mime
View raw message