You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Wei-dun Teng (JIRA)" <ji...@apache.org> on 2013/12/03 13:12:38 UTC

[jira] [Commented] (CASSANDRA-4880) Endless loop flushing+compacting system/schema_keyspaces and system/schema_columnfamilies

    [ https://issues.apache.org/jira/browse/CASSANDRA-4880?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13837624#comment-13837624 ] 

Wei-dun Teng commented on CASSANDRA-4880:
-----------------------------------------

I have encountered similar bug after I upgrade one of my 1.2.2 node to 1.2.12.
Was using FreeBSD 8.2 + diablo-jre-1.6.0.07.02_18.

Before I upgrading the node to 1.2.12, I changed JVM to openjdk-7.25.15_2 (in retrospective probably not a good idea ...), saw flipping Memtable flushes, changes JVM back to diablo-jre-1.6.0.07.02_18, and still seeing rapid flushes. Now I've taken the 1.2.12 node offline (but not decommissioned it).

After that I see tens of Memtable flushes on the 1.2.12 per second, while once or twice a second on nodes with 1.2.2.

Attached files are the flipping schema versions.

> Endless loop flushing+compacting system/schema_keyspaces and system/schema_columnfamilies
> -----------------------------------------------------------------------------------------
>
>                 Key: CASSANDRA-4880
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-4880
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 1.1.6, 1.2.0 beta 1
>         Environment: Linux x86_64 3.4.9, sun-jdk 1.6.0_33
>            Reporter: Mina Naguib
>            Assignee: Pavel Yaskevich
>             Fix For: 1.1.7, 1.2.0 beta 3
>
>         Attachments: 131203-schema-1.txt, 131203-schema-2.txt, CASSANDRA-4880-fix.patch, CASSANDRA-4880.patch
>
>
> After upgrading a node from 1.1.2 to 1.1.6, the startup sequence entered a loop as seen here:
> http://mina.naguib.ca/misc/cassandra_116_startup_loop.txt
> Stopping and starting the node entered the same loop.
> Reverting back to 1.1.2 started successfully.



--
This message was sent by Atlassian JIRA
(v6.1#6144)