You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "Sho Shimauchi (JIRA)" <ji...@apache.org> on 2012/11/19 07:13:01 UTC
[jira] [Created] (HBASE-7183) print WARN message if
hbase.replication.sizeOfLogQueue is too big
Sho Shimauchi created HBASE-7183:
------------------------------------
Summary: print WARN message if hbase.replication.sizeOfLogQueue is too big
Key: HBASE-7183
URL: https://issues.apache.org/jira/browse/HBASE-7183
Project: HBase
Issue Type: Improvement
Components: Replication
Reporter: Sho Shimauchi
A metric hbase.replication.sizeOfLogQueue may become big when replication is delaying.
It would be useful if HBase prints WARN log which tells hbase.replication.sizeOfLogQueue is too big.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HBASE-7183) print WARN message if
hbase.replication.sizeOfLogQueue is too big
Posted by "Jean-Daniel Cryans (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-7183?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13510068#comment-13510068 ]
Jean-Daniel Cryans commented on HBASE-7183:
-------------------------------------------
[~sho.shimauchi], do you have a preference as to what constitutes "too big"? 2? 5? Configurable?
Personally I'd be fine with alerting at 2 because any lag is usually bad but we can't alert at 1 since we'd get spurs of WARNs when the log is rolled.
Next we also need to decide when the alert is sent. In trunk I added a new stats thread that runs every 5 minutes but it's not very amenable to printing this kind of warning, so I'd rather do it when the log is rolled. In the worst case, like when there's an extended network partition, the WARN will be printed as often as logs are rolled (for each RS).
> print WARN message if hbase.replication.sizeOfLogQueue is too big
> -----------------------------------------------------------------
>
> Key: HBASE-7183
> URL: https://issues.apache.org/jira/browse/HBASE-7183
> Project: HBase
> Issue Type: Improvement
> Components: Replication
> Reporter: Sho Shimauchi
>
> A metric hbase.replication.sizeOfLogQueue may become big when replication is delaying.
> It would be useful if HBase prints WARN log which tells hbase.replication.sizeOfLogQueue is too big.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HBASE-7183) print WARN message if
hbase.replication.sizeOfLogQueue is too big
Posted by "Sho Shimauchi (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-7183?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13510116#comment-13510116 ]
Sho Shimauchi commented on HBASE-7183:
--------------------------------------
Hi JD,
2 is fine as default but it should also be configurable, because the best parameter depends on the size of a cluster.
Alerting when the log is rolled is good. I think the worst case scenario is acceptable because network partition itself is really bad and user have to fix the situation asap.
> print WARN message if hbase.replication.sizeOfLogQueue is too big
> -----------------------------------------------------------------
>
> Key: HBASE-7183
> URL: https://issues.apache.org/jira/browse/HBASE-7183
> Project: HBase
> Issue Type: Improvement
> Components: Replication
> Reporter: Sho Shimauchi
>
> A metric hbase.replication.sizeOfLogQueue may become big when replication is delaying.
> It would be useful if HBase prints WARN log which tells hbase.replication.sizeOfLogQueue is too big.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira