[
https://issues.apache.org/jira/browse/HDFS-3803?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13434724#comment-13434724
]
Andy Isaacson commented on HDFS-3803:
-------------------------------------
bq. LOG.debug seems reasonable level for it.
I disagree, without the "Starting a new period" message it's very difficult to
diagnose the block pool scanner in production scenarios. Production cannot run
with debug enabled due to other cases where debug logging is actually
expensive. So making this critical status message a LOG.debug makes it
effectively useless.
I'd be OK with replacing all of these logs with a useful set of (for example)
JMX metrics so that the scanner is more observable in an automatable/graphable
fashion, but until that happens, this message needs to stay in at LOG.info.
> BlockPoolSliceScanner new work period notice is very chatty at INFO level
> -------------------------------------------------------------------------
>
> Key: HDFS-3803
> URL: https://issues.apache.org/jira/browse/HDFS-3803
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: data-node
> Affects Versions: 2.1.0-alpha, 2.0.1-alpha
> Environment: Hadoop 2.0.1-alpha-SNAPSHOT
> Reporter: Andrew Purtell
> Priority: Minor
> Fix For: 2.1.0-alpha, 3.0.0
>
> Attachments: HDFS-3803.patch
>
>
> One line of ~140 chars logged every 5 seconds.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
