[
https://issues.apache.org/jira/browse/HBASE-20701?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16518411#comment-16518411
]
Mike Drob commented on HBASE-20701:
-----------------------------------
The patch makes sense, thanks for contributing [~mihir6692]!
Can you address the checkstyle concern about line length from QA-Bot and upload
a patch for branch-1? I'll take care of applying it to the other branches from
there.
> too much logging when balancer runs from BaseLoadBalancer
> ---------------------------------------------------------
>
> Key: HBASE-20701
> URL: https://issues.apache.org/jira/browse/HBASE-20701
> Project: HBase
> Issue Type: Improvement
> Components: Balancer
> Reporter: Monani Mihir
> Assignee: Monani Mihir
> Priority: Trivial
> Attachments: HBASE-20701-branch-1.3.patch,
> HBASE-20701-branch-1.3.patch, HBASE-20701-branch-1.3.patch,
> HBASE-20701-branch-1.4.patch, HBASE-20701-branch-1.4.patch
>
>
> When balancer runs, it tries to find least loaded server with better locality
> for current region. During this, we make debug level logging for each of
> those regions. It creates too much amount of logging at debug level , we
> should move this to trace level logging.
> {code:java}
> int getLeastLoadedTopServerForRegion (int region, int currentServer) {
> ...
> if (leastLoadedServerIndex != -1) {
> LOG.debug("Pick the least loaded server " +
> servers[leastLoadedServerIndex].getHostname()
> + " with better locality for region " + regions[region]);
> }
> ...
> }{code}
> This was fixed in branch-2.0 as part of -HBASE-14614-
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
