hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "stack (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HBASE-515) At least double default timeouts between regionserver and master
Date Fri, 14 Mar 2008 03:45:24 GMT

    [ https://issues.apache.org/jira/browse/HBASE-515?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12578588#action_12578588
] 

stack commented on HBASE-515:
-----------------------------

Review please (its one line change)

> At least double default timeouts between regionserver and master 
> -----------------------------------------------------------------
>
>                 Key: HBASE-515
>                 URL: https://issues.apache.org/jira/browse/HBASE-515
>             Project: Hadoop HBase
>          Issue Type: Improvement
>    Affects Versions: 0.1.0, 0.2.0, 0.16.0
>            Reporter: stack
>            Priority: Blocker
>         Attachments: 515.patch
>
>
> 501 added logging of how long we sleep at the end of the HRegionServer main run method
before we send heartback back to the master.  Last night during upload I saw this:
> {code}
> 2008-03-13 00:03:50,884 WARN org.apache.hadoop.hbase.util.Sleeper: We slept ten times
longer than scheduled: 3000
> {code}
> Above log has since been improved but its saying that we slept > 30 seconds, the default
timeout on master/regionserver communications (When the lease expires, master starts giving
regions to someone else and when this regionserver reports in, its told to close all its regions).
> Server was under load from other processes but still...upload rate was not that rabid.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message