hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jean-Daniel Cryans (Created) (JIRA)" <j...@apache.org>
Subject [jira] [Created] (HBASE-4574) "Node deleted but still in RIT" printed too often
Date Tue, 11 Oct 2011 22:05:11 GMT
"Node deleted but still in RIT" printed too often
-------------------------------------------------

                 Key: HBASE-4574
                 URL: https://issues.apache.org/jira/browse/HBASE-4574
             Project: HBase
          Issue Type: Bug
    Affects Versions: 0.92.0
            Reporter: Jean-Daniel Cryans
             Fix For: 0.92.0


Looking at the 0.92 master logs, I see I often get this message:

bq. WARN org.apache.hadoop.hbase.master.AssignmentManager: Node deleted but still in RIT:
TestTable,blah. state=OPEN, ts=1318369648361, server=blah

The issue seems to be due to a race between OpenedRegionHandler and watchers in AssignmentManager.
Specifically, ORH first deletes the znode then deletes the in-memory RIT data structure (via
regionOnline). Between the two steps a watcher is triggered and if it arrives first then it
will see the region still in RIT.

If the message is really supposed to be a warning then in the current form it's useless as
people will see this message 99% of the time because of this race.

--
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

        

Mime
View raw message