accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Eric Newton (JIRA)" <j...@apache.org>
Subject [jira] [Created] (ACCUMULO-798) large ParNew can cause long-term stop-the-world java GC
Date Tue, 09 Oct 2012 20:14:02 GMT
Eric Newton created ACCUMULO-798:
------------------------------------

             Summary: large ParNew can cause long-term stop-the-world java GC
                 Key: ACCUMULO-798
                 URL: https://issues.apache.org/jira/browse/ACCUMULO-798
             Project: Accumulo
          Issue Type: Bug
          Components: tserver
    Affects Versions: 1.4.1
         Environment: running a tserver with a very large memory foot-print (16G)
            Reporter: Eric Newton
            Assignee: Keith Turner
            Priority: Trivial
             Fix For: 1.4.2


While running a tserver with 16G of JVM memory, under continuous ingest, a ParNew collection
took longer than 40 seconds.  This happened on two different servers (out of 40) after performing
continuous ingest for more than 20 hours.

Limiting the ParNew generation to 2G, and never saw this failure again.

Recommend this setting in the examples, even though none of the examples approaches this memory
size.

--
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

Mime
View raw message