hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sergey Shelukhin (JIRA)" <j...@apache.org>
Subject [jira] [Created] (HBASE-8178) IPCUtil produces a LOT of WARN log spew
Date Fri, 22 Mar 2013 18:39:15 GMT
Sergey Shelukhin created HBASE-8178:
---------------------------------------

             Summary: IPCUtil produces a LOT of WARN log spew
                 Key: HBASE-8178
                 URL: https://issues.apache.org/jira/browse/HBASE-8178
             Project: HBase
          Issue Type: Bug
          Components: IPC/RPC
            Reporter: Sergey Shelukhin
            Priority: Critical


I was running some integration tests on trunk and noticed that IT log (which uses LoadTestTool
with 16k records) is overflowing with this:
{code}
13/03/21 23:02:15 WARN ipc.IPCUtil: Buffer grew from 16384 to 16971
13/03/21 23:02:15 WARN ipc.IPCUtil: Buffer grew from 16384 to 18906
13/03/21 23:02:15 WARN ipc.IPCUtil: Buffer grew from 16384 to 18486
13/03/21 23:02:15 WARN ipc.IPCUtil: Buffer grew from 16384 to 22896
13/03/21 23:02:15 WARN ipc.IPCUtil: Buffer grew from 16384 to 22253
{code}

So much in fact, that
{code}
reznor-mbp:stripe sergey$ grep -a -c "Buffer grew" it.log 
2769498
reznor-mbp:stripe sergey$ grep -a -v "Buffer grew" it.log > it2.log
reznor-mbp:stripe sergey$ ll it*
-rw-r--r--  1 sergey  staff  189440541 Mar 22 11:22 it.log
-rw-r--r--  1 sergey  staff    1114677 Mar 22 11:33 it2.log
{code}

There are two questions.
1) Should this log be TRACE level, or there at all? it happens a lot.
2) Is this condition normal? I am assuming the log level was set to WARN for a reason :)




--
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