Hi,
DFS trouble. Have you taken the recommended steps according
to this Wiki page:
http://wiki.apache.org/hadoop/Hbase/Troubleshooting
?
Try the steps for #5, #6, and #7.
And/or, try adding more data nodes to spread the load.
Hope that helps,
- Andy
> 2009-04-14 16:17:08,718 INFO
Hi all,
The insert operation is still executing, but there are region servers
getting down now and then. The log info shows that they are shutdown for
different reasons. Here is another failed region server's log:
2009-04-14 16:17:08,718 INFO org.apache.hadoop.hbase.regionserver.HLog:
removing
hi JD,
I tried your solution by upgrading hbase to 0.19.1 and applying the
patch. The inserting mapreduce application has been running for more than
half an hour, we lost one region server and here is the log on the lost
region server:
2009-04-14 16:08:11,483 FATAL
org.apache.hadoop.hbase.regi
I put up a v4 on this issue. You should use that one
instead. Please be advised this is still experimental.
- Andy
> From: Jean-Daniel Cryans
> Subject: Re: Region Server lost response when doing BatchUpdate
> To: hbase-user@hadoop.apache.org
> Date: Monday, April 13, 2009, 5:4
hi Jean-Daniel,
As you said, we were inserting data using sequential pattern, and if we
use random pattern there would not be such prolem.
I'm trying hbase 0.19.1 and the patch now.
Thanks!
2009/4/13 Jean-Daniel Cryans
> I see that your region server had 5188 store files in 121 store
I see that your region server had 5188 store files in 121 store, I'm
99% sure that it's the cause of your OOME. Luckily for you, we've been
working on this issue since last week. What you should do :
- Upgrade to HBase 0.19.1
- Apply the latest patch in
https://issues.apache.org/jira/browse/HBASE
