Hi,

When I do an hbase hbck, I'm getting this:

13/01/30 19:29:55 WARN hdfs.DFSClient: Failed to connect to
/192.168.23.7:50010, add to deadNodes and continuejava.io.IOException:
Got error for OP_READ_BLOCK, self=/192.168.23.7:57612,
remote=/192.168.23.7:50010, for file
/hbase/entry/df36c172b5b6523467eb676b70e94e26/.regioninfo, for block
-6941943399693151592_200976

Any idea what it means?

I'm getting many errors like that, and when I'm trying to run MR jobs,
I'm getting 2013-01-30 16:58:59,575 INFO
org.apache.hadoop.hbase.regionserver.HRegionServer: Scanner
1275691821504852773 lease expired on region....

Can it be related?

Thanks,

JM

So.

After many retries, repairs, etc, I was able to get rid of the
OP_READ_BLOCK exception.

But I still have this one in the hbck output:
13/01/30 21:42:38 WARN regionserver.StoreFile: Failed match of store
file name hdfs://node3:9000/hbase/.META./1028785192/.oldlogs/hlog.1341486023008
13/01/30 21:42:38 WARN regionserver.StoreFile: Failed match of store
file name hdfs://node3:9000/hbase/-ROOT-/70236052/.oldlogs/hlog.1341486022850

I'm not sure to get what  this one is.

Sorry about the title. I was going to write about the shortcircuit and
did not figured I did not removed it. Those issues happend when I
tried to activate it. I have reverted the changes for now...

JM

2013/1/30, Jean-Marc Spaggiari <[EMAIL PROTECTED]>:
> Hi,
>
> When I do an hbase hbck, I'm getting this:
>
> 13/01/30 19:29:55 WARN hdfs.DFSClient: Failed to connect to
> /192.168.23.7:50010, add to deadNodes and continuejava.io.IOException:
> Got error for OP_READ_BLOCK, self=/192.168.23.7:57612,
> remote=/192.168.23.7:50010, for file
> /hbase/entry/df36c172b5b6523467eb676b70e94e26/.regioninfo, for block
> -6941943399693151592_200976
>
> Any idea what it means?
>
> I'm getting many errors like that, and when I'm trying to run MR jobs,
> I'm getting 2013-01-30 16:58:59,575 INFO
> org.apache.hadoop.hbase.regionserver.HRegionServer: Scanner
> 1275691821504852773 lease expired on region....
>
> Can it be related?
>
> Thanks,
>
> JM
>