Home | About | Sematext search-lucene.com search-hadoop.com
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB
 Search Hadoop and all its subprojects:

Switch to Plain View
HBase >> mail # dev >> Re: infinite loop of RS_ZK_REGION_SPLIT on .94.2


+
lars hofhansl 2012-11-04, 00:37
+
rajesh babu chintaguntla 2012-11-04, 02:10
+
Matt Corgan 2012-11-04, 02:27
+
Ted Yu 2012-11-04, 22:07
+
Matt Corgan 2012-11-05, 01:16
+
Jean-Daniel Cryans 2012-11-05, 17:52
+
Matt Corgan 2012-11-05, 22:02
+
Jean-Daniel Cryans 2012-11-05, 22:15
+
Matt Corgan 2012-11-06, 01:33
+
Ted Yu 2012-11-06, 03:07
+
Matt Corgan 2012-11-06, 05:10
+
lars hofhansl 2012-11-06, 05:17
+
Matt Corgan 2012-11-06, 05:28
+
lars hofhansl 2012-11-06, 05:45
+
ramkrishna vasudevan 2012-11-06, 05:59
Copy link to this message
-
Re: infinite loop of RS_ZK_REGION_SPLIT on .94.2
The log shows that the first time the region was transitioned to SPLITTING
even then it was not populated in the Master's memory.

On Tue, Nov 6, 2012 at 11:29 AM, ramkrishna vasudevan <
[EMAIL PROTECTED]> wrote:

> Could you attach the master logs at this time
> 2012-11-05 00:24:55?
>
> Regards
> Ram
>
> On Tue, Nov 6, 2012 at 11:15 AM, lars hofhansl <[EMAIL PROTECTED]>wrote:
>
>> Took a brief look through all SPLIT related commits since 0.94.0... Found
>> these:
>>
>> HBASE-6854 *
>> HBASE-6713
>> HBASE-6329 *
>>
>> HBASE-6088
>>
>> HBASE-5986
>> HBASE-6070 *
>>
>>
>> The ones marked with * are (IMHO) more likely to be related.
>>
>> -- Lars
>>
>> ________________________________
>> From: Matt Corgan <[EMAIL PROTECTED]>
>> To: dev <[EMAIL PROTECTED]>; lars hofhansl <[EMAIL PROTECTED]>
>> Sent: Monday, November 5, 2012 9:28 PM
>> Subject: Re: infinite loop of RS_ZK_REGION_SPLIT on .94.2
>>
>> Yeah - we were running .94.0 since it came out but never saw it there.
>> I'll keep trying to narrow it down.  The only harm it's causing is log
>> spam and failing to move daughters to a new regionserver, which are
>> definitely problems, but it's not bringing down the cluster.
>>
>>
>> On Mon, Nov 5, 2012 at 9:17 PM, lars hofhansl <[EMAIL PROTECTED]>
>> wrote:
>>
>> > So it seems you can repeat this to some extend in 0.94.2, but you have
>> > never seen this before?
>> >
>> >
>> > -- Lars
>> >
>> >
>> >
>> > ________________________________
>> >  From: Matt Corgan <[EMAIL PROTECTED]>
>> > To: dev <[EMAIL PROTECTED]>
>> > Sent: Monday, November 5, 2012 9:10 PM
>> > Subject: Re: infinite loop of RS_ZK_REGION_SPLIT on .94.2
>> >
>> > It happened in this new table that I have all the logs for.  The region
>> in
>> > question this time is 6839663e4f8f79be3d7469784c21cbc2, and the first
>> trace
>> > of this region is on the regionserver with the "Intantiated
>> tableName..."
>> > message
>> >
>> > 2012-11-05 22:24:21,162 DEBUG
>> org.apache.hadoop.hbase.regionserver.HRegion:
>> > Instantiated
>> >
>> >
>> StatAreaModelLink,\x00\x00\x07\xD9\x00\x00\x00\x0C\x00\x00\x00\x004H\xC4\xB5\x00\x00\x00\x02\x00\x00\x00\x05\x00\x00\x00\x00G.l\x9B,1352172257535.6839663e4f8f79be3d74
>> > 9784c21cbc2.
>> >
>> > I also know this region came from a recent split, but I can't find any
>> log
>> > messages show the parent finishing the split that created this daughter
>> > region.  So my guess now is that the split is actually finishing and
>> > letting clients continue to write data, but something is failing to
>> print
>> > the log line and correctly tell the master about the new region.
>> >
>> > I've noticed that these regions are showing up on the clients in calls
>> to
>> > HTable.getRegionLocations(), so the clients continue to function, but
>> if I
>> > call HBaseAdmin.move() i get an UnknownRegionException.
>> >
>> >
>> > On Mon, Nov 5, 2012 at 7:07 PM, Ted Yu <[EMAIL PROTECTED]> wrote:
>> >
>> > > I think what Matt encountered was this:
>> > > https://issues.apache.org/jira/browse/HBASE-7101
>> > >
>> > > On Mon, Nov 5, 2012 at 5:33 PM, Matt Corgan <[EMAIL PROTECTED]>
>> wrote:
>> > >
>> > > > Beats me - i can't find any record of it before that.
>> > > >
>> > > > I'm importing data into another table now.  I disabled/enabled the
>> > table
>> > > > first to make sure we have the original 4 region locations logged
>> > > > everywhere.  Will report back...
>> > > >
>> > > >
>> > > > On Mon, Nov 5, 2012 at 2:15 PM, Jean-Daniel Cryans <
>> > [EMAIL PROTECTED]
>> > > > >wrote:
>> > > >
>> > > > > Wow, and can you figure how this happened?
>> > > > >
>> > > > > 2012-11-05 00:24:54,538 DEBUG
>> > > > > org.apache.hadoop.hbase.regionserver.HRegion: Instantiated
>> > > > >
>> > > > >
>> > > >
>> > >
>> >
>> ActiveListingRecord16,\x86\x07\xDC\x03\x17RealtyTrac\x0044737383,1352093084141.22f8fa73d8af837410ca270f344f6bb8.
>> > > > >
>> > > > > Split? Else how did the master told that RS to open a region that
>> > > doesn't
>> > > > > exist?
+
Matt Corgan 2012-11-06, 08:12
+
ramkrishna vasudevan 2012-11-06, 10:07
+
ramkrishna vasudevan 2012-11-06, 11:29
+
lars hofhansl 2012-11-06, 15:13
+
Matt Corgan 2012-11-06, 23:36
+
Matt Corgan 2012-11-05, 01:20
+
ramkrishna vasudevan 2012-11-04, 17:12
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB