Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Plain View
HBase, mail # user - Re: HBase Multiget taking more time


+
Ankit Jain 2013-06-25, 13:34
Copy link to this message
-
Re: HBase Multiget taking more time
Jean-Marc Spaggiari 2013-06-25, 13:36
Hi Ankit,

Attachements are not working well with the mailing list. Can you post
them on pastbin? Also, can you please post the servers (region and
master) logs?

thanks,

JM

2013/6/25 Ankit Jain <[EMAIL PROTECTED]>:
> Hi Jean-Marc/Michael,
>
> Thanks for the reply.
>
> Hardware detail:
> Processor: 8 core
> RAM: 16 GB.
>
> We have allotted 4GB of RAM to HBase and also we are ingesting data into
> HBase in parallel with rate of 50 records(Each record has 20 KB) per sec.
>
> Please find the attached GC log.
>
> Thanks,
> Ankit
>
>
> On Tue, Jun 25, 2013 at 6:03 PM, Ankit Jain <[EMAIL PROTECTED]> wrote:
>>
>> Hi Jean-Marc/Michael,
>>
>> Thanks for the reply.
>>
>> Hardware detail:
>> Processor: 8 core
>> RAM: 16 GB.
>>
>> We have allotted 4GB of RAM to HBase and also we are ingesting data into
>> HBase in parallel with rate of 50 records(Each record has 20 KB) per sec.
>>
>> Please find the attached GC log.
>>
>> Thanks,
>> Ankit
>>
>>
>>
>>
>>
>>
>> On Tue, Jun 25, 2013 at 5:43 PM, Michael Segel <[EMAIL PROTECTED]>
>> wrote:
>>>
>>> Double your timeout from 60K to 120K.
>>> While I don't think its the problem... its just a good idea.
>>>
>>> What happens if you drop the 50 down to 25?  Do you still fail?
>>> If not, go to 35,  etc... until you hit a point where you fail.
>>>
>>> As Jean-Marc said, we need a bit more information. Including some
>>> hardware info too.
>>>
>>> Thx
>>>
>>> -Mike
>>>
>>> On Jun 25, 2013, at 4:03 AM, Ankit Jain <[EMAIL PROTECTED]> wrote:
>>>
>>> > Hi All,
>>> >
>>> > HBase multiget call taking large time and throwing time out exception.
>>> > I am
>>> > retrieving only 50 records in one call. The size of each record is 20
>>> > KB.
>>> >
>>> > java.net.SocketTimeoutException: 60000 millis timeout while waiting for
>>> > channel to be ready for read. ch :
>>> > java.nio.channels.SocketChannel[connected
>>> > local=/192.168.50.122:48695remote=ct-0096/
>>> > 192.168.50.177:60020]
>>> >
>>> > hTable = new HTable(conf, tableName);
>>> > results = hTable.get(rows);
>>> >
>>> > Cluster Detail:
>>> > 1 master, 1 regionserver and 8 regions
>>> >
>>> > --
>>> > Thanks,
>>> > Ankit Jain
>>>
>>
>>
>>
>> --
>> Thanks,
>> Ankit Jain
>
>
>
>
> --
> Thanks,
> Ankit Jain
+
Ankit Jain 2013-06-25, 16:25
+
Ankit Jain 2013-06-25, 09:03
+
Michael Segel 2013-06-25, 12:13
+
Jean-Marc Spaggiari 2013-06-25, 11:59