hbase-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ramakrishna s vasudevan <ramkrishna.vasude...@huawei.com>
Subject RE: Regions in Transition?
Date Sat, 21 Jan 2012 06:28:49 GMT
Hi 

Check the master side logs and in the RS logs check what the RS is doinig w.r.t to the region
mentioned in the
below trace.

Hope you are trying out in 0.90.x and not 0.92?

Regards
Ram
________________________________________
From: Harsh J [harsh@cloudera.com]
Sent: Saturday, January 21, 2012 11:47 AM
To: user@hbase.apache.org
Subject: Re: Regions in Transition?

(Just to avoid confusion, Lars is talking about HBase Table CF's blocksize, unrelated to HDFS
block sizes, which would be fine if set to 512m.)

On 21-Jan-2012, at 9:49 AM, lars hofhansl wrote:

> Also, you probably do not want to set your blocksize to 512mb. The default is 64k.
> HBase has to load (either HDFS or cache) and scan this amount of data for each key value
lookup.
>
>
> -- Lars
>
>
>
> ________________________________
> From: Stack <stack@duboce.net>
> To: user@hbase.apache.org
> Sent: Friday, January 20, 2012 1:16 PM
> Subject: Re: Regions in Transition?
>
> On Fri, Jan 20, 2012 at 10:41 AM, Mark <static.void.dev@gmail.com> wrote:
>
>> I recently bumped up the region size memory configuration on our HBase
>> cluster and after doing a rolling restart of our 5 nodes I saw the
>> following in the HBase status page:
>>
>>
>>    Regions in Transition
>>
>> Region  State
>> 32cb0e36cfa326d0a431734ba93a16**df items,869239091/es-LA,**1323971864141.*
>> *32cb0e36cfa326d0a431734ba93a16**df. state=PENDING_OPEN, ts=Fri Jan 20
>> 10:30:56 PST 2012 (422s ago), server=hadoop3.mycompany.com,**
>> 60020,1327082744696
>> 3f70e7e8734ca7b772e2a4db8e3d27**3f items,287955002/fr-FR,**1324492747210.*
>> *3f70e7e8734ca7b772e2a4db8e3d27**3f. state=PENDING_OPEN, ts=Fri Jan 20
>> 10:08:07 PST 2012 (1791s ago), server=hadoop4.mycompany.com,**
>> 60020,1327082844453
>> 4851ec840abe39e3d70d834866fc62**e1 item_translations,589372741/**
>> es-LA,1323511796967.**4851ec840abe39e3d70d834866fc62**e1.
>> state=PENDING_OPEN, ts=Fri Jan 20 10:08:07 PST 2012 (1791s ago), server=
>> hadoop4.mycompany.com,**60020,1327082844453
>> 8966449d7bb08dcd060ff4bf9fd12d**36 item_translations,450498891/**
>> es-LA,1324512750098.**8966449d7bb08dcd060ff4bf9fd12d**36.
>> state=PENDING_OPEN, ts=Fri Jan 20 10:08:07 PST 2012 (1791s ago), server=
>> hadoop4.mycompany.com,**60020,1327082844453
>>
>>
>> What does this mean I why are they taking so long to transition? FYI I am
>> using Cloudera CDH3U2 and my block size is configured to 512m.
>>
>>
> Region is stuck.  Master assigned it to a node but its not opening.  Can
> you see in master logs where this region was assigned.  Then go to that
> regionservers's ui and see if you see the region?
>
> St.Ack

--
Harsh J
Customer Ops. Engineer, Cloudera


Mime
View raw message