You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@hbase.apache.org by Shahab Yunus <sh...@gmail.com> on 2015/02/04 21:04:01 UTC
Behavior of Split method in HBaseAdmin
If we programmatically split a table by async method in HBaseAdmin class
and even after waiting for quite a while, the split does not happen (there
is no difference in number of regions before and after the split call) and
there is no error or an exception either, does it mean that there is
something wrong or is there a valid scenario foe this behavior as well?
I am providing the split point as well.
Cdh5.3.0 is the version.
Thanks.
Regards,
Shahab
Re: Behavior of Split method in HBaseAdmin
Posted by Stack <st...@duboce.net>.
You should see your split request being logged in the master logs.
SPLITTING_NEW would seem to indicate splits happening.
Open master log, run your command, and then see what shows in the log
thereafter.
St.Ack
On Wed, Feb 4, 2015 at 3:47 PM, Shahab Yunus <sh...@gmail.com> wrote:
> Thank for the reply Stack.
>
> What exactly should be I looking for? I checked the logs and I don't see
> any error, 'can't, 'don't or 'fail' kind of message. What kind of error or
> text?
>
> I do see stuff like:
> 2015-02-04 18:16:25,754 INFO org.apache.hadoop.hbase.master.RegionStates:
> Transitioned *null* to {f9cea11544eb80b6aa7c3f5a2fe75181
> state=SPLITTING_NEW, ts=1423091785754,
> server=ip-10-12-0-111.vf1.internal,60020,1422918135701}
>
>
> Regards,
> Shahab
>
> On Wed, Feb 4, 2015 at 3:07 PM, Stack <st...@duboce.net> wrote:
>
> > See master log when split call comes in. It will say why the split did
> not
> > happen.
> > St.Ack
> >
> > On Wed, Feb 4, 2015 at 12:04 PM, Shahab Yunus <sh...@gmail.com>
> > wrote:
> >
> > > If we programmatically split a table by async method in HBaseAdmin
> class
> > > and even after waiting for quite a while, the split does not happen
> > (there
> > > is no difference in number of regions before and after the split call)
> > and
> > > there is no error or an exception either, does it mean that there is
> > > something wrong or is there a valid scenario foe this behavior as well?
> > >
> > > I am providing the split point as well.
> > >
> > > Cdh5.3.0 is the version.
> > >
> > > Thanks.
> > >
> > > Regards,
> > > Shahab
> > >
> >
>
Re: Behavior of Split method in HBaseAdmin
Posted by Shahab Yunus <sh...@gmail.com>.
Thank for the reply Stack.
What exactly should be I looking for? I checked the logs and I don't see
any error, 'can't, 'don't or 'fail' kind of message. What kind of error or
text?
I do see stuff like:
2015-02-04 18:16:25,754 INFO org.apache.hadoop.hbase.master.RegionStates:
Transitioned *null* to {f9cea11544eb80b6aa7c3f5a2fe75181
state=SPLITTING_NEW, ts=1423091785754,
server=ip-10-12-0-111.vf1.internal,60020,1422918135701}
Regards,
Shahab
On Wed, Feb 4, 2015 at 3:07 PM, Stack <st...@duboce.net> wrote:
> See master log when split call comes in. It will say why the split did not
> happen.
> St.Ack
>
> On Wed, Feb 4, 2015 at 12:04 PM, Shahab Yunus <sh...@gmail.com>
> wrote:
>
> > If we programmatically split a table by async method in HBaseAdmin class
> > and even after waiting for quite a while, the split does not happen
> (there
> > is no difference in number of regions before and after the split call)
> and
> > there is no error or an exception either, does it mean that there is
> > something wrong or is there a valid scenario foe this behavior as well?
> >
> > I am providing the split point as well.
> >
> > Cdh5.3.0 is the version.
> >
> > Thanks.
> >
> > Regards,
> > Shahab
> >
>
Re: Behavior of Split method in HBaseAdmin
Posted by Stack <st...@duboce.net>.
See master log when split call comes in. It will say why the split did not
happen.
St.Ack
On Wed, Feb 4, 2015 at 12:04 PM, Shahab Yunus <sh...@gmail.com>
wrote:
> If we programmatically split a table by async method in HBaseAdmin class
> and even after waiting for quite a while, the split does not happen (there
> is no difference in number of regions before and after the split call) and
> there is no error or an exception either, does it mean that there is
> something wrong or is there a valid scenario foe this behavior as well?
>
> I am providing the split point as well.
>
> Cdh5.3.0 is the version.
>
> Thanks.
>
> Regards,
> Shahab
>