Sumit:
You can find doc on upgrade here:
http://hbase.apache.org/book.html#hbase.rolling.restart
bq. Are there two versions of HBase
You can upgrade from 0.98.2 to 0.98.3
0.98.4 RC0 would come out this weekend.
Cheers
On Fri, Jul 11, 2014 at 12:51 PM, Sumit Mohanty
wrote:
> Nice.
>
> As such
e:
> >> >
> >> > > The action update, at its full capacity, is what the current build
> >> action
> >> > > does:
> >> > > user may want to change hdfs, hbase config values and use new image
> >> > > (compatible with the previous image)
ate, at its full capacity, is what the current build
>> action
>> > > does:
>> > > user may want to change hdfs, hbase config values and use new image
>> > > (compatible with the previous image).
>> > >
>> > > One solution is to give
current build
> action
> > > does:
> > > user may want to change hdfs, hbase config values and use new image
> > > (compatible with the previous image).
> > >
> > > One solution is to give action update the same capacity as action
> build -
> &g
at its full capacity, is what the current build action
> > does:
> > user may want to change hdfs, hbase config values and use new image
> > (compatible with the previous image).
> >
> > One solution is to give action update the same capacity as action build -
> >
on update the same capacity as action build -
> without the constraint of aforementioned check.
Perhaps that would be more appropriate. “Build” has certain semantic
implications that don’t seem to correlate to what is being attempted.
>
> Cheers
>
>
> On Fri, Jul 11, 2014 at 1
The action update, at its full capacity, is what the current build action
does:
user may want to change hdfs, hbase config values and use new image
(compatible with the previous image).
One solution is to give action update the same capacity as action build -
without the constraint of
Why would I want to leverage “build” to update the cluster rather than the
discussed “update” action? In both instances I don’t imagine the parameter
change should take place prior to a restart. Correct?
— Jon
On Jul 11, 2014, at 1:19 PM, Ted Yu wrote:
> Hi,
> When I issue action bu
Hi,
When I issue action build on an existing cluster, I get:
2014-07-11 17:15:05,377 [main] ERROR main.ServiceLauncher - Application
Instance dir already exists: hdfs://
hor10n03.gq1.ygridcore.net:8020/user/yarn/.slider/cluster/cl1
2014-07-11 17:15:05,378 [main] INFO util.ExitUtil - Exiting with