Thanks Stack.. will be trying this out.

I had one more question: What are the scenarios under which I can get null
regioninfo. Can it be attributed to manual maintenance tasks like
major-compact or offline merge ? If I want to do a online periodic merge of
all my empty regions inoder to keep the number of regions under control,
shud I also, always look for these regions with null-region info and delete
them from META periodically ?

On Wed, Nov 23, 2011 at 10:01 PM, Stack <[email protected]> wrote:

> On Wed, Nov 23, 2011 at 4:39 AM, kisalay <[email protected]> wrote:
> > This query is regarding the online_merge script from
> > https://issues.apache.org/jira/browse/HBASE-1621
> >
> > My setup is hbase - cdh3u1, hadoop-cdh3-u0.
> >
> > When I am running this script, I encounter the problem at the following
> > line:
> >
> >  info =
> >
> Writables.getHRegionInfo(result.getValue(HConstants::CATALOG_FAMILY,HConstants::REGIONINFO_QUALIFIER))
> >
> > The
> result.getValue(HConstants::CATALOG_FAMILY,HConstants::REGIONINFO_QUALIFIER)
> > returns a null and Writable.getHRegionInfo throws a
> null-pointer-exception.
> >
> > I am wondering in what cases can I get null returned
> > from
> result.getValue(HConstants::CATALOG_FAMILY,HConstants::REGIONINFO_QUALIFIER)
> > ?
>
> Change the script to use
>
> http://hbase.apache.org/apidocs/org/apache/hadoop/hbase/util/Writables.html#getHRegionInfoOrNull(byte[])
> instead.  If it returns null, continue rather than proceed.  Or clean
> out these rows with null regioninfo before running the script by
> deleting them from .META.  You should see a list of what to delete if
> you look in your master log.  It should have complaint about rows with
> empty info:regioninfos.
>
> St.Ack
>

Reply via email to