hbase-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Lukáš Drbal <lukas.dr...@gmail.com>
Subject Re: NPE in restore_snapshot
Date Tue, 03 Sep 2013 10:55:18 GMT
Thanks for response Matteo. I haven't problem with this NPE, i want only
report it. Good to know about HBASE-8760.

Lukas

2013/9/1 Matteo Bertozzi <theo.bertozzi@gmail.com>

> one region of the table has no SERVER_QUALIFIER, that means it was never
> assigned.
> and you get the NPE during the snapshot taken before restoring.
> The workaround is drop the table and use clone_snapshot instead.
>
> (The NPE was fixed as part of HBASE-8760, which should be part of 0.94.12)
>
>
>
> Matteo
>
>
>
> On Sat, Aug 31, 2013 at 7:45 PM, Lukáš Drbal <lukas.drbal@gmail.com>
> wrote:
>
> > Ach, sorry.
> > Hbase 0.94.11 and hadoop 1.0.3
> >
> >
> > 2013/8/31 Ted Yu <yuzhihong@gmail.com>
> >
> > > What version of HBase are you using ?
> > >
> > > According to output, the NPE seems to point to the following line
> > > in DisabledTableSnapshotHandler.java (tip of 0.94):
> > >
> > >       LOG.info("Starting to write region info and WALs for regions for
> > > offline snapshot:"
> > >           + SnapshotDescriptionUtils.toString(snapshot));
> > >
> > > It is unlikely the statement which caused NPE. Knowing the release
> would
> > > help pinpoint the exact location.
> > >
> > > Cheers
> > >
> > >
> > > On Sat, Aug 31, 2013 at 11:25 AM, Lukáš Drbal <lukas.drbal@gmail.com>
> > > wrote:
> > >
> > > > Hi,
> > > > i catch NPE when i try restore my snapshot.
> > > >
> > > > Snapshot is taken from online table and exported to another cluster
> > via:
> > > > hbase org.apache.hadoop.hbase.snapshot.ExportSnapshot -snapshot
> > > > fb_posts-2013-08-31 -copy-to hdfs://hadoops-master:9000/hbase -chuser
> > > hbase
> > > > -chgroup hbase -mappers 14
> > > >
> > > >
> > > > OT: files on hadoops-master cluster has permission hadoop:hbase
> (HDFS).
> > > > Maybe a mistake? I specify -chuser and -chroup in export : ) I fix it
> > > with
> > > > hadoop fs -chown -R hbase:hbase /hbase
> > > >
> > > > hbase(main):006:0> list_snapshots
> > > > SNAPSHOT
> > >  TABLE
> > > > + CREATION
> > > > TIME
> > > >
> > > >  fb_comments-2013-08-27
> > > > fb_comments (Tue Aug 27 08:42:01 +0000
> > > > 2013)
> > > >
> > > >  fb_posts-2013-08-26
> > > > fb_posts (Mon Aug 26 19:26:04 +0000
> > > > 2013)
> > > >
> > > >  fb_posts-2013-08-31
> > > > fb_posts (Sat Aug 31 17:22:52 +0000
> > > > 2013)
> > > >
> > > > 3 row(s) in 0.0220 seconds
> > > >
> > > > SnapshotInfo: http://pastebin.com/dLS6tzTJ
> > > >
> > > > I see here a lot NOT_FOUND files, duno why and that's imho reason for
> > > this
> > > > fail.
> > > >
> > > > hbase(main):007:0> list
> > > > TABLE
> > > >
> > > > fb_comments
> > > >
> > > > fb_posts
> > > >
> > > > fb_posts_old
> > > >
> > > > test
> > > >
> > > > 4 row(s) in 0.0180 seconds
> > > >
> > > > fb_posts_old are cloned snapshot from fb_posts-2013-08-26
> > > >
> > > > hbase(main):008:0> is_disabled 'fb_posts'
> > > > true
> > > >
> > > > 0 row(s) in 0.0080 seconds
> > > >
> > > > hbase(main):009:0> restore_snapshot 'fb_posts-2013-08-31'
> > > >
> > > > output error: http://pastebin.com/tFsnGLha
> > > >
> > > > Master log: http://pastebin.com/yfLjJyfc
> > > >
> > > > RS logs are clean :( Log level are set to DEBUG
> > > >
> > > >
> > > > Please excuse my bad english.
> > > >
> > > > --
> > > > Save The World - http://www.worldcommunitygrid.org/
> > > >
> > http://www.worldcommunitygrid.org/stat/viewMemberInfo.do?userName=LesTR
> > > >
> > > > LesTR
> > > >
> > >
> >
> >
> >
> > --
> > Save The World - http://www.worldcommunitygrid.org/
> > http://www.worldcommunitygrid.org/stat/viewMemberInfo.do?userName=LesTR
> >
> > LesTR
> >
>



-- 
Save The World - http://www.worldcommunitygrid.org/
http://www.worldcommunitygrid.org/stat/viewMemberInfo.do?userName=LesTR

LesTR

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message