Hi Esteban,
Thank you so much for the response, it was indeed the problem. Had a
followup question of this, why is this RPC call not respecting the HBase
RPC time-out
set for 90 seconds by us.
Thanks
On Thu, May 30, 2019 at 10:59 PM Esteban Gutierrez
<este...@cloudera.com.invalid> wrote:
> Hello Abhishek,
>
> The HBase master UI will try to list the snapshots if there are available.
> Since GetCompletedSnapshots goes to the filesystem to scan for all the
> possible snapshots that can take some time depending on the number of
> snapshots or the load of the NN.
>
> thanks,
> esteban.
>
>
> --
> Cloudera, Inc.
>
>
>
> On Thu, May 30, 2019 at 12:05 PM Abhishek Gupta <abhila...@gmail.com>
> wrote:
>
> > Hi,
> >
> > We are trying to debug an issue where multiple listSnapshots calls seem
> to
> > be running extremely slow RPC calls on HBase Master.
> > These calls appear to be happening from Master node itself. I wanted to
> > understand if there is any Master chore service that needs to
> periodically
> > do an operation which involves a lisSnapshots?
> >
> >
> *{"call":"GetCompletedSnapshots(org.apache.hadoop.hbase.protobuf.generated.MasterProtos$GetCompletedSnapshotsRequest)","starttimems":1559233674685,"responsesize":175421,"method":"GetCompletedSnapshots","processingtimems":420615,"client":"
> > 10.0.1.193:58296
> > <http://10.0.1.193:58296>","queuetimems":0,"class":"HMaster"}*
> > Some of these calls are taking 1000 seconds of processing time.
> >
> > Thanks in advance,
> > Abhishek
> >
>
