Hi,
We are using hbase-0.94.1
Recently we observed that Requests per second in HBase increased to 2 to 3 
fold.We have 15 regionServers & 20 data nodes in our cluster.

Thanks,Sandeep.

> CC: [email protected]
> From: [email protected]
> Subject: Re: Frequent scan time outs in multiple region servers in same time 
> intervals
> Date: Thu, 14 Nov 2013 06:43:17 -0800
> To: [email protected]
> 
> Which version of HBase are you using ?
> 
> If this problem only started to appear recently, was there noticeable change 
> in load on the cluster ?
> 
> Thanks
> 
> On Nov 14, 2013, at 3:08 AM, Sandeep L <[email protected]> wrote:
> 
> > Its not a Scan for Map/Reduce.
> > Internally we are using HBase for one of our application where our 
> > application scans data from HBase tables and we use the scan results.Due to 
> > regular scan time outs our service getting interrupted.
> > 
> > 
> > Thanks,Sandeep.
> > 
> >> Date: Thu, 14 Nov 2013 18:54:28 +0800
> >> Subject: Re: Frequent scan time outs in multiple region servers in same 
> >> time intervals
> >> From: [email protected]
> >> To: [email protected]
> >> 
> >> Is it a Scan for Map/Reduce? Or it's only a regular scan in which case
> >> maybe you could consider to set the start and end position of your scan to
> >> limit the operation.
> >> 
> >> Thanks
> >> Ramon
> >> 
> >> 
> >> On Thu, Nov 14, 2013 at 6:23 PM, Sandeep L 
> >> <[email protected]>wrote:
> >> 
> >>> Hi,
> >>> In recent times we are seeing frequent scan time outs in multiple region
> >>> servers of our production cluster. Due to this our HBase cluster is not
> >>> responding for 5 to 10 minutes for any queries & that causing huge problem
> >>> for us.
> >>> We are getting following error in multiple region servers almost in same
> >>> time intervals
> >>> 2013-11-14 00:16:30,178 INFO
> >>> org.apache.hadoop.hbase.regionserver.HRegionServer: Scanner
> >>> -8978491571406751432 lease expired on region
> >>> TableName,news,1383881528749.1e3f228e7a7a5a13d117af309f252c6c
> >>> This error is occurring on multiple tables.
> >>> Can suggest us what could be the cause and how to resolve the issue.
> >>> 
> >>> Thanks,Sandeep.
> >                         
                                          

Reply via email to