hbase-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andrew Purtell <apurt...@apache.org>
Subject Re: Hbase 0.19 failed to start: exceeds the limit of concurrent xcievers 3000
Date Thu, 29 Jan 2009 19:24:08 GMT
Hi Genady,

Based on information you have provided in the past, in my opinion
your cluster is too small and load on HDFS is insufficiently
spread. The fact that you exceed xceiver count even when it is 
set at 3000 is a good indication of this. 

Also I believe your cluster became considerably slower after you
upped the xceiver count in the first place, correct? I wonder if
the additional resources consumed by that configuration change
is causing the working set of some of the JVMs to swap.

   - Andy

> On Thu, Jan 29, 2009 at 4:59 AM, Genady wrote:
> Regarding the last questions: actually there are
> updates, but eventually it take to MR to finish about 1.30
> hour(instead of average 15 minutes) or it fails as result
> of timeouts.


View raw message