This kind of issue was discussed a couple of times on this mailing list. Basically, you can play with hbase.client.pause and hbase.client.retries.number but you won't find this satisfactory, which is why we opened https://issues.apache.org/jira/browse/HBASE-2445
J-D On Tue, Jul 13, 2010 at 10:38 AM, Justin Cohen <justin.co...@teamaol.com> wrote: > Is there a way to configure hbase.client call timeouts? If we have a > network outage or if hbase/zk goes down for some reason, we want our > table.puts to timeout reasonably (5-15 seconds) so we can queue them up for > later. We also want scans and gets to timeout so we can fail gracefully. > I've played with some of the retries and pause configs but I can't seem to > get a consistent timeout. It's often over 30 seconds, and occasionally it > never times out. > > Also, when hbase does come back, there seems to be a good 30-60 second delay > before we can contact region servers. > > Right now I'm testing this using hbase in standalone mode, bring up and down > stop-hbase.sh and start-hbase.sh > > Any tips? > Justin >