1. 2014-04-29 10:32:28,876 INFO org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation: This client just lost it's session with ZooKeeper, will automatically reconnect when needed.
What about ZK metrics, may be it's timing out or something (I mean echo srvr |nc host port)? -Mikhail 2014-04-29 14:52 GMT-07:00 Stack <[email protected]>: > Anything in the zk log when the RS connects JMS? > > > On Tue, Apr 29, 2014 at 7:37 AM, Jean-Marc Spaggiari < > [email protected]> wrote: > > > Hi, > > > > I face a "strange" issue today. > > > > I stopped my cluster to merge some regions. When I tried to restart it, > one > > server got stuck. > > > > If I try to stop the cluster, this server ignore the request and stay > > stuck. > > > > Then if I restart, the cluster come back without this server saying that > > it's already started. > > > > The server seems to be stuck with ZK. http://pastebin.com/6UcYEKN2 > > > > I have killed the RS and tried to restart it, same issue. I'm able to > > access the ZK server from the RS. I have restarted the 3 ZK server too, > > same result. The RS tries to connect to all of them but always fail. > > > > Any idea why ZK might faile that way? > > > > Thanks, > > > > JM > > > -- Thanks, Michael Antonov
