After we switched pool master to another host with command “xe pool-designate-new-master host-uuid=xxxx”, the CloudStack mgmt. server eventually connected to all hosts successfully with all hosts in Running state.
So I can only speculate that there is something wrong with the Xen Cluster itself, but I have no direct evidence to support this claim from Xen server logs. Once the master is switched over, I restarted xen toolstack on the old master node just to be safe. Yiping On 9/7/16, 7:15 PM, "Ezequiel Mc Govern" <ezequiel.mcgov...@gmail.com> wrote: We have the same problem,one Pool get in connecting or Alert state. The configuration is the same. > On Sep 7, 2016, at 17:50, Yiping Zhang <yzh...@marketo.com> wrote: > > Hi, > > In one of our xen clusters, all host except the pool master are stuck in alerting or connecting state. There are no scary log entries in catalina.out or management-server.log. > > I have tried to restart management service already, but it has no effect. What other steps can I take? I am thinking of restarting xen toolstack on the pool master (already tried on one of slave host, no effect), is this a safe operation to do with tons of VM running on all hosts and the pool master? > > Yiping > > PS: we are running CS 4.8.0 on RHLE 6.7 and Xen 6.5 with all recommended security patches installed.