Valery, you should report a bug since the nullpointer shouldn't occur. The reason that you get to the null pointer seems to be that there is no host up in your env, to deploy the vm on. Is that correct?
On Tue, Sep 24, 2013 at 2:49 PM, Valery Ciareszka <[email protected]> wrote: > Hi all, > > I try to test HA on CS 4.2/KVM and I have java.lang.NullPointerException. > > ENV used: > CS 4.2(rev 2963) management on centos 6.4 > CS 4.2(rev 2963) agent on centos 6.4 (node1,2) > separate nfs server as primary/secondary storage > zone type: KVM > > scenario: > I create several VMs with HA-enabled offering. > Then I switch power on node2 via IPMI. > expected behaviour: > ha-enabled vms from node2 should start on node1 > real behaviour: > vms remain in stopped state, unassigned to any host, NullPointerException > in mgmt log: > > 2013-09-24 11:21:25,500 ERROR [cloud.ha.HighAvailabilityManagerImpl] > (HA-Worker-4:work-4) Terminating HAWork[4-HA-6-Running-Scheduled] > java.lang.NullPointerException > at > com.cloud.storage.VolumeManagerImpl.canVmRestartOnAnotherServer(VolumeManagerImpl.java:2641) > at > com.cloud.ha.HighAvailabilityManagerImpl.restart(HighAvailabilityManagerImpl.java:516) > at > com.cloud.ha.HighAvailabilityManagerImpl$WorkerThread.run(HighAvailabilityManagerImpl.java:831) > > Full log is here: http://pastebin.com/upnEA601 > > Any thoughts ? > > > > > -- > Regards, > Valery > > http://protocol.by/slayer
