I have two rh el 6 systems configured with rhcs and clvmd. General cluster services seems to be ok. As I'm not able to successfully migrate a vm through clusvcadm, I'm now downsizing the problem to direct virsh command that fails when called from clusvcadm. The guest's storage is composed by two disks that are clustered logical volumes vm definition is <vm name="exorapr1" recovery="relocate" use_virsh="1" xmlfile="/etc/libvirt/qemu/exorapr1.xml"/> the xml file is the same at both hosts At first I verified correct startup on both nodes this way: - vm running on host2 with resource recovery policy set "relocate" - shutdown vm from inside its operating system - the cluster notices this and correctly restarts it on host1 - shutdown vm from inside its operating system - the cluster notices this and correctly restarts it on host1 I have also ssh equivalence in place (for the intracluster names) so that I can run from host2: [host2 ] # virsh -c qemu+ssh://intrarhev1/system list without need of password input. If I try the command used by the cluster itself (after stopping the vm from clusvcadm): # virsh migrate --live exorapr1 qemu+ssh://intrarhev1/system I receive: error: operation failed: Migration unexpectedly failed On host2: [host2 ] # virsh list Id Name State ---------------------------------- 3 exorapr1 running In messages: Mar 4 14:27:30 host2 libvirtd: 14:27:30.527: error : qemuDomainWaitForMigrationComplete:5394 : operation failed: Migration unexpectedly failed Setting this: [root at host2 libvirt]# export LIBVIRT_DEBUG=1 [root at host2 libvirt]# export LIBVIRT_LOG_OUTPUTS="1:file:/tmp/virsh.log" I get the file I'm going to attach (due to migration happening with intracluster network, the names are intrarhev1 and intrarhev2 on that LAN) It seems no more information in the file.... Any hints on further debugging? If there is not any big mistake at my side I could also open an official case, as these two systems are under subscription maintenance... Thanks in advance, Gianluca
Attachment:
virsh.log
Description: Binary data
-- Linux-cluster mailing list Linux-cluster@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/linux-cluster
