Hi, I have a question about config parameter of VirtualDomain after following commit.
High: VirtualDomain: Ensure it is possible to manage a libvirt domain defined outside of VirtualDomain https://github.com/ClusterLabs/resource-agents/commit/f00dcaf194 I'm using RHEL6.4 and bundled libvirt. $ rpm -qf `which virsh` libvirt-client-0.10.2-18.el6.x86_64 For example, when VM is created by virt-manager, it is generated as persistent domain. * persistent domain : the state outputted by 'virsh list --all' even if VM has stopped. $ virsh list --all Id Name State ---------------------------------------------------- - vm1 shut off $ ls -l /etc/libvirt/qemu/vm1.xml -rw------- 1 root root 3000 Nov 22 12:53 /etc/libvirt/qemu/vm1.xml When making this vm1 manage by VirtualDomain, start operation will become an error if /etc/libvirt/qemu/vm1.xml is specified as config. Nov 28 17:10:33 bl460g1n6 crmd[7368]: notice: te_rsc_command: Initiating action 14: start prmVM1_start_0 on bl460g1n6 (local) Nov 28 17:10:33 bl460g1n6 crmd[7368]: info: do_lrm_rsc_op: Performing key=14:4:0:682fec50-0325-46e1-9a59-ec77ccc7277e op=prmVM1_start_0 Nov 28 17:10:33 bl460g1n6 lrmd[7365]: info: log_execute: executing - rsc:prmVM1 action:start call_id:47 Nov 28 17:10:33 bl460g1n6 VirtualDomain(prmVM1)[7727]: DEBUG: Virtual domain vm1 is currently shut off. Nov 28 17:10:33 bl460g1n6 VirtualDomain(prmVM1)[7727]: ERROR: Failed to start virtual domain vm1. Nov 28 17:10:33 bl460g1n6 lrmd[7365]: notice: operation_finished: prmVM1_start_0:7727:stderr [ error: Failed to open file '/etc/libvirt/qemu/vm1.xml': No such file or directory ] Nov 28 17:10:33 bl460g1n6 lrmd[7365]: info: log_finished: finished - rsc:prmVM1 action:start call_id:47 pid:7727 exit-code:1 exec-time:152ms queue-time:0ms Nov 28 17:10:33 bl460g1n6 crmd[7368]: notice: process_lrm_event: LRM operation prmVM1_start_0 (call=47, rc=1, cib-update=75, confirmed=true) unknown error Nov 28 17:10:33 bl460g1n6 crmd[7368]: warning: status_from_rc: Action 14 (prmVM1_start_0) on bl460g1n6 failed (target: 0 vs. rc: 1): Error Nov 28 17:10:33 bl460g1n6 crmd[7368]: warning: update_failcount: Updating failcount for prmVM1 on bl460g1n6 after failed start: rc=1 (update=INFINITY, time=1385626233) I think that there are the two methods of avoiding. A) Move /etc/libvirt/qemu/vm1.xml to another directory, and specify the directory as config. $ mv /etc/libvirt/qemu/vm1.xml /qemu/vm1.xml primitive prmVM1 ocf:heartbeat:VirtualDomain \ params config="/qemu/vm1.xml" \ * Also when I change configuration of vm1 by virt-manager, it's necessary to move /etc/libvirt/qemu/vm1.xml again. B) After undefine the vm1, move the vm1.xml which evacuated to /etc/libvirt/qemu/ and specify the /etc/libvirt/qemu/vm1.xml as config. $ mv /etc/libvirt/qemu/vm1.xml /tmp/vm1.xml $ virsh undefine vm1 $ mv /tmp/vm1.xml /etc/libvirt/qemu/vm1.xml primitive prmVM1 ocf:heartbeat:VirtualDomain \ params config="/etc/libvirt/qemu/vm1.xml" \ Which is the method assuming? Best Regards, Kazunori INOUE _______________________________________________________ Linux-HA-Dev: [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev Home Page: http://linux-ha.org/
