Hide Forgot
Description of problem: fence-virt currently uses QMFv1. It needs to be updated to QMFv2
Patches posted to cluster-devel mailing list: https://www.redhat.com/archives/cluster-devel/2011-August/msg00015.html https://www.redhat.com/archives/cluster-devel/2011-August/msg00016.html
Patches look good; do we want to update the config file stuff (I'll do it), or will it just work as-is ?
Nevermind, looks like the config is compatible so we'll leave it alone.
http://fence-virt.git.sourceforge.net/git/gitweb.cgi?p=fence-virt/fence-virt;a=commit;h=e4658c1ae237dd8b0987ce0060d2a996a8990f49 http://fence-virt.git.sourceforge.net/git/gitweb.cgi?p=fence-virt/fence-virt;a=commit;h=6dbcc695052d0035a0da5671f37f24210cdb5b2b http://fence-virt.git.sourceforge.net/git/gitweb.cgi?p=fence-virt/fence-virt;a=commit;h=a2aff2aeb8815a8cf190623af3d54e39717f356c
Technical note added. If any revisions are required, please edit the "Technical Notes" field accordingly. All revisions will be proofread by the Engineering Content Services team. New Contents: Do not document.
Created attachment 533656 [details] Configuration file for fence_virtd I used the attached configuration when attempting to run. Without specifically setting the host and/or port, fence_virtd crashed, which will be a bug we need to address later. I started qpidd with the following options: --auth no -m yes I started libvirt-qmf with the following options: --broker localhost -p 5672 -d Starting fence_virtd in the foreground allowed me to watch the following occur: ipv4_recv_sk: success, fd = 5 Request 2 seqno 955657 domain f16-1 Plain TCP request ipv4_connect: Connecting to client ipv4_connect: Success; fd = 6 tcp_response: no-op (AUTH_NONE) tcp_challenge: no-op (AUTH_NONE) Request 2 seqno 955657 src 127.0.0.1 target f16-1 [libvirt-qpid] REBOOT operation on f16-1 [libvirt-qpid] OFF operation on f16-1 f16-1 running 2011-11-14 19:51:12 warning Connection [127.0.0.1:59733-127.0.0.1:5672] closed [libvirt-qpid] ON operation on f16-1 f16-1 shutoff 2011-11-14 19:51:15 warning Connection [127.0.0.1:59734-127.0.0.1:5672] closed Sending response to caller... On the caller side: [root@zealand lhh]# fence_xvm -o reboot -H f16-1 Could not read /etc/cluster/fence_xvm.key; trying without authentication [root@zealand lhh]# echo $? 0 On the VM side, I observed the virtual machine move to the powered-off state then restart itself.
I tracked down the crash to a port not being specified; it is possible a simple port change will cause things to work as expected (though it is likely libvirt-qpid.cpp should handle the exception being thrown by the QMF API) https://bugzilla.redhat.com/show_bug.cgi?id=753974
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. http://rhn.redhat.com/errata/RHBA-2011-1566.html