Description of problem: I'm trying to configure power management in order to enable the high availability feature of ovirt. Without proper fencing, it appears from my research that ovirt doesn't auto-migrate VMs from failed nodes, which is a prerequisite for using ovirt at our site. From a host's power management interface, I enter the IP, user/pass, "apc" as type, the port and slot, then check "secure". When I click "Test", I receive an error: Test Failed, Host Status is: unknown. The fence-agent script reported the following error: Failed: You have to enter plug number Please use '-h' for usage High-availability failover therefore isn't attempted. Version-Release number of selected component (if applicable): ovirt-engine-3.1.0_0001-1.8.el6.x86_64 How reproducible: Steps to Reproduce: 1. add host in ovirt-engine web interface 2. in "power management", enable and fill in all fields for APC type 3. click test Actual results: Error: Test Failed, Host Status is: unknown. The fence-agent script reported the following error: Failed: You have to enter plug number Please use '-h' for usage Expected results: Host's power management effectively configured, thereby enabling ovirt's high-availability features. Additional info:
Here's the vdsm and engine log output when I click "Test": ------------------------------------------------------------ ==> ../ovirt-engine/engine.log <== 2012-04-25 12:18:50,302 INFO [org.ovirt.engine.core.bll.FencingExecutor] (http--0.0.0.0-8443-4) Executing <Status> Power Management command, Proxy Host:heilig, Agent:apc, Target Host:, Management IP:134.174.x.x, User:ovirt, Options:port=22,plug=1,secure=true,slot=1 2012-04-25 12:18:50,305 INFO [org.ovirt.engine.core.vdsbroker.vdsbroker.FenceVdsVDSCommand] (http--0.0.0.0-8443-4) START, FenceVdsVDSCommand(vdsId = 6de5e3fa-8a33-11e1-b3f9-003048c85226, targetVdsId = 60087c5e-8a3b-11e1-b15d-003048c85226, action = Status, ip = 134.174.x.x, port = , type = apc, user = ovirt, password = ******, options = 'port=22,plug=1,secure=true,slot=1'), log id: 56a240f9 ==> vdsm.log <== Thread-52872::DEBUG::2012-04-25 12:18:50,311::API::954::vds::(fenceNode) fenceNode(addr=134.174.x.x,port=,agent=apc,user=ovirt,passwd=XXXX,action=status,secure=,options=) Thread-52872::DEBUG::2012-04-25 12:18:50,348::API::980::vds::(fenceNode) rc 1 in agent=fence_apc ipaddr=134.174.x.x login=ovirt option=status passwd=XXXX out err Failed: You have to enter plug number Please use '-h' for usage Thread-52873::DEBUG::2012-04-25 12:18:50,356::API::954::vds::(fenceNode) fenceNode(addr=134.174.x.x,port=,agent=apc,user=ovirt,passwd=XXXX,action=status,secure=,options=) Thread-52873::DEBUG::2012-04-25 12:18:50,392::API::980::vds::(fenceNode) rc 1 in agent=fence_apc ipaddr=134.174.x.x login=ovirt option=status passwd=XXXX out err Failed: You have to enter plug number Please use '-h' for usage ------------------------------------------------------------ See the thread in ovirt-users here: http://lists.ovirt.org/pipermail/users/2012-April/001699.html
Please specify the version of fence-agents package that is on the host (the one with VDSM on it).
fence-agents-3.1.5-10.el6_2.2.x86_64
Any updates on this? I can't setup a highly-available cluster of VM hosts until this apparently minor bug is resolved or I hear of a suitable workaround.
this was resolved by Roy Golan (vdsm bug) probably requires vdsm upgrade Roy, please specify the version on which that was resolved
please check vdsm version and see if it is below vdsm-4.9.6-29.0 If it below vdsm-4.9.6-29.0 , please close the bug as this was laready resolved
BZ was resolved in vdsm-4.9.6-29.0 Please note that for verification you should only check the vdsm.log abd see that all parameters are passed correctly. Please also note that since our lab was upgraded to APC 5.x you will not be able to get it working (the workaround is to use apc_snmp agent) , however, as stated above only parameters passing verification is needed