[Users] Host status non-responsding

2012-04-20 Thread Rahul Upadhyaya
Hi,

While performing some operations (vdsm restart and other operations to
debug SSL settings on the host) , my host goes to non-responding state.
Now it has happened for the third time. For the previous two times I did a
engine-cleanup and did a engine-setup to resolve this issue. Is there some
other way ... m guessing removing some host entry from database or
something to recover faster. On a non-responding host I can't put it
to maintenance mode because it says it has running Vms on it... also I cant
remove it and add it again, so basically my host is in a state of limbo
where I cant perform any operation on it. Is there any other way to recover
from such condition other that preforming a engine cleanup and engine-setup
again ?

-- 
Regards,
Rahul
===
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users


Re: [Users] Host status non-responsding

2012-04-20 Thread Eli Mesika


- Original Message -
 From: Rahul Upadhyaya rak...@gmail.com
 To: users@ovirt.org
 Sent: Friday, April 20, 2012 4:03:02 PM
 Subject: [Users] Host status non-responsding
 
 
 
 
 Hi,
 
 
 While performing some operations (vdsm restart and other operations
 to debug SSL settings on the host) , my host goes to
 non-responding state. 
Well , if you stop vdsm service , your host is in non-responding state and 
that's OK
 Now it has happened for the third time. For
 the previous two times I did a engine-cleanup and did a engine-setup
 to resolve this issue. Is there some other way ... m guessing
If you can define Power Management on that hots, your host will be 
automatically rebooted in such a case 

 removing some host entry from database or something to recover
 faster. On a non-responding host I can't put it to maintenance mode
 because it says it has running Vms on it... also I cant remove it

So, it is the only running host in your system. If you define one more host and 
run it 
you can send a non-responding host to maintainance since HA VMs will migrate to 
the other host

 and add it again, so basically my host is in a state of limbo where
 I cant perform any operation on it. Is there any other way to
 recover from such condition other that preforming a engine cleanup
 and engine-setup again ?
 
 
 --
 Regards,
 Rahul
 ===
 
 ___
 Users mailing list
 Users@ovirt.org
 http://lists.ovirt.org/mailman/listinfo/users
 
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users


[Users] high availability via fencing

2012-04-20 Thread Ian Levesque
Hello,

I'm testing ovirt for potential deployment and one of the metrics for its 
success relies on the high availability feature. In my research on this 
feature, I found scattered documentation indicating that fencing is a 
prerequisite. On my test hardware, I don't have any LOM/IPMI but I see that APC 
managed PDUs are supported, which I do have.

The problem is when I try to configure Power Management to use the apc type, 
I get this 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

ovirt-engine/engine.log tells me:

2012-04-20 11:32:44,595 INFO  [org.ovirt.engine.core.bll.FencingExecutor] 
(http--0.0.0.0-8443-5) Executing Status Power Management command, Proxy 
Host:heilig, Agent:apc, Target Host:, Management IP:134.174.x.x, User:ovirt, 
Options:port=22,secure=true,slot=1

2012-04-20 11:32:44,598 INFO  
[org.ovirt.engine.core.vdsbroker.vdsbroker.FenceVdsVDSCommand] 
(http--0.0.0.0-8443-5) 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,secure=true,slot=1'), log id: 57f86a56

2012-04-20 11:32:44,696 INFO  
[org.ovirt.engine.core.vdsbroker.vdsbroker.FenceVdsVDSCommand] 
(http--0.0.0.0-8443-5) FINISH, FenceVdsVDSCommand, return: 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

I've tried to add port=1 to the Options field but that seems to have no 
effect.

Any ideas? Is there any way to configure a dumb power management / fencing 
configuration for testing?

Cheers,
Ian
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users