On Wednesday 01 December 2010 23:13:23 Frank-Matthias Kuschel wrote:
> On 12/01/2010 06:57 PM, Michael Schwartzkopff wrote:
> > On Wednesday 01 December 2010 17:21:21 Frank-Matthias Kuschel wrote:
> >> Hi Folks,
> >> 
> >> i have a problem using stonith with suicide device.
> >> First my use case:
> >> 2 SLES 11 nodes in active active configuration using clones.
> >> The main approach is to load balance http requests to a application
> >> server. There is a virtual ip adress which is used by both nodes.
> >> 
> >> I've try to keep the services alive and if any service didn't start the
> >> node should reboot for the first time.
> >> 
> >> Now here comes the problem.
> >> I see in the logs that a request for fence (reboot) are rejected with
> >> the message
> >> info: process_remote_stonith_query: Ignoring reply from srvhfosamno1,
> >> hosts are not permitted to commit suicide
> > 
> > suicide is not a good idea in production environments. I don't know if
> > they work at all, since I read somewhere that nodes cannot commit
> > suicide.
> 
> Maybe you right with this point. Let me ask a alternative way.
> It is possible to move a resource (ipaddr2) when a cloned service like
> apache failed (and actually restarting is in progress)?
> 
> My main problem is that a cloned ipaddr2 work correctly, but when i use
> something like
> order test inf: nfs_clone jboss_clone apache_clone ip, to handle the
> correct start and shutdown order,
> then the half of the clients can work correctly and the other is waiting
> for restarting is successfully finished.

Set migration-threshold=2 for apache.

-- 
Dr. Michael Schwartzkopff
Guardinistr. 63
81375 München

Tel: (0163) 172 50 98

Attachment: signature.asc
Description: This is a digitally signed message part.

_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to