On 10.11.2009 12:37, Krzysztof Mazurek wrote:
[email protected] — zapisano w dniu 2009-11-10 12:11:45:
> Finnur Örn Guðmundsson <[email protected]>
> Wysłane przez: [email protected]
>
[....]
>
> Every thing is fine - I can relocate service, reboot nodes, use web
> console - all toghether a lot of fun.
> Up to the moment I want to test the Failover function.
> I just simply reset/turn off the node working the service is
> working. I expect the Cluster to try localte node, the as it is not
> available, Cluster should run the service on the other working node.
> Instead I receive never ending timeouts and no FailOver actions is
performed!
>
[...]
> Hi,
>
> You are using manual fencing, that's why nothing really happens, the
> cluster is waiting for a manual ack before doing anythings.
>
> See: http://sources.redhat.com/cluster/wiki/FAQ/Fencing#fence_manual2
Ok! You found the problem I had with manual fencing - I didn't know
about manual confirmation - which in fact is the 'essence' of manual
fencing ("fence_ack_manual"). Thanks for wiki page link.
It works now - services were moved to the working node.
And I can see that there are no easy way of having HA Cluster running
withoud hardware fencing.
Krzysztof
_______________________________________________
rhelv5-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/rhelv5-list
Sadly, no....RHCS really needs someway of fencing mechanism....if you
are using a server with a onboard BMC you can probably use IPMI, or buy
DRAC card (Dell), iLO2 (HP) or RSAII (IBM).
Hope this helps.
Bgrds,
Finnur
_______________________________________________
rhelv5-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/rhelv5-list
