[Pacemaker] very urgent

[Andrew Beekhof]

On Feb 16, 2009, at 11:53 AM, Glory Smith wrote:

>
>
> having an unreliable stonith mechanism is worse than not having one  
> at all.
>
> what if your resource fencing has a bug? its the same problem.
>
> reliable fencing is a fundamental requirement of the cluster.
>
> i understand . but If i can do both (node as well as resource  
> fencing) then there is very very rare chance that both will fail at  
> same time.

You can have multiple stonith devices.  If one fails, we just try the  
next.

>
> i will again say it would be really good if something can be done to  
> implement resource fencing in openais-pacemaker. dont you agree with  
> this.

I get the feeling that by "resource fencing", you just mean scsi  
reservations which are already possible in the current framework.


> 2) no, it cant.
> it wont have quorum and therefor isn't allowed to start cluster  
> resources
>
> if i set no-quourm to stop (Which is default)  then i feel following  
> things will happen
>
> 1) resouces will be stopped at all nodes who belong to  no-quourm  
> partition and fail over to a nodes which have quorum.
>
> 2) All nodes of no-quorum partion will be stonithed by  nodes having  
> quorum.
>
> am i right?

correct.  although 2) usually happens a lot faster than 1).

>
> if openais -pacemaker provide something  for resouce fencing we  
> would have completely ruled out  above possiblities  Please share  
> your view.
>
>
>
> _______________________________________________
> Pacemaker mailing list
> Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> _______________________________________________
> Pacemaker mailing list
> Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.clusterlabs.org/pipermail/pacemaker/attachments/20090216/a050b886/attachment-0002.htm>