"Schmidt, Florian" <[EMAIL PROTECTED]> writes:

> I already thought of this.
>
> I know that this script (drbddisk) is recommended by the DRBD-developers.
>
> But testing v2-style cluste configuration I always used the ofc-script drbd 
> and it always worked.
> So I do not know why it fails now....it would be great if anyone could 
> explain this.
>
> I'll reconfigure my CIB now to use the drbddisk-script
>
> Where exactly are the advantages/disadvantages of the ocf-script drbd against 
> the drbddisk-script?
>
> Thanks
> Florian
>
> PS: During testing I think I found some inconsistency in naming of the states 
> of master/slave resources: In my opinion they should know the states 
> "stopped", "master" and "slave". For slave there is another name "started" 
> but during my testing both of these states appeared in crm_mon. It would be 
> fine if you could choose one (started or slave) and make this consistent, 
> because for me this really means the same. If I'm wrong please correct me. :)

I also have this issue.

And i also think about using drbddisk rather than the ocf RA.

I have a similar setup, but i haven't put drbd resources in groups.
Also half the drbd resources are supposed to run master on node1 and
the other half master on node 2.
Unfortunatly for a reason i don't understand yet, it doesn't really work
properly. (for instance some resources don't get a slave)

Maybe you could try without groups.
iirc hb_gui refuses to put master/slave primitive in a group.

-- 
_______________________________________________
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