Hi,
On Thu, May 20, 2010 at 05:26:34PM +0100, Alexander Fisher wrote:
> Hi.
>
> I'm setting up a two node corosync+pacemaker cluster. My servers each
> have dual PSUs.
> I've got them connected to a pair of APC networked PDUs. This gives
> me the option of using both external/ipmi
> and external/rackpdu.
>
> I've configured both and this seems to work fine. If I unplug the
> PDUs from the network, ipmi is used and likewise
> external/rackpdu is used if the servers' ipmi feature is disabled.
>
> But how do I specify the order in which they get tried?
Use the priority instance attribute. It's listed with every meta
data of stonith plugins:
priority (integer, [0]):
The priority of the stonith resource. The lower the number,
the higher the priority.
> Which plugin do people think will be the most reliable?
Unfortunately, most power management devices are not really
robust. That being said, a fencing device failure would be a
second failure in the cluster. The probability that that happens
within a short period is really low. And if you have two devices,
you should be OK. However, as discussed in another thread, a
fencing device reporting success in case the node hasn't been
reset is no good. That's not a failure, but a bug. And lies are
definitely out.
Thanks,
Dejan
> Thanks,
> Alex
> _______________________________________________
> Linux-HA mailing list
> [email protected]
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
