[Pacemaker] Fixed! - Re: Problem with dual-PDU fencing node with redundant PSUs

[Andrew Beekhof]

On 30/06/2013, at 4:48 AM, Lars Marowsky-Bree <lmb at suse.com> wrote:

> On 2013-06-29T09:22:20, Andrew Beekhof <andrew at beekhof.net> wrote:
> 
>>> This doesn't help people who have dual power rails/PDUs for power
>>> redundancy.
>> I'm yet to be convinced that having two PDUs is helping those people in the first place.
>> If it were actually useful, I suspect more than two/three people would have asked for it in the last decade.
> 
> I admit that after thinking about it some more, I do see the appeal for
> certain configurations.
> 
> Like Digimer wrote, you do need a double failure for this two-layer
> fencing mechanism to fail (node burned + one grid down).
> 
> While I personally prefer the self-fence mechanisms, I think this does
> make sense for other environments where multiple redundant power
> supplies are desired.
> 
> And apparently, this is one of the scenarios for which fence topology
> was created and supports multiple devices per level. I'd venture the
> opinion that the current implementation of "multiple devices per level"
> is broken (since it requires work-arounds like digimer posted).

No, it was designed for things like: 
   http://clusterlabs.org/doc/en-US/Pacemaker/1.1-pcs/html/Pacemaker_Explained/_advanced_fencing_configurations.html

> 
> I don't have a very strong opinion, but if this is something RHEL
> customers had before, I'd suggest you don't want to force that complex
> configuration on them.

I don't know what all the location constraints were doing there, but otherwise its no less complex than what they were configuring before.

https://access.redhat.com/site/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/Cluster_Administration/s2-dualpower-fence-config-conga-CA.html

Setting action= appears to be "normal" in RHCS... I guess thats why they're trying to do it in Pacemaker too.

> And on you to support. ;-)

Far easier to support this than extra code to do it automagically.