On Thu, Feb 7, 2013 at 8:22 PM, Marek Grac <mg...@redhat.com> wrote: > On 02/07/2013 07:07 AM, Andrew Beekhof wrote: >> >> This is right, fence agents (all of them) can be used only with one port >> value. >> Would this be a good feature to add? >> Particularly for people using "reboot" as the action. >> >> Otherwise the caller (ie. me :-) ) has to split it up into stop 1, >> stop 2, start 1, start 2 - which is messy. > > It is possible to add this feature to most of the fence agent as it should > be part of common fencing library. After brief investigation, it can be > fairly simple change and I'm not completely against it. But we would have to > better define what should be output of fence agent when entering multiple > plug values. > > * action = status: (single) return code 0 (ON); return code 2 (OFF) -> > (multiple) ?
I suspect we want rc=0 if any of the ports are on. > * action = off/on: (single) fail if operation was not completed -> > (multiple) fail if any of these operation was not completed yep > * action = reboot: (single) fail if OFF fails -> (multiple) fail if any OFF > fails yep > Some fence agents can have 'reboot' implemented without OFF/ON procedure > (direct usage of command 'reboot') for these multiple plug should be > forbidden. Could we not fall back to off off on on ? _______________________________________________ Pacemaker mailing list: Pacemaker@oss.clusterlabs.org http://oss.clusterlabs.org/mailman/listinfo/pacemaker Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org