On Fri, Feb 4, 2011 at 3:45 PM, Alain.Moulle <[email protected]> wrote: > Hi, > > I've to test a new option of fence_ipmilan added by RH , which is "diag" > , meaning that the > fence will send via the IPMI a DIAG signal to the node to be fenced so > that it enters a dump > phase. > > With fence_ipmilan executed manually with -o diag , it works fine. > > In pacemaker, I've set action="diag" instead of action="reboot" like this : > > primitive restofencenode1 stonith:*fence_ipmilan* \ > params ipaddr="<my ipaddr>" login="<mylogin>" passwd="<mypasswd>" > pcmk_host_check="static-list" pcmk_host_list="node1" *action="diag"* \
The action is set globally based the stonith-action cluster property (which allows the values: reboot and poweroff). I can see that it might be useful to allow per-device settings. If you file an upstream bug I'll get to it. > > but it does always a reboot when fencing the node1 , and I can see in > syslog : > tengine_stonith_notify: Peer chili1 was terminated (*reboot*) by chili0 > for chili0 (ref=ead5a817-034e-4559-9064-8288d2ce11f5): OK > > thus metadata seems ok : > <parameter name="action" unique="1"> > <getopt mixed="-o" /> > <content type="string" default="reboot"/> > <shortdesc lang="en">Operation to perform. Valid operations: on, > off, reboot, status, list, diag, monitor or metadata</shortdesc> > </parameter> > > Did I miss something to change so that it really send a *diag* instead > of *reboot* ? > > Thanks for your response. > Regards > Alain > > > _______________________________________________ > 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
