On Tue, Aug 19, 2008 at 13:29, Satomi Taniguchi <[EMAIL PROTECTED]> wrote: > > Hi Lars, > > Thank you for your reply! > > > Lars Marowsky-Bree wrote: > (snip) >> >> There is one missing bit though; a node not doing kdump needs to be >> STONITH'ed; so, failure of the kdump-stonith plugin should "escalate" to >> the next plugin. I'm not sure the current STONITH subsystem can handle >> this. > > I think it is possible by making a plugin which has ability to spoil and > a kdump-check plugin group. > And by making kdump-check plugin return ERROR when it detects that > the target node is not dumping. > As far as I confirm, if the first plugin in a group is failed, > the second one is executed. > (If it is an unexpected behavior, please let me know the correct one.) > > However, a timeout-problem comes here. > Under present STONITH, when a plugin takes long time, > parent-stonithd's timeout occurs, > and STONITH function starts again from the beginning > (from creating a graph). > This means that the behavior in the case of a plugin takes long time is > different from that in the case of it returns ERROR. > > >> >>> Second, I would like to hear your opinion about the following. >>> I think a timeout setting shuold be necessary for STONITH plugin. > > (snip) >> >> Yes, being able to somehow specify a per-plugin "fence" timeout would be >> useful. The "start" and "stop" timeouts can be set, but not the actual >> stonith ops ... >> >> I think it would make sense if they could be specified as regular >> operations in the CIB, and then would be passed to stonithd somewhow. >> > > Now I'm pondering exactly on it. > It is easy way to add fixed attributes (like "fence-timeout") > to each plugin's instance_attributes, and get it's value from params in > stonith_rsc_t... > But maybe it is more proper way to add new op "fence" and add an element > (like "child_timeout" or "fence_timeout" or something) > to stonith_ops_t, I think... > > In addition, maybe it is necessary to pass "fence_timeout" > to tengine too, not only to stonithd. > TE has its own timeout setting for STONITH, so > it is necessary to extend that depending on each plugin's > timeout setting,
No. This is not possible since the TE does not (should not, and can not) know which stonith device (or series of devices) is going to be used. > even if two or more plugins are set in a group. > > > > > Best Regards, > Satomi Taniguchi > > > _______________________________________________________ > Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.org > http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev > Home Page: http://linux-ha.org/ > _______________________________________________________ Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev Home Page: http://linux-ha.org/
