Committed as 3010 and 3011 Regards, Honza
Fabio M. Di Nitto wrote: > Looks ok to me (untested). > > Fabio > > On 7/28/2010 4:08 PM, Jan Friesse wrote: >> Attached is better implementation of SUBJ patch. It also modifies init >> script file to NOT create pid file. >> >> Second patch is for better integration with cman. If corosync was >> started by cman, corosync refuses to stop. >> >> Regards, >> Honza >> >> >> Fabio M. Di Nitto wrote: >>> On 7/22/2010 7:56 PM, Steven Dake wrote: >>>> On 07/22/2010 10:25 AM, Steven Dake wrote: >>>>> On 07/22/2010 08:49 AM, Jan Friesse wrote: >>>>>> Patch uses flock to ensure that only one instance of corosync is running. >>>>>> >>>>>> Regards, >>>>>> Honza >>>>>> >>>>>> >>>>>> >>>>>> _______________________________________________ >>>>>> Openais mailing list >>>>>> [email protected] >>>>>> https://lists.linux-foundation.org/mailman/listinfo/openais >>>>> g >>>>> reat work! good for merge >>>>> >>>> Honza, >>>> >>>> After talking with Fabio, he mentioned that the proper place for this >>>> file is /var/run/corosync.pid, and the contents of this file should be >>>> the active process ID of the corosync child. >>>> >>>> Sorry for not getting more details earlier. >>> My bad, it did pass under my radar. >>> >>> Honza, the implementation you did is ok, i also suggest to look into >>> dm_create_lockfile(const char*) implementation that´s been recently >>> written to be "strong". >>> >>> Fabio > _______________________________________________ Openais mailing list [email protected] https://lists.linux-foundation.org/mailman/listinfo/openais
