On Nov 23, 2016, at 5:49 AM, Jan Pokorný <jpoko...@redhat.com> wrote: > > On 18/11/16 08:22 +0100, Ulrich Windl wrote: >>> 1) is there a way to set up a "kill script", such that before trying to >>> launch a new copy of a process, pacemaker will run this script, which would >>> be responsible for making sure that there are no other instances of the >>> process running? >>> >>> [...] >> >> I wonder: Wouldn't a monitor operation that reports the resource as >> running as long as the port is occupied resolve both issues? > > What if the port is occupied by something entirely different?
Not only that (which is an entirely valid concern), but in this particular instance the failure mode was such that the service was non-functional but was still occupying the port - so such a monitor instance would say it is running, when in fact it isn't - not really. Somehow the current monitor operation *did* catch the fact that it was non-functional, and at least tried to restart it, which is a good thing. The only problem was that the old instance was only *mostly* dead. > > Note also that on Linux at least, there's some time-to-dispose-socket > after respective process termination, so SO_REUSEADDR socket option > on the respective daemon's side is a potentional player in this game. > > -- > Jan (Poki) > _______________________________________________ > Users mailing list: Users@clusterlabs.org > http://clusterlabs.org/mailman/listinfo/users > > Project Home: http://www.clusterlabs.org > Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf > Bugs: http://bugs.clusterlabs.org _______________________________________________ Users mailing list: Users@clusterlabs.org http://clusterlabs.org/mailman/listinfo/users Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
