Hi, On Fri, Feb 17, 2012 at 11:16 AM, MengHuan Chiang <a00477@xxxxxxxxx> wrote: > Hi Dan > Thank you for your response > > 2012/2/16 Dan Frincu <df.cluster@xxxxxxxxx> >> >> Hi, >> >> On Thu, Feb 16, 2012 at 10:34 AM, MengHuan Chiang <a00477@xxxxxxxxx> >> wrote: >> > Dear all: >> > I have been using pacemaker-1.0.10 and corosync-1.4.2 to constuct two HA >> > nodes >> > But when I turn down one of the nodes' network interface (eth0) and >> > restart >> > it. >> >> I'd recommend you also take a look at >> http://www.corosync.org/doku.php?id=faq:ifdown >> >> Might be related to your issue. > > Thanks for that advice. >> >> >> > The corosync service would dead but pid file exists >> > and if I check the processes status (ps -eaf | grep "heartbeat") >> > there are 2 processes left >> > >> > root 4998 1 0 13:59 ? 00:00:00 >> > /usr/lib64/heartbeat/lrmd >> > 101 5000 1 0 13:59 ? 00:00:00 >> > /usr/lib64/heartbeat/pengine >> > >> > Can anyone give me a hint about what would cause this situation? >> > >> > _______________________________________________ >> > discuss mailing list >> > discuss@xxxxxxxxxxxx >> > http://lists.corosync.org/mailman/listinfo/discuss >> > >> >> >> >> -- >> Dan Frincu >> CCNA, RHCE > > I have another question about if corosync would handle the network restart > situation after the interface has been manually shutdown. > Thanks Short answer: no. Long answer: - if you ifdown an interface corosync is running on, you end up with it gone and pacemaker processes running at 100% - if you bring back corosync, the pacemaker processes don't notice this - if you try to stop gracefully pacemaker processes, not all of them listen, so at least the cib has to be killed forcefully - starting pacemaker processes after corosync brings back the cluster to a working state This would be a situation without fencing configured/enabled. HTH, Dan > -- Dan Frincu CCNA, RHCE _______________________________________________ discuss mailing list discuss@xxxxxxxxxxxx http://lists.corosync.org/mailman/listinfo/discuss