We have rebooted both at the same time. They do come up as expected with the primary holding the virtual IP. I did bang my head on the learning curve and each node had its own name in the haresources so they each grabbed the IP. That is not the case now.
It performs as I expect it to perform but I want to get better control over that performance. Tim On Thu, 2010-08-12 at 10:56 -0500, Igor Chudov wrote: > Tim, just curious if you tried to reboot both A and B at EXACT same time. > > i > > On Thu, Aug 12, 2010 at 10:54 AM, Tim Pickard <tpick...@crossref.org> wrote: > > Hi all, > > > > I have a pair of machines I am going to deploy as a proxy/smtp/dns > > server. > > > > These are RHEL 5.5 with heartbeat 2.1.4-11.el5. > > > > They are set up with machine A as the primary for virtual IP x.x.x.29 in > > the haresources. "machinea.domain.com x.x.x.29" > > > > In simple testing, machine A network is unplugged and machine B then > > takes on the IP x.x.x.29 after the deadtime passes (currently set to > > "10"). > > While disconnected, machine A will have the IP x.x.x.29 in a NOCARRIER > > state. > > When the machine A is reconnected, machine B will immediately surrender > > the IP x.x.x.29. Machine A ALSO surrenders the IP upon connection as it > > seems to detect the machine B has the IP when it reconnects. > > Machine A will then take the IP x.x.x.29 after the deadtime passes. > > > > Is there a way to configure a gentler or negotiated hand-off of the IP > > when the primary returns to service? auto_failback does not seem to be > > exactly what I need and is deprecated for ver. 3 (when we get there). > > > > Thanks, > > Tim > > > > > > My ha.cf: > > > > keepalive 2 > > deadtime 10 > > udpport 694 > > bcast bond0 > > mcast bond0 225.0.0.1 694 1 0 > > ucast bond0 machineBIP > > udp bond0 > > node machinea.domain.com > > node machineb.domain.com > > logfile /var/log/heartbeat.log > > > > _______________________________________________ > > Linux-HA mailing list > > Linux-HA@lists.linux-ha.org > > http://lists.linux-ha.org/mailman/listinfo/linux-ha > > See also: http://linux-ha.org/ReportingProblems > > > _______________________________________________ > Linux-HA mailing list > Linux-HA@lists.linux-ha.org > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems _______________________________________________ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems