Thank you for your reply. My requirement is like this: The httpd service runs on every node in the cluster and is monitored by watchdog. VIP only runs on one node at a time. Heartbeat will check the status of httpd on each node and make sure the VIP runs on the node that has httpd running. Note that the Heartbeat is not expected to control the httpd service but just to monitor.
Say I have Heartbeat, then I have the following configuration questions: 1) Should I use clone for monitoring httpd? 2) Which operation should I specified for the action of httpd service? "fence" or "block" or another? Is the combination clone+action+colocation enough for the requirement above? If not, what else special configuration do I need? Thank you for any advices! Hongyu On Tue, May 24, 2011 at 12:48 AM, RaSca <[email protected]> wrote: > Il giorno Gio 19 Mag 2011 19:25:54 CET, 吴鸿宇 ha scritto: > > Hi All, >> I have a 2 node cluster. My intention is ensuring the VIP is always on the >> node that has httpd running, i.e. if service httpd on the VIP node is >> stopped and fails to start, the VIP should switch to the other node. >> With the configuration below, I observed that when httpd stops and fails >> to >> start, the VIP is stopped also but is not switched to the other node that >> has healthy httpd. I appreciate any ideas. >> > [...] > > Some questions: > Why httpd is cloned? Are you sure you want an INFINITY stickiness? Are logs > saying anything helpful? > > Anyway, like Nikita said, consider upgrading Heartbeat to version 3. > > -- > RaSca > Mia Mamma Usa Linux: Niente è impossibile da capire, se lo spieghi bene! > [email protected] > http://www.miamammausalinux.org > > _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
