[Linux-HA] Three Node Cluster Problem

[Dejan Muhamedagic]

you can't have more than two nodes in the v1 style cluster. either
reduce the number of machines or use crm.

On Fri, Nov 03, 2006 at 11:22:40AM +0100, Mc Linux wrote:
> Hi!
> 
> I've got the following configuration:
> 
> 3 home PCs (hb1,hb2,hb3), with one ethernet. (This cluster is for
> testing purposes)
> Installed OS: SuSE Linux Enterprise Server 9.0
> Heartbeat: 2.0.7 rpm package made with "ConfigureMe package"
> 
> same ha.cf on each maschine:
> 
> keepalive 2
> deadtime 30
> warntime 10
> initdead 120
> 
> udpport 694
> mcast eth0 225.0.0.1 694 1 0
> auto_failback off
> 
> node hb1
> node hb2
> node hb3
> 
> same haresources on each maschine:
> 
> hb1 192.168.70.155 saslauthd
> 
> I've just use saslauthd for testing.
> 
> same authkeys on each maschine:
> 
> auth 3
> 
> 3 md5 hello
> 
> When I start the cluster software on the maschines every maschine can
> see the other two. Logs from hb1:
> 
> heartbeat[10390]: 2006/10/30_16:41:21 info: Local status now set to: 'up'
> heartbeat[10390]: 2006/10/30_16:41:22 info: Link hb2:eth0 up.
> heartbeat[10390]: 2006/10/30_16:41:22 info: Status update for node
> hb2: status up
> harc[10397]:    2006/10/30_16:41:22 info: Running /etc/ha.d/rc.d/status 
> status
> heartbeat[10390]: 2006/10/30_16:41:23 info: Link hb3:eth0 up.
> heartbeat[10390]: 2006/10/30_16:41:23 info: Status update for node
> hb3: status up
> 
> If I turn off hb1 the other two maschines relocate the service in the same 
> time!
> 
> Ha2 logs:
> 
> info: Link hb1:eth0 dead.
> info: Running /etc/ha.d/rc.d/status status
> info: No local resources [/usr/lib/heartbeat/ResourceManager listkeys
> hb2] to acquire.
> info: Taking over resource group 192.168.70.155
> info: Acquiring resource group: hb1 192.168.70.155 saslauthd
> INFO: IPaddr Resource is stopped
> info: Running /etc/ha.d/resource.d/IPaddr 192.168.70.155 start
> INFO: eval /sbin/ifconfig eth0:0 192.168.70.155 netmask 255.255.255.0
> broadcast 192.168.70.255
> INFO: Sending Gratuitous Arp for 192.168.70.155 on eth0:0 [eth0]
> 
> 
> Ha3 logs:
> 
> WARN: node hb1: is dead
> WARN: No STONITH device configured.
> WARN: Shared disks are not protected.
> info: Resources being acquired from hb1.
> info: Link hb1:eth0 dead.
> info: Running /etc/ha.d/rc.d/status status
> info: No local resources [/usr/lib/heartbeat/ResourceManager listkeys
> hb3] to acquire.
> info: Taking over resource group 192.168.70.155
> nfo: Acquiring resource group: hb1 192.168.70.155 saslauthd
> INFO: IPaddr Resource is stopped
> info: Running /etc/ha.d/resource.d/IPaddr 192.168.70.155 start
> INFO: eval /sbin/ifconfig eth0:0 192.168.70.155 netmask 255.255.255.0
> broadcast 192.168.70.255
> INFO: Sending Gratuitous Arp for 192.168.70.155 on eth0:0 [eth0]
> 
> And after a few seconds they realize that both of them got the resources:
> 
> ERROR: Both machines own our resources!
> 
> What did I wrong?
> I think I missed something from the ha.cf, but I don't know what. I've
> got dozens of working two-node clusters.
> 
> Please help me!
> 
> Thanks,
> 
> Mr McLinux
> _______________________________________________
> Linux-HA mailing list
> Linux-HA at lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems