[Linux-HA] Resource appears to be active on two nodes

[Dejan Muhamedagic]

Hi,

On Wed, Dec 01, 2010 at 09:35:15AM +0000, bharat khandelwal wrote:
> hi, 
> I have to do this without using pacemaker as pacemaker is last option for me to 
> achieve this task.
>  So, i need help with this version of heartbeat only without pacemaker.
>  As i am architecture follows:
>  Heartbeat version : 2.0.5

There's nobody using this version anymore for a very long time.
Please upgrade.

Thanks,

Dejan

>  No of nodes       : Three nodes A,B,C
>  Type of cluster   : Active/Passive/Passive
>  
> Desired cluster behaviour is.
>  A then B then C (as B and C are backup servers)
> 
>  it does'nt matter if previous active node is up or down it should move ahead as 
> per the node list in crm_mon and no failback required.
> I am trying this with default setting of cluster as in my cib i am having this 
> configuration:
> 
> default_resource_stickiness --- INFINITY (TO STOP FAILBACK)
> 
> three nodes A,B,C
> 
> one resource ocf ipaddr (for floating ip address)
> 
> and no constraint.
> 
> With this configuration i am trying to achieve high availibility but getting 
> these errors.
> 
>      On node A: 
> info: stop_all_resources:lrm.c Making sure all active resources are stopped 
> before exit
> crmd[7411]: 2010/12/01_14:29:02 ERROR: lrm_get_all_rscs(596): failed to receive 
> a reply message of getall.
> 
>      On node B:
> 
> ERROR: native_add_running:native.c Resource ocf::IPaddr:ip_1 appears to be 
> active on 2 nodes.
> 
> As such resulting in resource failback on node A which is not desired,
> please tell how to avoid this problem 
> 
> Thanks 
> Bharat
> 
> 
> _______________________________________________
> 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