[Pacemaker] Error when managing network with ping/pingd.

[Francis SOUYRI]

Hello,

Thank for your reply Andreas, I set the multiplier to 1000, cut the 
netword nothing append.

I have these score:

Resource           Score     Node           Stickiness #Fail
IPaddr2_dhcpd      1000      noeud2.apec.fr 100        0
IPaddr2_dhcpd      1350      noeud1.apec.fr 100        0
IPaddr2_named      1000      noeud2.apec.fr 100        0
IPaddr2_named      1350      noeud1.apec.fr 100        0
IPaddr2_named2     1000      noeud1.apec.fr 100        0
IPaddr2_named2     1350      noeud2.apec.fr 100        0
IPaddr2_samba      1000      noeud1.apec.fr 100        0
IPaddr2_samba      1350      noeud2.apec.fr 100        0
ping-gateway:0     100       noeud2.apec.fr 100        0
ping-gateway:0     -INFINITY noeud1.apec.fr 100        0
ping-gateway:1     0         noeud2.apec.fr 100        0
ping-gateway:1     100       noeud1.apec.fr 100        0

Best regards.

Francis

On 09/17/2013 10:21 AM, Andreas Kurz wrote:
> On 2013-09-17 09:45, Francis SOUYRI wrote:
>> Hello,
>>
>> Some help about my problem ?
>>
>> I have a corosync/pacemaker with 2 nodes and 2 nets by nodes,
>> 192.168.1.0/24 for cluster access, 10.1.1.0/24 for drbd in bond, both
>> used by corosync.
>> I try to used ocf:pacemaker:ping to monitor the 192.168.1.0/24 I have
>> the configuration below, but when I remove the cable of the noeud1 the
>> named group resource do not migrate to noeud2.
>
> Looks like the extra score 100 from pingd of node2 is not high enough to
> overrule the location constraints with score 50 and the
> resource-stickiness of 100 ... you can e.g. increase the "multiplier"
> value of your pingd resource to 1000 to be sure it overrules these
> constraints and the stickiness.
>
> Regards,
> Andreas
>