Hi,

On Fri, Jul 25, 2008 at 04:03:02PM +0530, nishant kumar wrote:
> Hi
> im using R1 type 2 node cluster. They are connected via one ethernet
> connection only ( i will remove that SPOF between two nodes later ).
> Here after removing ethernet cable from node 1 (who was holding resorce at
> that time ) , cluster of-course goes in partition and node two takes over
> resources too. But IPfail of node 1 should have seen that since my ping to
> the ping node "192.168.10.1" has died, and should have released its
> resources ? this doesnt happen becase HBComm between two nodes fail and they
> cant communicate ? If yes then this policy should be prevailing that since
> "I have lost communication to the ping node i should release resources" ?
>  What changes do i need to do in my below config to have such policy, even
> if its incorrect?

ipfail won't do anything if it can talk to the other node. That
is a v1 option and v1 is not developed anymore. I'm afraid that
you're out of luck.

Thanks,

Dejan

> 
> 
> NODE 1 Configuration:
> 
> ------------------ha.cf---------------
> respawn hacluster /usr/lib/heartbeat/ipfail
> logfile /var/log/ha-log
> keepalive 2
> deadtime 10
> warntime 4
> initdead 10
> udpport 694
> node slave master
> ucast neoeth0 192.168.10.46
> auto_failback off
> ping 192.168.10.1
> 
> ----------------haresources---------------
> slave IPaddr2::192.168.10.99/24/neoeth0
> 
> Node 2 Config:
> 
> ----------ha.cf--------------------
> respawn hacluster /usr/lib/heartbeat/ipfail
> logfile /var/log/ha-log
> keepalive 2
> deadtime 10
> warntime 4
> initdead 10
> udpport 694
> node slave master
> ucast neoeth0 192.168.10.156
> auto_failback off
> ping 192.168.10.1
> 
> 
> thanks
> nish
> _______________________________________________
> Linux-HA mailing list
> [email protected]
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to