Hi,

On Thu, Jan 29, 2009 at 04:06:47PM +0430, Avnika Nautiyal wrote:
> Problem: Heartbeat takeover doesn't happen when eth0 goes down
> OS: Red Hat Enterprise Linux ES release 4 (Nahant Update 6)
> Kernel: 2.6.9-67.ELsmp #1 SMP
> Heartbeat packages: heartbeat-2.0.4-1.el4.i386.rpm, 
> heartbeat-pils-2.0.4-1.el4.i386.rpm, heartbeat-stonith-2.0.4-1.el4.i386.rpm
> DRBD packages: drbd-0.7.25-1.i386.rpm, drbd-debuginfo-0.7.25-1.i386.rpm, 
> drbd-km-2.6.9_67.EL-0.7.25-1.i386.rpm, 
> drbd-km-2.6.9_67.ELsmp-0.7.25-1.i386.rpm
> 
> Primary Machine Configuration
> -------------------------------
> Attached primary-ha.cf, primary-haresources, primary-drbd.conf
> IP: 192.168.75.110
> 
> Secondary Machine Configuration
> -------------------------------
> Attached primary-ha.cf, primary-haresources, primary-drbd.conf
> IP: 192.168.75.151
> 
> Virtual IP: 192.168.75.152
> 
> All my scenarios worked fine, except this particular and most critical one.
> How to generate the scenario:
> When I bring down the eth0 interface of primary server, the
> secondary server does not take control of resources (Takeover
> doesn't happen). The interface can be brought down either by
> executing ifconfig eth0 down command or by plugging off the
> cable from eth0 port.

The cluster can't recover from split brain (complete loss of
communication) unless there is a fencing device (stonith). Looks
like you don't have one.

Thanks,

Dejan

> The logs on secondary server shows:
> Jan 29 15:46:01 swestd2 kernel: r8168: eth0: link down
> Jan 29 15:46:26 swestd2 heartbeat: [24615]: info: Link swestd1.domain:eth0 
> dead.
> Jan 29 15:46:26 swestd2 ipfail: [24639]: info: Link Status update: Link 
> swestd1.domain/eth0 now has status dead
> Jan 29 15:46:27 swestd2 ipfail: [24639]: info: We are dead. :<
> Jan 29 15:46:27 swestd2 ipfail: [24639]: info: Asking other side for ping 
> node count.
> Jan 29 15:46:28 swestd2 ipfail: [24639]: info: Ping node count is balanced.
> Jan 29 15:46:28 swestd2 ipfail: [24639]: info: No giveup timer to abort.
> 
> Please help this is very critical to me!!!
> Regards,
> Avnika







> _______________________________________________
> 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