SUMMARY Re: TruCluster V1.6 member startup problems

From: Narendra Raavi <narendra_at_spiff.mt.att.com>
Date: Sun, 02 Dec 2001 10:13:14 -0500

Managers,

Original question is below. First thanks to Christian Wessely.
He suggested that we may have forgotten to reconfigure asemgr and that we
may have a bad /etc/CCM.

Fortunately, (or unfortunately) we do NOT create the cluster configuration
map because

 (a) we never use cmon (I suspect only cmon uses /etc/CCM)
 (b) we knew we had to change to a different network
 (c) we were told that CCM causes problems between the agent and cmon

The real problem turned out to be that the asecdb on the three servers went
out of sync. Same number of bytes, but showed different checksums. Copying
the asecdb from one of the systems to others solved the problem.

On further investigation and testing, we found that this problem was
created because, initially when these systems were booted up, the switches
they were connected to had spanning tree protocol enabled.

After disabling the spanning tree protocol, and copying the asecdb from one
system, we were back up.

-- 
 -- Narendra Ra[a]vi     Email : narendra_at_spiff.hr.att.com
    MT  A4-4B27       Phone : 732-420-7792
On Wed, Nov 28, 2001 at 02:55:09PM -0500, Narendra Raavi wrote:
> 
> Managers,
> 
> We have a 3 node ES40 cluster in a 2+1 configuration running V4.0F PK#4 and
> TCRV1.6 PK#4. We have 4 services on this cluster. The clustering is over
> the (2) NICs.
> 
> Today we had to move to a new network. We shutdown all services, moved to
> single user mode, changed IP addresses in /etc/rc.config, modified
> /etc/routes, updated /etc/hosts and updated /etc/resolv.conf. We connected
> to the new network switches and rebooted the systems, all systems came up
> fine with full network connectivity. However, we have problems with ASE.
> 
> When I boot up ONE member of the cluster, then all ase processes come up
> fine. In fact, the ase service on that member also comes up fine. However
> when I boot the other member, I see the following in the daemon.log and
> ase does not respond. Anyone have similar problems?

Received on Sun Dec 02 2001 - 15:08:39 NZDT