In Jan we purchased two large 2100 4.0b machines in tru-cluster and ase
mode to give greater resilience to our internet service. As agreed with
the private vendor, we had to immediately change the host names and ip
addresses etc and reconfigure dns and so on. We did this using CDE and
were successful in operating both systems in stand alone mode with the
built-in private disks. However, clearly none of the changes propagated
to the cluster and ase info and we have spent two months now trying to get
ase back on line without success (the shared disks are externally mounted
sw racks).
If anybody has had to do this and has guidelines then of course we would
be delighted to hear from you.
Eventually we had to reinitialise ase from scratch and rerun
cluster_map_create to get a new /etc/CCM. This has given us a valid
cluster and clu_ivp shows no errors. However we cannot get the services
back on line. It looks as though the problem is our not understanding how
to reconfigure the lsm volumes with new host names if this is necessary,
as a precursor to getting advfs filesets to be recognised on these disks.
We would like to recover the original structure if possible but had to
junk the asecdb when we reinited ase as otherwise it would not start at
all. I should say that lsm starts up the private disks ok and the advfs on
root and usr.
What am I looking for please? Any ideas on how to get out of the mess a
I am in?
As a separate question but assuredly related, we cannot start the cluster
monitor even though clu_ivp is passed. The error given is "CCM has not
been initialised OK". I thought a gui might help! Both the prime and
secondary hosts are in the xhost list and the rest of CDE works fine.
I can mail the CCM file or any other info if somebody is able to help.
Thankyou.
Stuart McKenzie
sysop MANNET
Received on Wed Mar 10 1999 - 15:04:18 NZDT