scinstall error

I completed scinstall on the first node in the cluster and did a reconfiguration boot. But when I executed scinstall on the second node, I ran in to the following problem:

scrconf: RPC: Timed out

scinstall: Unable to retrieve a copy of the cluster config

scinstall: scinstall did NOT complete successfully!

The IP addresses, hostnames are correctly set in the /etc/hosts file. What could be wrong?

[428 byte] By [dwal_007] at [2007-11-26 10:29:51]
# 1
HiCheck the time different on both node....if any please set first and reinstall with fresh configuration to all.....:)GoodLUCKMohammed Tanvir
Tanvir@SCSA at 2007-7-7 2:35:56 > top of Java-index,Solaris Operating System,Solaris Essentials - General Technical Questions...
# 2

Another thing to check: Set domainname on both machines. Suggest putting it

in /etc/defaultdomain as well.

I know, you would say "I am not using NIS/NIS+, so why should this matter?". You

are right, it shouldn't matter, but there are reports that it does. Something to

try anyway.

Please do share your experience on the forum.

HTH,

-ashu

ashu15 at 2007-7-7 2:35:56 > top of Java-index,Solaris Operating System,Solaris Essentials - General Technical Questions...
# 3

Mohammed,

The system time on the second node that gave me the error was approximately 2 minutes behind that on the first node. So I took your advice and did a fresh installation of SOE, patches, and JES. And before running scinstall I configured NTP to synch system time on the cluster nodes with another server. I'm not sure if that is what fixed the problem because there were bunch of other things I did different from the first attempt (list below). How come you sounded so sure about the NTP configuration? If this is really the issue then the installation documents should make this a pre-requisite for cluster installation. As it stands now, NTP configuration is suggested to be a post-cluster installation procedure.

The things I did differently included:

- Switched the first node to second and vice versa

- The RAID (StorEdge 3320) I'm using, requires a patch (115663-03 according to the release notes) on Sun Fire V440 servers. The patch is obsoleted by the kernel patch 118558-33. But according to SC release notes the required patch level for Solaris 9 is 117949-25. This time I did not install 118558-33.

- Did not enable jumbo frames on /kernel/drv/ce.conf

The domain name was set up in the first attempt also so I'm not sure that has anything to do with it.

The 2 node cluster finally seems to have been setup correctly but setting it up has been one frustrating experience.

Thank you to all of you for responding to my queries in a timely fashion.

waitio at 2007-7-7 2:35:56 > top of Java-index,Solaris Operating System,Solaris Essentials - General Technical Questions...