How can I verify a system disk?

[Heinz W Genhart]

Hi Paul

after changing the Cluster Group and Password, it's required to have the whole cluster down.
You will not be able to boot the second node, as long the first node is not also down at same time. (all nodes must be down at same time)

Regards

Heinz

[Paul Beaudoin]

Heinz,

OK - I should have read your original note a bit better - yeou did tell me that but I didn't hear it :-)

I'll (holding my breath) try that tonight. Comes back to the original question - As near as I can tell this sys disk is OK but would love to prove that before reboot. The real problem here is that I can't boot off CD anymore (hardware prob) and if I run into problems it gets a bit complicated.

Many thanks - I'll let you know.

Paul

[Paul Beaudoin]

Heinz,

Partial success. As part of other things, I installed the PCSI and upodate003 patches and rebooted. Bugchecked on the first try very early in the sequence but came up fine on the second try. All seems well. The sattelite still shows the same result though - it gets its first file(s?) and then starts requesting cluster membership endlessly. I was a bit rushed so didn't get the boot screen up and see what it was saying so I'll have to try that over the weekend to get the complete picture.

Thanks for your help (and patience)

Paul

[Paul Beaudoin]

Gents,

I'm really stuck now. Back to the original problem: Sattelite finds initial files then requests cluster membership. Boot node appears to accept (received Request followed by cluster tansition completed) and then sattelite requests again repeating this cycle. I have now: changed the cluster pass and rebooted, checked VAXCLUSTER param (2) and licensing - all OK. I have also booted the sattelite standalone (it has a sys disk) and that is OK. I searched but could not find a basic guide (check list) to cluster building. This has to be something simple and so obvious that I can't see it. Straw grabbing: was there any licence requirement changes in 8.2? Any suggestions?

Thanks

Paul

[Paul Beaudoin]

Gents,
Just to close this out properly - I have found the problem. The new root created by cluster_config.com included setting the sysgen param expected_votes to 2 in the new root. Once changed to its proper value of 1 , it booted. This is a vms/procedure error. The vote param was set (correctly) to 0 and from the local box the correct value of 1 could and should have been calculated. This was a brand new root creation and I did not even check, never mind change any params as until now, it was a safe assumption that all were correct.
Many thanks to all, especially Heinz who helped me through this.

Paul

[Paul Beaudoin]

See previous post.