upgrading from 4.1 to 6.1

pbaustin · ‎02-02-2009

Hi All, I'm upgarding from 4.1 to 6.1 but several of the servers are out on oil rigs so i was thinking of doing the upgrade for these servers in the lab and then sending the hard disks to site to be installed in the existing servers and then running the bios update. Question - would this work provising the existing hardware supports ver 6.X?

Thanks

Tommer Catlin · ‎02-02-2009

I'd probably test that in the lab first also before shipping it out to a remote site. In theory it sounds fine. If you have exact same server models in the lab, it should be pretty simple then. If they are slightly differnt, the drives may not boot up.

Jonathan Schulenberg · ‎02-02-2009

You'll need to be exact with the firmware versions when matching them up manually.

Also remember that UCM is keyed to NIC1's MAC address. You'll need to get each license re-issued when you put the drives in the final server.

JeffG1 · ‎02-02-2009

Are these "Subscriber" nodes or all publishers?

If subscribers, how about an IP based KVM along with the Install DVD?

If publishers, you need a place to store the DMA data...

pbaustin · ‎02-02-2009

Both, because they are on rigs in the north sea, and the min round trip delay, I have 5 clusters - pub and sub in each cluster. I think I'll have to do these on-site.

Thanks

Tommer Catlin · ‎02-03-2009

Prepare for odd things to happen after you are done. If it's a simple cluster (no AC, no MOH, etc) you should be good to go. We have had issues where the 4.1 just never would update. If there are VG224 or VG248s involved, you will need to REMOVE these first, then do the TAR file. Things like this can cause an upgrade to take forever! Keep spare drives around. Basically, pull your 4.1 drives, insert new drives, then install 6.x. At least you can rollback by inserting the old drives quickly if the TAR file does not upgrade correctly.

Jonathan Schulenberg · ‎02-03-2009

I'll echo tcatlinins here: always have spare drives on hand!

UCM will not install unless the RAID array is healthy and drives have an amazing ability to sense when an upgrade is about to occur, and then fail right before you start. I've had it happen several times in just the last few months.