From:	MERC::"uunet!CRVAX.SRI.COM!RELAY-INFO-VAX" 20-APR-1992 20:51:41.67
To:	INFO-VAX@SRI.COM
CC:	
Subj:	Re: Clusters at VMS 5.5???

We upgraded our LAVC to VMS 5.5 a few weeks ago and it went quite smoothly.
I attribute some of that to the helps and hints I collected from INFO-VAX before
we upgraded. Shoot - I had a whole list of "helpful hints from INFO-VAX" that
I had collected and was going to post and now I can't find it. I'm sure others
will share their warnings.

The only issues and problems we had off the top of my head:

1. Your password history file will be deleted - save it before hand if you
   want to continue it.

2. We had some problems with our license databases and I'm still not sure
   just what happened; on a few nodes they appeared to "be overwritten".
   I still doubt that that actually happened; but I had made copies of those
   elsewhere as well (of LMF$LICENSE.LDB) and was glad I did. I just threw the
   old ones in place and quit asking questions.  :-\

3. We ended up with a "corrupted" IMAGELIB.OLB, but I can't say that this is
   attributed to the update. We had problems with linking (image date mismatch)
   and inserting/replacing/deleting modules. It *may* have been corrupted
   beforehand and we just ran into problems when VMSINSTAL tried to update it.
   I doubt it is a 5.5 problem and was our own local problem, but I bring it
   up just in case....I rebuilt it with all our current images and we appear
   to be fine.

4. A few had posted that that answering "YES" to the question of, "Do you wish
   to upgrade to the new JOB_CONTROL" while installing the
   mandatory update would cause the installation to fail. I took their word
   for it and answered NO which leaves you at VMS A5.5 and did the standalone
   queue upgrade a few minutes later per the documentation. Went smoothly.

5. Things look a little different when you run LATCP. Not a problem it just
   suprised me!

We have only one system disk so I'm not sure what problems might be encountered
with a rolling upgrade; but the documentation does state that you can't use the
new queuing system unless all nodes in the cluster are at 5.5 with multiple
system disks.

Thanks for posting hints as you go, people. It really helped me out.

My only curiosity was our IMAGELIB problem. If someone else encountered a
problem please post for the benefit of future sites. If no one else mentions
a problem with it, I'd say don't worry about it!

Peggy Mogush
Taylor University
pgmogush@tayloru.bitnet