Thanks for the replies!
Russ Fish
Alan Rollow
Harold Gabel
Summary:
I did the actual update on Saturday - 5/15. I definitely did the vdump 0 of / and /usr. The root
directory actually I tar'ed over to one of the "untouchable" Oracle partitions as well.
I followed all the update instructions I could find, using the last Installation Guide I had - 4.0B, as well
as the 4.0D Release Notes. When I ran "verify -r root_domain", It did say that it found an errored
directory tag, and deleted(repaired) it, although when I ran it again, I got the same error.
I decided to trust that it was indeed fixed. I checked to make sure that the listed layered products
that would halt the update were not installed - none were. Ran '/sbin/installupdate /dev/rz6c' and it
barfed because the step to mount /usr was missing from the Installation Guide ! I actually 'mount -u /' as
well. The installupdate then took off, warned about some layered products that "may" have to be
reinstalled afterwards, but none actually did.
That was basically it. The update didn't seem to affect anything at all functionally. I did notice that the
sysconfigtab is now HUGE, where once it was actually only an empty skeleton of a file, to which I
had added a few configs manually over the past year, it's now several pages on top of my now
relatively miniscule entries.
One note: I haven't investigated this yet, but a "defragcron" entry was added to root's crontab. I got
several repetitious runs of cron jobs or at least the mailings that follow at the end of these crons were
repeated. I have yet to check it out, but it was curious. I hadn't run defragcron ever before, and haven't
found the MAN page for it yet.
BTW - The Oracle data/partitions were untouched.
***********************
Wayne Sweatt
Principal Software Analyst
Litton / PRC
505.827.9288
***********************
-----Original Message-----
From: Wayne Sweatt [SMTP:sweatt_at_dps.state.nm.us]
Sent: Friday, May 14, 1999 6:38 PM
To: 'DU Mail List'
Subject: installupdate to 4.0d question
A quick question:
I've actually got to do my first REAL update from 4.0B to 4.0D. I have notes here from past responses about problems. I will for instance run verify on my advfs domains and vdump level 0 my / and /usr directories. I'll watch out for the change in UID 12, which is our Oracle user ID and put it back when done, etc,etc. here's a simpler question:
I want to physically isolate my RAID 5 set(s) of Oracle RDBMS data. I'm sure that the update won't touch those directories, but I would rather be prudent about this. I used to unplug non-OS drives when doing massive upgrades on SunOS and Solaris systems, but I have not dealt with one on DU/Tru64 UNIX. I don't want to upset the ol' SWXCR 200 either. so....
Should I just change my fstab (copy and then change) to only list / and /usr ? Should I completely power down everything, including BA-350 cabinet, et al and unplug the RAID 5 member drives, leaving only the / and /usr ones (RAID 1) ?
What should I do? NO! I don't want to, and shouldn't have to vdump all 25GB of RAID 5/Oracle stuff. It took me 6-8 hours last time!! Bogus.
It seems the physical unplugging is definitely the most foolproof, but I don't want to tork off the SWXCR.
Please Advise...
Thanks!
P.S. We're talking 2100 AlphaServer 4/200 W/ RAID Array 200 V2 (I think - I know I upgraded from 1.0)
***********************
Wayne Sweatt
Principal Software Analyst
Litton / PRC
505.827.9288
***********************
Received on Mon May 17 1999 - 16:43:33 NZST