Funny this should come up since we're planning to do the exact same thing
very soon. HP advises that we do a FULL backup followed by creating a SLT.
Then, move the clock forward and test the heck out of yer software.
When rolling the date back to reality, HP advises that you shut the system
down, and either do an INSTALL or an UPDATE from your SLT followed by a
:RESTORE / (without the ;KEEP option). Apparently, some of the post-Y2K
dates are written in places that you & I don't understand and that a full
:RESTORE *should* (fingers crossed) correct the "time warp" factor.
By all means, please keep us posted of any problems you encounter when
rolling the date back and I'll do likewise. Insofar as an INSTALL is
concerned, I've been handed this solution many times in the past by HP
(disk drive mechs go out, disk controllers fail, etc.). I sometimes wonder
if HP is aware of the amount of time needed to re-INSTALL a 50 or 100 GB
system and if users can simply accept a day of lost production, paying an
operator or some other person(s) to watch the green light on a DAT flash
for 12 hours or so.
Best of luck!!
Tom Madigan
HP Systems Manager
International Communications Research, Inc.
605 West State St.
Media, PA 19063-2620
At 03:48 PM 11/11/99 -0800, Deane Bell wrote:
>We're planning to do an actual full-system/application test on
>an HP 3000/969 running MPEiX 6.0 PP1. The test will involve
>shutting down and restarting, setting the system clock ahead to
>Feb 28,2000 @ 23:30, noting how applications behave on
>Feb 29,2000. Because we are a 7x24 medical operation, down-time
>is very scarce. After test completion I was planning to purge
>system log files, restore databases, rebuild image log files,
>shutdown and restart with the system clock set to current time.
>Checking in with the RC they pointed me to their
>MPE Testing White Paper and EMPHASIZED that the entire SYSTEM
>must be reloaded from scratch, not simply shutdown and restarted.
>That'll take down-time that my user-community won't appreciate
>unless it's absolutely essential. I can't seem to get a straight
>answer from RC as to what their actual experience may have shown
>as problems if this is not done, ie why shouldn't I simply
>use SETCLOCK to roll back the clock or do the restart previously
>planned. It seems HPs "Y2K attorneys" have issued a "gag-order"
>wrt Y2K issues. Has anyone else done these kind of tests without
>a full RELOAD and what if anything untoward happened?
>Other words of wisdom on the topic? TIA.
>---------------------------------------------------------------------------
---
> Deane Bell Computer Services Coordinator, Affiliate Instructor
> The University of Washington Vmail : (206) 598-6009
> Medical Centers Pharmacy switchboard: (206) 598-6060
> 1959 N.E. Pacific Street, Box 356015 facsimile : (206) 598-4901
> Seattle, WA 98195-6015 U.S.A. internet : [log in to unmask]
>---------------------------------------------------------------------------
---
>
|