Yesterday was a *VERY* long day!  Started with a 4 a.m. idle-system
backup, and then started update from 4.0 to 5.0.
 
At the end of the SLT update, during autoboot, we got DCC Startup
errors pointing to a problem with NMCONFIG file.  Yes, I *DID*
cross-validate, several times during the past week, under 4.0,
and never got an error.
 
After a couple of time-consuming RSes, with many more similar
errors, we were able to force another cross validate (using NMMGR
that comes with 5.0 this time), and got warning messages to the
effect that NM STACK and NM HEAP parameters were set above the
maximum and had been dynamically reduced to the maximum.  So
why does 5.0 have different maximums?  Would have appreciated a
forewarning about this--took an extra 2 hours out of what was
already a very *TIGHT* window.  We run a public hospital, not a
sweets factory.
 
Just to make matters worse--we dropped NS/3000 off our contract
(as well as some other things), and the plain-vanilla 5.0 LAN
LINK software wouldn't let me restart our network.  Had the
perfectly *AWFUL* feeling that HP had left something off our
SUBSYS tape--but it turned out to be some old files left over
from the full NS/3000 product.
 
The 5.0 installation/update/add-on manual is missing a crucial
start norecovery instruction--only of concern to a neophyte,
since it's plainly obvious to me that one has to do one at the
ISL prompt after a RS.  More serious is the lack of instruction
to run NMMGRVER and/or stream JCONFJ (which failed anyway, due
to hang-over files, according to the NARC).
 
Outbound telnet works (after tweaking the SERVICES.NET and
RESLVCNF.NET files)--nice.  I think there are problems with
SQL client access--but I'm too tired (went to the opera immediately
after the update--got home at 1 a.m.--22 hours after getting up)
to investigate.  SERVICES mentions the port, the relevant services
are up and running.  They can wait.
 
Ron
[log in to unmask]