RE: ES40's won't boot!

-----Original Message-----
From: John Santos [mailto:john.santos@xxxxxxxxxxxxxxxx]
Sent: July 28, 2007 4:25 PM
To: Info-VAX@xxxxxxxxxxxx
Subject: ES40's won't boot!

Long story made longer...

Customer bought a pair of used ES40's for in-cabinet upgrades of a
pair of AlphaServer 4100's. They're having scheduling
difficulties, so
they want to get the new equipment tested before the 30-day
warrantee
from the dealer expires next week.

They also got a pair of MSA1000's, about 40 72GB disks, a pair of
fiber switches and a pair of DS-KGPSA-DA fiber host interfaces for
each ES40, which they would also like to test.

They have an old, unused AlphaServer 4100 (VMS 7.3-2) that they
wanted to use as a testbed. (Production systems are also 7.3-2.)
Everything patched through UPDATE V11 (we could install UPDATE V12
if needed, but at another site they've been running with MSA's
and, until recently, much older patch levels, without any obvious
problems.)

I'm trying to diagnose this from 3000 miles away, and I'm on
vacation, supposedly, starting today.

Anyway, I walked the customer through configuring the MSA's,
temporarily, as a pair of 1.5TB RAID5 sets partitioned into
4 units of 750GB each. They installed 2 of the 4 HBA's in
the AS4100, which sees them, but doesn't see the MSAs and
produces an error on each FGA device at boot time. From
some HSG80 docs I found at HP, it looks like the KGPSA-DA's
aren't supported on AS4100's. Does anyone know for sure?
At the dead sergeant, the AS4100 says the HBA's have an
NVRAM formatting problem, and wwidmgr doesn't see any of
the units.


Reference:
http://tinyurl.com/3ac6fo (scroll down to known issue - NVRAM failure in V56 console and earlier)

The ES40's power up and look okay at the console, and when
they cabled the MSAs into one of them, ">>> show device"
and wwidmgr show all 4 of the units, so I think that part's okay.

But the ES40's (neither of them) will boot. They each
have a KZPCM (dual SCSI plus Ethernet) in them, and using
a spare external storageworks shelf, they've tried booting
a copy of the AS4100's system disk, the actual AS4100
system disk, and on the internal IDE CD drive, the
VMS V7.3-1 distribution CD (they couldn't find V7.3-2,
didn't look very hard), and the V6.x or V7.x (not sure what
version they were trying, but something plenty recent enough
to include ES40's) of the Alpha firmware CD.

All 8 cases say something like "transferring to bootstrap"
and then just hang. (Waited up to 20 minutes.) It sounds
like they think they are successfully loading the boot
from the disk (APB.EXE?) and then are getting stuck. They
have to press the halt button to get the >>> back.

So what stupid misconfiguration could be causing this? Or
are both ES40's identically broken?

(Every console variable I could think of seems to be okay.
console is set to "serial". os_type is "OpenVMS". Boot
flags, etc. seem to be correct. Memory seems to be okay
(passes power-on self-test.)

About 2 weeks ago, they upgraded another ES40 at another
site (in-box upgrade of CPUs from 667MHz to 833MHz), added
a new pair of MSA1000's to an existing pair, and had some
weird CPU problems before they actually started, and HP
had to reload the SROM from a floppy (fail-safe loader?)
System had been up for many months, rebooted fine, but
when HP tried a power-down and reboot, it broke. (This was
before they touched anything.) The customer had HP do that
upgrade; I have no idea why they decided to do this one
themselves. Anyway, having the experienced HP guys onsite
to do it dodged a major disaster in that case! Is this
a common ES40 problem? Do we need to do the same thing
to both the new ones?


Could there be something (backplane jumper, etc.) putting
them in "test mode" that just needs to be cleared or removed
or reset?


Some dumb configuration problem with the console terminal
they are attempting to use (some kind of Wyse supposed
VT320 emulator that they have zillions of?) They did
report seeing some backwards question marks, which I would
interpret as a parity or framing or character size problem,
but it seems to be working okay in console mode. There
were some console variables that seemed to reflect terminal
characteristics, but they were 9600 baud, 8-bit, no parity,
no modem control, I.e. pretty standard.


If they could get one of the ES40's to boot, we could
run some disk bashing tests on the MSA's before the
warrantee expires. (Apparently their current H/W
maint. vendor, not HP, won't put the new hardware under
their maintenance agreement until it is demonstrated to
be installed and working, and the clock is ticking on
the seller's warrantee. Arggh!) (And why didn't they
do this two or three weeks ago, before my vacation??)


I'll try to check in over the next week, supposedly the
vacation house (holiday house as my Aussie niece calls
it) has WiFi...


--
John

John,

Re: supported options .. good link:
http://h18002.www1.hp.com/alphaserver/products/options.html (click on system in question)
http://h18002.www1.hp.com/alphaserver/options/ases40/ases40_options.html (ES40 - beware wrap)

ES40 supported options:
http://tinyurl.com/2mrcat

Note - I always like to keep it simple when troubleshooting.

Make sure you have V7.3-2 CDrom as I seem to recall an issue with V7.3-1 CDrom booting on some servers. Don't bother trying to troubleshoot the SAN stuff until local CDrom booting works. Also, one option is to remove SAN adapters to make sure they are not causing some local issue as well (especially if not sure if the older adapters are supported).

What graphics card is installed?

Note - there was also an older ES40 HW compatibility fco - suggest get CS to verify if this is installed or not.

Regards


Kerry Main
Senior Consultant
HP Services Canada
Voice: 613-592-4660
Fax: 613-591-4477
kerryDOTmainAThpDOTcom
(remove the DOT's and AT)

OpenVMS - the secure, multi-site OS that just works.






.

Loading