cancel
Showing results for 
Search instead for 
Did you mean: 

Problems booting a GS160

Telia BackOffice
Valued Contributor

Problems booting a GS160

Hi

I have a problem booting a GS160. To make a long story short:

the machine consisted of 2 QBBs each with a I/O drawer. I reconfigured it to use only 1 QBB due to a problem with the HS-LINK between the QBBs.

The server will boot once in a while, but not always. It is really annoying.

Output from a boot sequence

P00>>>show device
dgb2110.1001.0.4.1 $1$DGA2110 EMC SYMMETRIX 5670
dgb2110.1002.0.4.1 $1$DGA2110 EMC SYMMETRIX 5670
dgc2110.1003.0.7.1 $1$DGA2110 EMC SYMMETRIX 5670
dgc2110.1004.0.7.1 $1$DGA2110 EMC SYMMETRIX 5670
dgd2110.1001.0.1.2 $1$DGA2110 EMC SYMMETRIX 5670
dgd2110.1002.0.1.2 $1$DGA2110 EMC SYMMETRIX 5670
dkb0.0.0.2.0 DKB0 COMPAQ BD0096349A 3B0A
dqa0.0.0.15.0 DQA0 COMPAQ CDR-8435 0013
pga0.0.0.3.0 PGA0 WWN 1000-0000-c923-b9de
pgb0.0.0.4.1 PGB0 WWN 1000-0000-c924-5013
pgc0.0.0.7.1 PGC0 WWN 1000-0000-c92c-d775
pgd0.0.0.1.2 PGD0 WWN 1000-0000-c928-64d9
pka0.7.0.1.0 PKA0 SCSI Bus ID 7 5.57
pkb0.7.0.2.0 PKB0 SCSI Bus ID 7 5.57
P00>>>boot -fi genvmunix


OpenVMS PALcode V1.98-2, Tru64 UNIX PALcode V1.92-1

system = QBB 0 (Hard Partition 0)
QBB 0 = CPU 0 1 2 3 + Mem 0 1 2 3 + Dir + IOP + PCA 0 1 + GP (Hard QBB 0)
micro firmware version is V6.4
shared RAM version is 1.6
hose 0 has a standard I/O module
starting console on CPU 0
initialized idle PCB
initializing semaphores
initializing heap
initial heap 340c0
memory low limit = 220000 heap = 340c0, 1ffc0
initializing driver structures
initializing idle process PID
XDELTA not enabled
initializing file system
initializing timer data structures
lowering IPL
CPU 0 speed is 731 MHz
create dead_eater
create poll
create timer
create powerup
access NVRAM
QBB 0 memory, 8 GB
total memory, 8 GB
probe I/O subsystem
probing hose 0, PCI
probing PCI-to-ISA bridge, bus 1
bus 0, slot 1 -- pka -- QLogic ISP10x0
bus 0, slot 2 -- pkb -- QLogic ISP10x0
bus 0, slot 3 -- pga -- KGPSA-C
bus 0, slot 15 -- dqa -- Acer Labs M1543C IDE
probing hose 1, PCI
bus 0, slot 4 -- pgb -- KGPSA-C
bus 0, slot 7 -- pgc -- KGPSA-C
probing hose 2, PCI


* * * Illegal Configuration with Memory Channel 2 [MC2] in Hose 2
* * * * Only MC2 devices can share same PCI Bus with MC2 Rev. 22

bus 0, slot 1 -- pgd -- KGPSA-C
bus 0, slot 3 -- mca -- DEC PCI MC
probing hose 3, PCI
starting drivers
entering idle loop
starting console on CPU 1
initialized idle PCB
initializing idle process PID
lowering IPL
CPU 1 speed is 731 MHz
create powerup
starting console on CPU 2
initialized idle PCB
initializing idle process PID
lowering IPL
CPU 2 speed is 731 MHz
create powerup
starting console on CPU 3
initialized idle PCB
initializing idle process PID
lowering IPL
CPU 3 speed is 731 MHz
create powerup
initializing GCT/FRU at 220000
Initializing pka pkb dqa pga pgb pgc pgd
Testing the System
Testing the Disks (read only)
AlphaServer Console V6.5-605, built on Jun 5 2003 at 15:17:34

CPU 0 booting

(boot dgb2110.1001.0.4.1 -file genvmunix -flags a)
failed to open dgb2110.1001.0.4.1
(boot dgb2110.1002.0.4.1 -file genvmunix -flags a)
failed to open dgb2110.1002.0.4.1
(boot dgc2110.1003.0.7.1 -file genvmunix -flags a)
block 0 of dgc2110.1003.0.7.1 is a valid boot block
reading 19 blocks from dgc2110.1003.0.7.1
bootstrap code read in
base = 400000, image_start = 0, image_bytes = 2600(9728)
initializing HWRPB at 2000
initializing page table at 1fffd0000
initializing machine state
setting affinity to the primary CPU
jumping to bootstrap code

cb_open: failed SCSI3 1 7 0 3 0 0 0 @wwid1 3d2b00 0000000000000800, dgc2110.1003.0.7.1

halted CPU 0

halt code = 5
HALT instruction executed
PC = 20000030
P00>>>

Could one of you wizards tell me how to debug/proceed?
8 REPLIES
Vladimir Fabecic
Honored Contributor

Re: Problems booting a GS160

I would start with checking FC connections:

cb_open: failed SCSI3 1 7 0 3 0 0 0 @wwid1 3d2b00 0000000000000800, dgc2110.1003.0.7.1
In vino veritas, in VMS cluster
Telia BackOffice
Valued Contributor

Re: Problems booting a GS160

yeah. But can I narrow it down between

bad fiber

or

problems logging into the switch?

It works "once in a while". The server booted on the 6th attempt this time.
Vladimir Fabecic
Honored Contributor

Re: Problems booting a GS160

It can be several things.
First check logs on FC switches.
You can also check some things with WWIDMGR.

>>> init
>>> wwidmgr -show wwid -full
>>> wwidmgr -show reachability
In vino veritas, in VMS cluster
Rob Leadbeater
Honored Contributor

Re: Problems booting a GS160

Hi,

> AlphaServer Console V6.5-605

Start off by upgrading your firmware. You're quite a few revisions behind... Latest is 7.3.

http://h18002.www1.hp.com/alphaserver/firmware/readmes/gs320.html

I've not had experience booting from SAN on a Symmetrix, but with the latest XP/HDS models, you have to be on 7.3 to support boot from SAN...

Hope this helps,

Regards,

Rob
cnb
Honored Contributor

Re: Problems booting a GS160

Hi,

This looks suspicious:

* * * Illegal Configuration with Memory Channel 2 [MC2] in Hose 2
* * * * Only MC2 devices can share same PCI Bus with MC2 Rev. 22

Do you have mixed or unsupported Memory Channel devices?

This *usually* indicates that either MC1 or MC1.5 are mixed with MC2 on the same PCI bus OR an unsupported MC2 hardware revision.


Since Fibre (pgd) is also on PCI 2 and it's complaining about an illegal configuration on PCI 2, why not pull the mc card(s) out and see if it stabilizes?


hth,
Vladimir Fabecic
Honored Contributor

Re: Problems booting a GS160

cnb said:
"Since Fibre (pgd) is also on PCI 2 and it's complaining about an illegal configuration on PCI 2, why not pull the mc card(s) out and see if it stabilizes?"

This may help, but also may not.
Anyway, I would do it.
Was this server a cluster member?
Is it still a cluster member or standalone?
In vino veritas, in VMS cluster
Telia BackOffice
Valued Contributor

Re: Problems booting a GS160

It is a cluster member, but do not use the memory channel anymore, as far as I remember.

It is official day off in Denmark today, but I'll investigate the memorychannel and bad fiber problems on monday.
Rob Leadbeater
Honored Contributor

Re: Problems booting a GS160

Hi,

You might also want to check whether pga0 is plugged in... If all the fibre HBAs go to off to the Symmetrix, I'd have thought you'd see a couple of dga2110... devices in "show dev".

Cheers,

Rob