AlphaServer 4100 disk problems

[John Appleby_1]

Hi,

I apologise if this is the wrong forum for this message - there doesn't seem to be an AlphaServer forum unfortunately.

I have just purchased an AlphaServer 4100 in a rack - the specs are as follows:

1x AlphaServer 4100 Rack
2x 21164-400-4Mb
2x 256Mb EDO
LSI Logic RAID with 2x Power, 6x4Gb SCSI disks
HSZ50 RAID with 2x(2x Power 6x4Gb SCSI disks)

After a bit of fiddling around (some blown power modules and someone put the CPUs in the wrong slots) I have gotten it to post to VGA, but none of the SCSI disks show up.

I booted the SCSI and RAID enclosures before turning on the unit and the lights on the disks go off one by one. Is this supposed to happen?

I'm not an expert with this unit though I have some experience of Alpha workstations; I'd expect the disks/containers to come up when you do show dev.

If I boot into AlphaBIOS it says that there are no disks installed. I have checked all the cables and they seem to be fine.

Any pointers would be much appreciated.

Regards,

John

[G. Vrijhoeven]

Hi John,

I found a document that might help you:

http://ftp.digital.com/pub/Digital/Alpha/systems/as4100/docs/4100_ug.pdf

link

http://www.google.nl/search?q=cache:G_0EdyfvSn8J:ftp.digital.com/pub/Digital/Alpha/systems/as4100/docs/4100_ug.pdf+AlphaServer+4100+disk+config&hl=nl&ie=UTF-8

Its section 6.2.?

HTH,

Gideon

PS i had a hanging IE when i opened it.

[Michael Schulte zur Sur · Accepted solution]

Hi John

you would have three choices. This one, because it probably involes Tru64 Unix or the storage forum, since it concerns the storage unit or maybe, the openvms.
The lights are supposed to go out.
You should be able to see the hsz50 disks with show dev. What happened to it? Has it been running before? Is that a single controller configuration or a redundant one?
Look at the controller(s) or the disks. Are there any amber lights? Next, connect a vt100 terminal or a notebook to the console port of the controller and use these commands:
show this full
show disk
show mirror
show unit
Post the results as attachment.

What os are you running?

greetings,

Michael

[Michael Schulte zur Sur]

Hi Gideon,

did you mean opening the pdf file?
It may be quite big and take it's time to load.

John, you might download it for reference.

greetings,

Michael

[John Appleby_1]

Hi,

Thanks for the reply. I have that document and the storageworks HSZ50 troubleshooting manual too but neither leave me any the wiser.

Even if I connect a single storageworks container to either onboard SCSI controller, I get nothing on them at all. I'd expect for the drives to show up as devices, albeit without the RAID configuration.

Regards,

John

[Michael Schulte zur Sur]

John,

what model of disk did you connect to what model of controller?

Michael

[G. Vrijhoeven]

Michael,

IE hanged opening the HTML link, not the pdf (Not responding)

John,

I would advice you to answer Michaels questions, your in luck he is online to help you.

Gideon

PS 0 Points please

[John Appleby_1]

Hi,

Sorry Michael, your replied after I had started writing my first reply to G.

The history of this is that it is a pull from Esprit Telecom that was sold to me. From what I can see the machine hasn't been messed with much but I have no idea what is on the machine yet or if I will need to install an OS.

Onto the configuration, from what I can see there is the following:

QLogic ISP10X0 connected to the backplane of the battery backup (I have no clue why this is, maybe there were some extra disks in there once).

DEC KZPSA connected to a pair of HSZ50/HSX inline. Terminated at both ends (is that right?). 3 of the channels on the HSZ50 module are connected to StorageWorks containers with 6 disks each in them.

All diagnostic lights on the unit are correct (HSZ50s are not amber and reset light is flashing once per second on each unit). The HSZ50 seems to correctly power up the drives one by one.

No need for an attachment :)

P00>>>show this full
P00>>>show disk
P00>>>show mirror
P00>>>show unit
P00>>>

I attached one of the storageworks arrays directly to both of the controllers in turn; the disks are RZ29B-VW.

Thanks very much for taking the time and I hope I have given you all the information
you asked for.

Regards,

John

[John Appleby_1]

I don't know if this is any use too:

P00>>>show dev
polling ncr0 (NCR 53C810) slot 1, bus 0 PCI, hose 1 SCSI Bus ID 7
dka500.5.0.1.1 DKA500 RRD45 0436
polling isp0 (QLogic ISP10X0) slot 0, bus 2 PCI, hose 1 SCSI Bus ID 7
polling pks0 (DEC KZPSA) slot 4, bus 0 PCI, hose 1
pkc.7.0.4.1 dkc TPwr 1 Fast 1 Bus ID 7 P01 A12
polling floppy0 (FLOPPY) PCEB - XBUS hose 0
dva0.0.0.1000.0 DVA0 RX23
polling tulip0 (DECchip 21140-AA) slot 2, bus 0 PCI, hose 1
ewa0.0.0.2.1 00-00-F8-30-03-30 FastFD (Full Duplex)

J

[Michael Schulte zur Sur]

Hi,

don't ask me about terminators. I haven't dealt with them too much. But I think, termination on both ends is right. The hsz50 flash roughly once per second at out site too. The commands I gave you, should be entered on the console of one of the hsz50. since it seems to be a dual redundant configuration add
show other full
show fail
to it.
Connect a simple vt100 terminal to it.

please post also
show config

thanks,

Michael

[John Appleby_1]

Ohhh onto the console of the HSZ50. Sorry!

Um I don't have a console cable for it so I will have to make one. I have a couple of different RJ45 console cables for cisco routers and stuff so maybe one of them will work - if not I'm sure the tech reference has the pinout.

Regards,

J

[Michael Schulte zur Sur]

John,

the show config was meant for the alpha console. Have you tried booting? Does the alpha have builtin disks?

Michael

[John Appleby_1]

Hi,

The 4100 doesn't have builtin disks, no; nor are there any disks not attached via the HSZ50. It claims its boot device to be dkc100.1.0.4.1 which sounds about right.

P00>>>boot
(boot dkc100.1.0.4.1 -flags a)
device dkc100.1.0.4.1 no longer valid

Show conf is below. I will have to make up a RJ11->RJ45 cable with the right pin connections tomorrow.

Regards,

John


P00>>>sh conf
Digital Equipment Corporation
AlphaServer 4100

Console V5.7-4 OpenVMS PALcode V1.21-2, Digital UNIX PALcode V1.23-2

Module Type Rev Name
System Motherboard 0 0000 mthrbrd0
Memory 512 MB EDO 0 0000 mem0
CPU (4MB Cache) 3 0001 cpu0
CPU (4MB Cache) 3 0001 cpu1
Bridge (IOD0/IOD1) 600 0032 iod0/iod1
PCI Motherboard 8 0000 saddle0

Bus 0 iod0 (PCI0)
Slot Option Name Type Rev Name
1 PCEB 4828086 0015 pceb0
2 S3 Trio64/Trio32 88115333 0000 vga0

Bus 1 pceb0 (EISA Bridge connected to iod0, slot 1)
Slot Option Name Type Rev Name

Bus 0 iod1 (PCI1)
Slot Option Name Type Rev Name
1 NCR 53C810 11000 0002 ncr0
2 DECchip 21140-AA 91011 0012 tulip0
3 PCI-PCI Bridge 11011 0002 pcb0
4 DEC KZPSA 81011 0000 pks0

Bus 2 pcb0 (PCI-PCI Bridge connected to iod1, slot 3)
Slot Option Name Type Rev Name
0 QLogic ISP10X0 10201077 0002 isp

[Michael Schulte zur Sur]

Hi,

if there aren't any disks attached, you can't see one.
You also can't just insert a disk into the hsz50 and expect the alpha to see anything. You have to run a configuration on the hsz50 and supply a unit for the alpha to see.

Here it late too. See you tomorrow,

Michael

[John Appleby_1]

Hi,

Am I to understand that if the configuration on the HSZ50 has been wiped, then despite the fact that there may be some disks attached, until containers have been made via the HSZ50 console, you won't be able to see any disks visible?

That does make sense of course, although it doesn't explain why when I connected the disk array to the other controller, I couldn't see any SCSI devices.

It seems that my main problem as it stands is finding a pinout for the console or finding a cable.

Thanks for the help; I don't think I can get any further until I've ordered a console cable.

Regards,

John

[John Appleby_1]

Bit of an update here - I've gotten it to boot into single user mode by putting the root disk into the NCR backplane and booting dkb500.

I don't think this helps me medium term because I still can't configure the HSZ50 in there.

Is there any way of configuring it without using the console port? Maybe you can access the console from within DU?

Regards,

J

[Jeff Wolfe_1]

John,

You are stuck on the config of the HSZ without a cable or a least 1 unit defined. With a unit defined, you can set host/scsi . But without that you'll have to find the proper cable. The units usally come with an RJ-45 to MMJ pigtail that will then accept a standard VT cable. I can post the part # for this cable on Monday AM EST.

[John Appleby_1]

Hi,

According to http://www.xenya.si/sup/info/digital/MDS/jun99/Cd3/STORAGE/HSZ50SVC.PDF (section 2.4, image 2.1) it's s 17-03511-04

I've searched on the net and a few people claim to have them in stock (no prices defined mind you).

I presume that unless I know the RAID configuration, there's no chance of being able to restore the old configuration and I'm going to be reinstalling from scratch.

In any case, from what I can see of the root partition, the machine is running DU4.0G and Oracle and is in a bit of a mess.

Regards,

John

[Michael Schulte zur Sur]

John,

the old company, who owned this machine probably put some energy into it to make sure, that no company data will be left on the disks. This includes deconfiguring the units/mirrors/disks.

In the documentation you should see the product number of the console cable. Look for this small adapter as Jeff mentioned.
Do you have a vt terminal?

What makes me wonder is, that it seems to see no internal disks.
What disks have you connected to the internal controller?

What are going to do with it? Using it for Tru64? If so, you may try booting from cdrom and see, if the installation programme sees any disks.

good morning,

Michael

[Jeff Wolfe_1]

John,

You'll also need part #12-43346. I never realized they were sold seperatly. My HSZ's always came with the 2 pieces as 1 piece.

Once you get access to the HSZ, you can run the config utility which will discover and add the disks into the configuration. You'll have no way of knowing what type of storage sets previoulsy existed. And as the previous post state, most users would have erased the disks. You can then create a 1 to 1 unit to disk map and see what the OS sees.

BTW..is the flash card installed in the slot of the HSZ controller(s)? If not then you have a bigger problem.

[John Appleby_1]

Hi,

The servers were decommissioned by the guy who sold it to me, who doesn't appear to have erased any data (and said he hadn't). This is probably a bit naughty on his part.

I don't know why the HSZ50 config appears to have been erased, but perhaps it had something to do with the PSU module that blew up with a flame when I first switched it on?

There are no internal disks (all disks on the HSZ50), but I moved the root partition to the "internal" backplane, which booted straight off. All the usernames are still intact, as is the fstab etc.; I don't see a HSZ50 config tool there though.

Yeah the HSZ50s are complete with program card.

I've not decided what OS to run on it yet, but since HP seem to offer a cheap non-commercial license for Tru64, I'm considering going with that.

I also have a stack of 18Gb disks which I'm considering replacing the 4Gb ones with. I've read the HSZ50 manual and setting it up looks pretty simple in the scheme of things.

I will have to go and buy one of these cables tomorrow (I need a PowerStorm graphics card too so I will get that at the same time). I'll duely report back if (when!) it's up and running.

Regards,

John

[Michael Schulte zur Sur]

Hi,

can you tell us, which version is currently running? Tru64 Unix has two versions of programmes to configure the hsz50. hszterm or hsxterm5 depending on whether it is 5.X or not. Look for them.
Can you post
#if 4.X
scan scan edt;
scu show edt
#if 5.X
hwmgr -scan scsi;
hwmgr -show -scsi
hwmgr -view hierarchy

greetings,

Michael

[Michael Schulte zur Sur]

Hi,

could you solve the problem?

greetings,

Michael

[John Appleby_1]

Hi,

Sorry about the delay in replying, I've not had time to have a proper go.

It's running 4.0G.

What I find strange is that when I pull the root disk off the HSZ50 and put it onto the other SCSI controller (NCR), it boots, but if I have the HSZ50 switched on, it crashes at boot time. I get:

...
rz13 at scsi1 target 5 lun 0 (LID=1) (DEC RZ28D (C) DEC 0010) (Wide16)
pza0 at pci1 slot4

And that's it. I've left it for quite a long time but that seems to make no difference.

When I boot without the HSZ50 backplane switched on, it boots. I am booting single user... otherwise it fails looking for the other partitions (obviously).

I found a copy of hszterm in the root folder of the root drive. It wasn't chmod +x so I did that, but I get:

hszterm: /sbin/loader: Fatal Error: cannot map libc.so

libc.so seems to live in /shlib, but I can't get it to link to it.

Regards,

John

[John Appleby_1]

Hi,

I'm back again. I was too cheap to buy a console cable set ($100+ are you kidding!). So, I guessed and made my own for $2. Here are the pinouts in case anyone is ever interested:

HSZ Signal DB9
1 DTR 4
2 TX 3
3 TX GND 5
4 RT GND 5
5 RX 2
6 DSR 6

This is for a 9-pin serial. Pins 7,8 of the serial cable need to be tied together at the DB9 end. Pin 1 of the HSZ is the left most when looking down at the RJ11 connector with the tab up. I used a Cat-5e cable and chopped off the brown, but whatever. And yes you do need to put TX and RX ground both into pin 5. I don't think it's actually necessary but hey, Digital kit is notoriously sensitive.

I found on my laptop that 9600, 1 stop, no parity and software flow control worked best.

I'm reading through the HSZ console commands that you guys asked me to run and I'll paste them all a bit later on.

Regards,

John