disk init fail

[Ruisheng Peng]

Hi All,

we have a cluster of two DEC3000 M400 alpha (running Openvms 6.2) and one VAXstation 3100 M76 (running OpenVMS 6.0). On one of the alpha there's a shadow set with two RZ29B disks. One of the set failed lately. Since I can't find another RZ29B disk to replace the failed one, I was going to set up another shadowed set with a pair of IBM DDRS-3456 (4.56GB) disks to backup the existing shadow set, and then replace it.

I attached the new set on the other alpha. Both disks correctly show up, at system power up as well as after boot. However, when I tried to initialize one of the disk:
$ init $3$dkb100: shad1
init failed, with a message saying the device is offline. The same is true for the other disk. I'm positive that the disks are good. Shutdown the system and placed jumper on pin #5 of the disks (enable auto spin, the factory setting, but I took the jumper off the first time around). Attached the enclosure to the alpha and boot up. The same thing. I'm puzzled on what I could have done wrong.

Thanks for the help,

--Ruisheng Peng

[Hoff]

There is information on disk integration in the OpenVMS FAQ; you're setting up to perform product integration and testing. V6.2 is OK at SCSI, and is the oldest that deals with ARRE and AWRE settings. V7.1-2 is better. Current (V8.3) is best. (This in terms of adapting to SCSI disk devices, and the configurations that can exist; the further back you go, the more you really want to use DIGITAL-branded disks.) Since I'm probably not going to talk you out of this approach (at least for now), look in the error log. The standard SKEY/ASC/ASCQ codes are usually listed at www.t10.org. As for the jumper, OpenVMS expects the power-on spin-up to be disabled; OpenVMS will send the spin-up request to the drive.

The alternative is to get a set of supported disks.

There's a Compaq StorageWorks array of fourteen 9 GB disks (and a rack) for about US$220 available for auction now.

There's a whole rack of RZ29 drives (7 RZ29B-VA 4.3GB drives) in a BA356 StorageWorks array for US$50 to US$100. This is the beige-colored narrow SCSI rack.

I'd be tempted to replace the whole configuration with newer gear. Particularly if you have the source code to the applications.

Stephen Hoffman
HoffmanLabs LLC

[Steven Schweda]

The usual answer is that IBM disks are not on
the supported hardware list. I have a
(newer) IBM-labeled Seagate drive on an
XP1000: "Disk ALP$DKB300:, device type
IBM-ESXS ST336605LC !#, [...]", but I can't
say that I've ever used a DDRS-3456.

It's possible that you have SCSI ID, cable,
and/or termination problems, or it may be
that DKDRIVER just doesn't like those drives.
I would not expect the auto-spin-up setting
to make any real difference, unless you can
hear that the drives just aren't spinning up.
On all my DEC/Compaq VMS systems (Alpha and
VAX), if the drive doesn't spin up by itself,
someone (console firmware or VMS) will send
it a "start unit" command to get it going.

[Steven Schweda]

Re:

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1134521

> [...] ST15150N [...]

As I recall, those are early Barracuda drives,
and run hotter than some toasters I've used.
If they're still working, I'll assume that
they've been well ventillated up to now, but
if it were up to me, I'd look for something
newer and cooler, especially if I were
interested in reliability.

[Vladimir Fabecic · Accepted solution]

Just one question.
Did you install this patch?
ftp://ftp.itrc.hp.com/openvms_patches/alpha/V6.2X/ALPSCSI08_062.A-DCX_AXPEXE
It was a long time ago, but after installing this patch some disks worked that did not work before.
At least you can try.

[Ruisheng Peng]

Hi all,

Thanks for the replies.

I indeed thought "S" stands for standard :) It turned out, as Vladimir pointed out, I didn't have the aplscsi08_062 patch on the system. Thanks for the link to the patches--I never thought they'd still be available on the net. Get the patch applied, and it was about time I started from scratch again as one of Seagate Barracuda disks I was using to built the shadowset failed. The short story is that the IBM DDRS-3456 disks worked after applying the patch.

It is a struggle to maintain "legacy" hardware, especially at 14000ft. I literally have a backup disk for every disk on the live system. But the RZ29s, RZ26s are running hot and air is thin up there. Hopefully the IBM DDRs-3456 will do better, as they draw less current than the Seagate ST15150N/RZ29N bricks.

With the SCSI patch apllied, I'm attempting to try bigger drives, such as 9GB ones as they'd be easier to come by and tend to be relatively recent. Is here a particular brand that works well with alpha? Also, would be be problem when mounting these drives hosted on the alpha on the Vax (OpenVMS 6.0)? Should the Vax be similarly patched?

Thanks again,

--Ruisheng Peng

[Steven Schweda]

> [...] they draw less current than the
> Seagate ST15150N/RZ29N bricks.

Practically _everything_ draws less power
than an ST15150N. (Some small _towns_ draw
less power than an ST15150N.)

At a former employer, we bought some of those
drives when they were new, and installed them
in some external boxes, same kind as we had
been using for the older, slower (and, as it
turned out, cooler) Hawk (5400r/m) drives.
It was not long before they started failing
(under warranty). Maybe a year or so. When
we opened one of the enclosures and learned
how hot the new drives were getting, we
started adding extra fans which would blow
directly at the cases of the new drives.
This cooled them down a lot, and we had no
more premature failures. (We also bought no
more of those drives, and we started looking
very closely at the power numbers when we
bought anything else.)

At 14000 feet, I'd be looking around for
places to put additional fans to blow on all
the hot things.

[Steven Schweda]

> Is here a particular brand that works well
> with alpha?

I've used Fujitsu and Seagate with no
trouble. (I've been at VMS V7.x for a long
time, so patches are less of a worry than for
V6.x. Seagate drives seem to last longer,
but it's a small sample, so my statistics
may be unreliable.)

> Also, would be be problem when mounting
> these drives hosted on the alpha on the
> Vax (OpenVMS 6.0)?

If the Alpha is the system with the SCSI
interface, then the VAX shouldn't care, but,
as with any unsupported hardware, the testing
is your job. (The VAX would see an
MSCP-served disk, and would not care if it's
SCSI or DSSI or SDA or whatever.)

[Steven Schweda]

Trying to consolidate the two threads:

> I also need to hunt down some 1GB drives
> for the system disk on the VAX...

I use Seagate ST11200N drives in my old
VAXstations, but it's getting harder to find
nice 1GB drives. You may need to watch eBay
for some time before you see anything
attractive.