SUMMARY: raid failure during configuration...

From: System Janitor <hubcap_at_hubcap.clemson.edu>
Date: Wed, 7 Aug 1996 14:13:55 -0400 (EDT)

I asked about a problem I had with a Mylex DAC960 Three Channel
PCI Backplane RAID Controller where a failed configuration attempt
left me with a 2100 that wouldn't even boot into SRM mode,
instead looping endlessly with the message:

      waiting for dra.0.0.6.0 to poll

Thanks for responses from:

  From: rlangrid_at_progress.co.uk (Roy Langridge)
  From: "Dr. Tom Blinn, 603-881-0646" <tpb_at_zk3.dec.com>
  From: Roger Joss <giarjo_at_gia.ch>
  From: jme_at_haydn.generale.be (Jan Mevis)

I also got some help from DEC. Somehow, instead of getting sent to
the normal overworked general purpose telephone help-line, they
sent me to someone with extensive RAID knowledge somewhere in Florida.

With the help from the list, and a little extrapolation of my own,
I had gotten past the problem I reported here before the DEC guy got
back to me. He mentioned a couple of pretty sage things, though,
and I'll pass at least one on: Some RZ style drives with older
firmware spin down if they haven't been accessed in a while.
This makes raid controllers think they are broken. If anyone has
disks in their raid sets that maddeningly go bad on
occasion for seemingly no good reason, this could be something
to check. He mentioned another problem that won't affect me, so
I only half remember it: filling up a 3 channel array (21 disks) with
disks that all spin at 7200 RPM (many spin at 5400) could cause either
heat problems or over stress power supplies, especially if
the controller is configured to spin too many (more than 4 I think)
up at power-on.

Anyway:

The main help I got from the list was to wait out the looping
error message, it would finally stop, leaving me in a functional
SRM mode.

Once I got to SRM mode, I figured my problems were over, I could
just go to ARC mode and reconfigure. Wrong, cranking up ARC didn't
bring me to the main ARC menu, but rather to the menu I was left
at when the original configuration failed. This menu allowed
for nothing but the ability to choose between saving or not saving
the failed configuration. I figured of course I didn't want to save it.
Choosing the ``not save'' option left me with a message to reboot,
and a reboot left me right back where I was: unable to get to
the ARC main menu. Several of the people from the list who helped me
said that they had to reinstall their raid controller firmware. I am
assuming they reinstalled to get past what I was facing here.

What I did instead was to save the failed configuration, which
was kind of counter-intuitive. But that got things going again,
allowing me to go to the main ARC menu on reboot, and from there
I just went to the ``tools'' section and told the controller to
forget about the failed configuration.

After that I just started over with the cookbook configuration
instructions in the StorageWorks User's Guide, and now I have
a functional raid array.

Thanks...

-Mike
Received on Wed Aug 07 1996 - 20:59:12 NZST

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:46 NZDT