Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 17 Dec 2000 14:06:24 +0200
From:      "Noor Dawod" <noor@comrax.com>
To:        <freebsd-questions@freebsd.org>
Cc:        <scott_long@btc.adaptec.com>
Subject:   Adaptec SCSI RAID 2100S problems.
Message-ID:  <PHEBIOJOBJJLIIJCOINKIELCCPAA.noor@comrax.com>

next in thread | raw e-mail | index | archive | help
Hello all,

We've just bought a brand-new server for our company and chose Adaptec
SCSI RAID 2100S to do RAID 5 on 4 IBM Ultrastar 18GB disks. 3 of the IBM
disks have similar revisions (MLC: DS0S80) and one of the disks has a
different revision (MLC: DS0S93). The other physical characteristics are
the same (tracks, sectors, clusters, disk size; the size on all 4
disks).

Before designing the RAID Array, the 2100S manages to identify all the
disks, their sizes, their features and their speed. Even if I reboot the
system few times, shut it down and then starts it up again, the RAID
2100S card still identifies the disks wonderfully (you'll know later why
this was done.)

Next, I go into the 2100S BIOS (called SMOR) and design the required
Array (I chose RAID 5 fault-tolerance). Once I save the new settings,
and the RAID Array starts building, I reboot the system.

Now, the problems start... One out of 3-4 reboots, the 2100S card
manages to identify all disks, reads the configured settings, and boots
the system (into FreeBSD which I managed to install on a previous reboot
under the configured Array). The other attempts do not succeed, and on
the Disk Cage I see 3 red lights out of 4, indicating that the 2100S
failed to identify disks in those bays. In addition, I hear a
high-volume alarm coming out of the 2100S card indicating a problem.

Next, I switched the power off, changed positions of the disks (1st disk
to 4th bay, 2nd disk to 3rd bay, and so on), and turned the power on.
The same error described above is repeated, but this time, 3 different
red lights out of 4 are led on the cage!! I've continued to change
positions of the disks in this fashion (and even after I deleted this
Array and created a new RAID-1 Array), but it's always that one disk is
identified (and its light is green), and 3 are not identified (and their
lights are red).

The strangest thing in this whole scenario is that the disk with the
different revision is ALWAYS identified, while the 3 similar revision
disks are not identified and FAIL wherever you place them in the bays!!
Another strange thing is that when I remove the disk with the different
revision, and I design a new Array using only the 3 similar disks, ALL
goes well. Even after 10 (!!!) reboots, the system identifies the Array,
the disks, and then boots the OS with no noticeable errors.

By the way, sometimes the system boots normally while the Array is
designed and created. Although most of the times, the erroneous scenario
described above happens more occasionally.

Does anyone know what is causing this problem, and if revision changes
might cause this problem? Does the Adaptec 2100S RAID controller have a
problem dealing with revision changes, or with IBM disks in particular?

Any help would be appreciated.

Thank you.
Noor

P.S.: Meanwhile, I've ordered a new disk with the same revision as the 3
to see what'd happen with 4 similar disks.



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?PHEBIOJOBJJLIIJCOINKIELCCPAA.noor>