RAID 5 array on Promise SX6000 failing after fsck at boot

[Fredrik Lindberg]

Hi,

I have a weird(?) problem with FreeBSD 4.8-RELEASE and a Promise SuperTrak
SX6000 (device pst), running a RAID 5 setup with 4 180gb IDE-drives
(identical drives).

The RAID controller itself and the logical disk is detected fine
--snip from dmesg---
pstpci0: <Promise SuperTrak SX6000 ATA RAID controller> mem
0xedc00000-0xedfffff
pst0: 530100MB <PROMISE TECH. I2O RAID DEVICE> [67578/255/63] on pstpci0

I ran newfs and diskelabel as with a normal drive (no errors) and I can
mount and unmount the drive without problem.
Writing/reading does also work.
However, when I add the drive to /etc/fstab and set the pass-field to >0
to make fsck check the filesystem at bootup the problems shows up.

At boot, fsck "hangs" when running on the /dev/pst0s1e drive and the raid
card beeps.
If reboot the machine, 1 or more channels on the array has failed (rebuild
is needed).
If I use 0 in the pass-field avoiding fsck to run on the drive everything
works just fine.

I've searched the mailing archives and the web but I haven't found
anything similar.
I don't have much experience with raid and freebsd.
Am I missning something vital? Am I not supposed to run fsck on a logical
raid "drive"?

Any help is really appreciated.

Thanks,
       Fredrik Lindberg


This is the output of disklabel. There is no output of fsck.

# /dev/pst0s1e:
type: unknown
disk: amnesiac
label:
flags:
bytes/sector: 512
sectors/track: 63
tracks/cylinder: 255
sectors/cylinder: 16065
cylinders: 67577
sectors/unit: 1085640507
rpm: 7200
interleave: 1
trackskew: 0
cylinderskew: 0
headswitch: 0           # milliseconds
track-to-track seek: 0  # milliseconds
drivedata: 0

8 partitions:
#        size   offset    fstype   [fsize bsize bps/cpg]
  c: 1085640507        0    unused        0     0               # (Cyl.
0 - 67577*)
  e: 1085640507        0    4.2BSD     4096 16384   158         # (Cyl.
0 - 67577*)