Random Kernel Panic on Boot

[Mike Baptiste <mike.baptiste@xxxxxxxx>]

I've got a new system I'm working to put into production and I have one 
nagging problem that I'm trying to resolve and perhaps one of you has 
seen this or can point me in the right direction.  I know this may not 
be an XFS problem but since I'm using the XFS installer I figured I'd 
start here.

I've got a Dual Athlon server (APPRO 1124 - Tyan Thunder K7 firmware 
2.08) with a Mylex AccelRaid 170 (32MB) with dual 18GB and dual 36GB 
Seagate Cheetahs in a RAID1 config.

DAC960: ***** DAC960 RAID Driver Version 2.4.10 of 23 July 2001 *****
DAC960: Copyright 1998-2001 by Leonard N. Zubkoff <lnz@xxxxxxxxxxxxx>
DAC960#0: Configuring Mylex AcceleRAID 170 PCI RAID Controller
DAC960#0:   Firmware Version: 6.00-13, Channels: 1, Memory Size: 32MB
DAC960#0:   PCI Bus: 0, Device: 8, Function: 1, I/O Address: Unassigned
DAC960#0:   PCI Address: 0xF4004000 mapped at 0xF8832000, IRQ Channel: 10
DAC960#0:   Controller Queue Depth: 512, Maximum Blocks per Command: 2048
DAC960#0:   Driver Queue Depth: 511, Scatter/Gather Limit: 128 of 257 
Segments
DAC960#0:   Physical Devices:
DAC960#0:     0:0  Vendor: SEAGATE   Model: ST318406LC        Revision: 0108
DAC960#0:          Wide Synchronous at 160 MB/sec
DAC960#0:          Serial Number: 3FE00G5L00002216GSGY
DAC960#0:          Disk Status: Online, 35807232 blocks
DAC960#0:     0:1  Vendor: SEAGATE   Model: ST318406LC        Revision: 0108
DAC960#0:          Wide Synchronous at 160 MB/sec
DAC960#0:          Serial Number: 3FE00KDJ00007216A142
DAC960#0:          Disk Status: Online, 35807232 blocks
DAC960#0:     0:2  Vendor: SEAGATE   Model: ST336706LC        Revision: 0108
DAC960#0:          Wide Synchronous at 160 MB/sec
DAC960#0:          Serial Number: 3FD08JH200007216E1CD
DAC960#0:          Disk Status: Online, 71651328 blocks
DAC960#0:     0:3  Vendor: SEAGATE   Model: ST336706LC        Revision: 0108
DAC960#0:          Wide Synchronous at 160 MB/sec
DAC960#0:          Serial Number: 3FD08H4000002217FADN
DAC960#0:          Disk Status: Online, 71651328 blocks
DAC960#0:     0:7  Vendor: MYLEX     Model: AcceleRAID 170    Revision: 0600
DAC960#0:          Wide Synchronous at 160 MB/sec
DAC960#0:          Serial Number:
DAC960#0:     0:9  Vendor: QLogic    Model: GEM359            Revision: 1.07
DAC960#0:          Asynchronous
DAC960#0:          Serial Number: 1
DAC960#0:   Logical Drives:
DAC960#0:     /dev/rd/c0d0: RAID-1, Online, 35807232 blocks
DAC960#0:                   Logical Device Initialized, BIOS Geometry: 
255/63
DAC960#0:                   Stripe Size: 64KB, Segment Size: 8KB
DAC960#0:                   Read Cache Disabled, Write Cache Disabled
DAC960#0:     /dev/rd/c0d1: RAID-1, Online, 71651328 blocks
DAC960#0:                   Logical Device Initialized, BIOS Geometry: 
255/63
DAC960#0:                   Stripe Size: 64KB, Segment Size: 8KB
DAC960#0:                   Read Cache Disabled, Write Cache Disabled


I installed RH 7.2 using the SGI XFS Installer with no problems.  I 
configured a fairly normal setup for partitions (all XFS):

Filesystem           1k-blocks      Used Available Use% Mounted on
/dev/rd/c0d0p2         4188164   1296152   2892012  31% /
/dev/rd/c0d0p1           59428      8476     50952  15% /boot
none                    513540         0    513540   0% /dev/shm
/dev/vg00/homelv       9432384       180   9432204   1% /home
/dev/vg00/varlv        9432384     42548   9389836   1% /var
/dev/rd/c0d0p5          260240     53040    207200  21% /tmp

Note the LVM partitions were added later (/var & /home used to be 
contained in c0d0p2) and this problem showed up prior to that so 
disregard LVM in this case.

The problem is this:  When booting either the smp or enterprise kernels 
(stock RH 2.4.9-XFS from the XFS installer CD), I get random kernel 
panics.  Maybe 3 out of 4 boots, at varying points in Interactive boot 
(from LVM activiation forward), I'll get a Kernel Panic from the DAC960 
driver:

DAC960#0: SegmentNumber != SegmentCount

These panics never happen in the same place, but its always after the 
root filesystem has been mounted while the init.d scripts are running 
(though a couple times I've seen it happen during LVM activiation in 
rc.sysinit)  I've never seen it happen booting the uni-proc kernel.

I've got APIC disabled due to the AMD Interrupt Errata (#22) with the 
760MP chipset.

The RAID Segment size is 8KB.

If the system boots without a panic - its been rock solid - no super 
heavy load, but lots of compiling and pkg installation without a single 
hiccup.  But rebooting is always a shot in teh dark since the majority 
of the time it'll panic, but if I hard reset once or twice, it'll boot 
normally and run fine.

Any ideas?  Could this be XFS related?  I searched all over with Google 
and Marc - Couldn't find any mention of a problem like this.  I know 
Mylex 170's are fairly common in Linux configs so if this was a common 
non-XFS problem I figure I'd have come across something, but then again, 
maybe not :)

Mike
--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Mike Baptiste            202 Hudson Hall, Box 90271, Durham, NC 27708
Director of Information Technology             mike.baptiste@xxxxxxxx
Pratt School of Engineering @ Duke University      Phone:919-660-5404