RHEL AS3 crashes on iSeries - More info

["Len Goldenstein" <len ghy com>]

After the logical partitions crashes, the iSeries gives a reference code of:
B600 F104

According to the IBM FAQ:
Why did my Linux IPL hang with a B600 F104 SRC?

You may have too much memory assigned to the Linux partition. The 32 bit
kernel can only use 4 Gigabytes of RAM. If you assign more than that, your
Linux partition may hang with the following SRCs (seen in the LPAR Monitor
Partition Status screen): B600 F104, 03C0 0000, 0000 0001. The recovery is
to reduce the amount of memory assigned to the Linux partition to be less
than 4 Gigabytes.

However, this partition only has 256M allocated to it (and I tried the same
virtual disk on two different partitions).

Very weird...

-----------------------------------------
Len Goldenstein
Network Administrator

Geo. H. Young & Co. Ltd.
809 - 167 Lombard Ave.
Winnipeg, MB R3B 3H8
Phone: (204) 947-6700 Ext. 203
Fax: (204) 947-3306

len ghy com
http://www.ghy.com

Due to changing circumstances and the complexity of Customs Regulations, our
opinions are based on the available information and should not be considered
binding. If you would like a binding ruling, please do not hesitate in
contacting our Professional Service Department via email at
consulting ghy com 

GHY International is PIP Approved (Canada) and C-TPAT Certified (USA)

> -----Original Message-----
> From: Len Goldenstein [mailto:len ghy com]
> Sent: September 28, 2004 8:23 AM
> To: Linuxppc-Iseries Lists  Linuxppc. Org; Taroon-List Redhat  Com
> Subject: RHEL AS3 crashes on iSeries
>
>
> Good morning everyone,
> I have the unfortunate situation that what was a stable install
> of RHEL AS 3 on the IBM iSeries platform has taken a turn for the worse.
>
> It was quite stable at AS3 Update 2 but during the update to U3,
> the system locked up and now the system (Logical Partition) is
> plagued by frequent crashes anywhere from 15 minutes after boot
> to crashing during boot.
>
> Even just starting the system a moment ago gave the following during boot:
> ...
> Mounting other filesystems:                                [  OK  ]
> Starting named:                                            [  OK  ]
> Starting sshd:                                             [  OK  ]
> nfs lockd sunrpc audit veth loop ext3 jbd lvm-mod viodasd sd_mod
> scsi_mod
> NIP: c00000000009a048 XER: 0000000000000000 LR: c00000000009a01c
> REGS: c00000000f89b070 TRAP: 0300    Not tainted
> NIP is at .__kmem_cache_alloc [kernel] 0x88 (2.4.21-20.EL)
> MSR: a000000000008032 EE: 1 PR: 0 FP: 0 ME: 0 IR/DR: 11
> TASK = c00000000f898000[754] 'named'
>
> I did try the U3 update on a test system and it worked fine.  I
> really don't know why the test system is OK and the production
> system is no longer stable.  I've tried using the U2 and U1
> kernels but the same problem still persists.
>
> Also, I've tried moving the iSeries Logical Partition virtual
> disk to another Logical Partition and the problem still persists.
>  This leads me to believe that the problem is not with the
> Logical Partition itself but rather the operating system.
>
> I've attached some of the crash logs that were on the console.
> It's really over my head so hopefully someone smarter can make
> more sense of it ;)
>
> If anyone can offer advice, it would be greatly appreciated.
> -----------------------------------------
> Len Goldenstein
> Network Administrator
>
> Geo. H. Young & Co. Ltd.
> 809 - 167 Lombard Ave.
> Winnipeg, MB R3B 3H8
> Phone: (204) 947-6700 Ext. 203
> Fax: (204) 947-3306
>
> len ghy com
> http://www.ghy.com
>
> Due to changing circumstances and the complexity of Customs
> Regulations, our
> opinions are based on the available information and should not be
> considered
> binding. If you would like a binding ruling, please do not hesitate in
> contacting our Professional Service Department via email at
> consulting ghy com 
>
> GHY International is PIP Approved (Canada) and C-TPAT Certified (USA)