PE 2650 Perc3/DI aacraid Kernel 2.6

Dominik L. Borkowski dom at vbi.vt.edu
Wed Jul 6 22:31:37 CDT 2005


Hello,
I figured I'd start here before I bug the kernel mailing list. Seems I'm not 
the first person to encounter the infamous 'rejecting I/O to offline device' 
bug. 

Long story short, since the advent of 2.6.11 and up, we've had both 2650's and 
2850's encountering this issue. 2.6.12 seems to be holding up ok on the 
2850's with the MSI raid controllers, but the 2650's with aacraid driver are 
still dying after a week of running 2.6.12. 

The latest bios [a20] and the raid firmware [2.8-6095] do not seem to help. 
I'm downgrading back to 2.6.10, which seemed to be working ok. 

Thus the question prompted by the fact that some people reported that 2.8-6095 
firmware seems to be working out for them: is this an issue that I should be 
bugging Dell about, or the kernel developers? Like most, I have noticed that 
2.4.x kernels worked just fine, and earlier 2.6.x were also [at least if I 
recall correctly, the bug wasn't triggered that often]. 

The average uptime of my machines, regardless of the bios and firmware seems 
to be roughly a week. I haven't found a pattern of heavy disk I/O that would 
cause it, or at least I haven't been able to trigger it manually.

machine1:
Red Hat/Adaptec aacraid driver (1.1.2-lk2 Jun 20 2005)
ACPI: PCI Interrupt 0000:04:08.1[A] -> GSI 30 (level, low) -> IRQ 30
AAC0: kernel 2.8-0[6095]
AAC0: monitor 2.8-0[6095]
AAC0: bios 2.8-0[6095]
AAC0: serial a92c41d3
scsi0 : percraid
  Vendor: DELL      Model: PERC RAID5        Rev: V1.0
  Type:   Direct-Access                      ANSI SCSI revision: 02

machine2:
Red Hat/Adaptec aacraid driver (1.1.2-lk2 Jun 22 2005)
ACPI: PCI Interrupt 0000:04:08.1[A] -> GSI 30 (level, low) -> IRQ 30
AAC0: kernel 2.7-1[3170]
AAC0: monitor 2.7-1[3170]
AAC0: bios 2.7-1[3170]
AAC0: serial 12dc41d3
scsi0 : percraid
  Vendor: DELL      Model: PERC RAID5        Rev: V1.0
  Type:   Direct-Access                      ANSI SCSI revision: 02

I guess I wonder if anybody has found a stable solution for this annoying bug. 
Any hints/suggestions would be greatly appreciated.

Sincerely,
Dominik L. Borkowski



More information about the Linux-PowerEdge mailing list