samba - Oct 2003 - Kernel Panic, memory leak?

Hi,

It's become almost routine now to reboot the server every morning.  If 
we are able to log in, after a lag of ~10 seconds, we see the load 
average is usually > 3.0.
As I am writing this, it just went down again.  The console had a bunch 
of text on it, the very bottom reading:

<o> Kernel panic : Aiee, Killing Interrupt handler!
In interrupt handler - not syncing

Here is the machine info.  Currently running Red Hat 8, Samba 3.0, with 
a PIII, and a RAID.

# uname -a
Linux BENT00 2.4.18-14 #1 Wed Sep 4 13:35:50 EDT 2002 i686 i686 i386 
GNU/Linux

# less /proc/meminfo
         total:    used:    free:  shared: buffers:  cached:
Mem:  1055735808 1043091456 12644352        0 15740928 935366656
Swap: 2089209856        0 2089209856
MemTotal:      		1030992 kB
MemFree:         		12348 kB
MemShared:           	0 kB
Buffers:         		15372 kB
Cached:         		913444 kB
SwapCached:          	0 kB
Active:         		661920 kB
Inact_dirty:     		61648 kB
Inact_clean:    		222832 kB
Inact_target:   		189280 kB
HighTotal:      		130992 kB
HighFree:         		1024 kB
LowTotal:       		900000 kB
LowFree:         		11324 kB
SwapTotal:     		2040244 kB
SwapFree:      		2040244 kB
Committed_AS:    	42644 kB

# less /proc/interrupts
            CPU0
   0:    1455741          	XT-PIC  timer
   1:          4          		XT-PIC  keyboard
   2:          0         		XT-PIC  cascade
   5:   10806406          XT-PIC  eth0
   8:          1          		XT-PIC  rtc
  10:         60          	XT-PIC  eth1
  11:     394317          	XT-PIC  aic7xxx, usb-uhci, usb-uhci
  12:         20          	XT-PIC  PS/2 Mouse
  14:          1          	XT-PIC  ide0
NMI:          0
ERR:          0


is it possible there is a memory leak?  if so, what are some good leak 
detection tools?  seems to me it's most likely an IRQ problem.  great, 
looking forward to dealing with that.

thanks for reading.

Jared

On Wed, Oct 15, 2003 at 03:01:19PM -0700, Francis Buxton
wrote:
> Hi,
> 
> It's become almost routine now to reboot the server every morning.  If 
> we are able to log in, after a lag of ~10 seconds, we see the load 
> average is usually > 3.0.
> As I am writing this, it just went down again.  The console had a bunch 
> of text on it, the very bottom reading:
> 
> <o> Kernel panic : Aiee, Killing Interrupt handler!
> In interrupt handler - not syncing
Not a Samba problem I'm afraid. We can't kill your kernel (easily,
anyway :-).

Jeremy.