Re: Server hang (aiee:scheduling in interrupt)

From: Bill Wichers <>
Date: Mon, 16 Feb 1998 23:44:54 -0500 (EST)

Although it sounds to me like it's probably a software problem, I have
often resolved interrupt problems that seem to be random glitches (which
is typical of hardware of course) by doing the following:

(these are all BIOS parameters)
        * Set IRQ triggering to LEVEL (increases resistance to noise on
          the PCI bus)

        * Add a few more delays to PCI transactions (32 -> 64, 64 -> 96,

        * Disable any power management settings that may be active

You might also want to play around with some of the other PCI settings. I
have solved some IRQ problems that I could have sworn were software
problems by doing one of more of the above tweaks. If your system is
running right on the edge between working and crashing you want to give it
that little nudge to keep it from falling on the (dictated by Murphy's
law) crashing side. Hope this helps!


On Tue, 17 Feb 1998, lopy wrote:

> Dear Squid Users,
> We have setup a new linux box (slackware-2.0.30) running squid-1.1.20. We
> found that the machine will hang up sometimes with this error message
> "Aiee:scheduling in interrupt 00123ee9" on console. We can reboot the
> machine ONLY by power off and on again. Also the machine just run squid
> process ( without nfs, rpc, ppp, slip, etc).
> Sorry that maybe the problem is coming from LINUX OS, please forgive
> little bit off-the-topic. Can anyone tell me what's going wrong from my
> setup ?
> Here is the information :
> P5 - 133 MHz
> 128 MRAM
> Aha2940 UW card (Detect -> IRQ = 11)
> U-Fast-Wide 4 GB scsi disk
> 3c905 (run at 100M Full-Duplex)
> OS - Linux slackware 2.0.30
> Squid - 1.1.20
> Squid configure - cache_mem -> 24 M
> cache_swap -> 4 G
> Sorry for any inconvenience caused,
> Lopy
> Hong Kong Baptist University
Received on Mon Feb 16 1998 - 20:51:01 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:38:54 MST