I think I may have identified the culprit (or at least another possibility). It sort-of froze this morning, but perhaps due to changes in the 2.4.13-7 kernel, did not completely freeze me out yet :). I switched to a vc and saw this message repeating every several seconds after trying to log in as root:
PDC202XX: Secondary channel reset. ide_dmaproc: chipset supported ide_dma_lostirq func only: 13 hdg: lost interrupt hde: lost interrupt I'm wondering if either my controller is bad (I will probably re-seat it) or if there is a long-standing bug with the Promise Ultra/66 driver in a 586 SMP computer that might have become more noticeable recently due to possible enhancements in recent kernels that make it more efficient and able to drive the hardware closer to its limits. I'm going to look into forwarding this to the Linux IDE Person. When I get crashes with the new 2.4.13-x series, I usually boot back with a vanilla 2.4.6 kernel, which doesn't lock up unless if I transfer gigs of data from a Mac using netatalk (I'm now wondering if that's related to this same error... I'll have to test it.. This also froze 2.4.13-7, though I couldn't switch VCs on that one and had to reboot, where it froze during the fsck, so I had to turn it off and turn it back on). On Thursday 08 November 2001 08:05 am, you wrote: > This is probably more of a kernel issue or possibly hardware, but my system > has locked up twice within the last week while I was asleep :-(. In the > log file for the latest one (this morning), it appeared to happen during > the cron job run: > > Nov 8 04:02:00 penguin CROND[18995]: (root) CMD (run-parts > /etc/cron.daily) Nov 8 04:02:00 penguin anacron[18998]: Updated timestamp > for job > `cron.daily' to 2001-11-08 > Nov 8 04:02:10 penguin kernel: APIC error on CPU1: 08(00) > Nov 8 04:02:10 penguin kernel: APIC error on CPU0: 02(00) > Nov 8 04:02:10 penguin kernel: APIC error on CPU1: 01(00) > Nov 8 04:02:45 penguin kernel: APIC error on CPU1: 08(01) > Nov 8 04:02:45 penguin kernel: APIC error on CPU0: 02(00) > Nov 8 04:02:45 penguin kernel: APIC error on CPU1: 00(00) > Nov 8 04:02:49 penguin kernel: APIC error on CPU1: 09(00) > Nov 8 04:02:49 penguin kernel: APIC error on CPU0: 02(00) > Nov 8 07:01:56 penguin syslogd 1.4-0: restart. > Nov 8 07:01:56 penguin kernel: klogd 1.4-0, log source = /proc/kmsg > started. Nov 8 07:01:56 penguin kernel: Inspecting > /boot/System.map-2.4.13-2mdksmp Nov 8 07:01:56 penguin rpc.statd[855]: > Version 0.3.3 Starting > Nov 8 07:01:57 penguin nfslock: rpc.statd startup succeeded > Nov 8 07:01:57 penguin kernel: Loaded 15972 symbols from > /boot/System.map-2.4.13-2mdksmp. > > This system has the latest cooker as of 11/6/2001, and is running on a dual > Pentium 233/MMX (Pentium 1). > > I have actually got these APIC errors ever since I started using 2.4.x > (about a year I think..), with occasional freezes. With 2.2.x, I didn't > see the APIC errors, though it sometimes froze without warning. Back when > it had NT on it (3 years ago), the only thing that froze it was the > Creative Labs Sound Blaster AWE64 driver (froze NT during sound playback > and listed the driver at the top of the stack in the BSODs). > > Does anyone have any tips on how I might go about debugging the kernel? On > every crash that I've had since running Mandrake, the SysRq key > combinations to sync, unmount, and reboot don't work, but they do work when > the system is not crashed. I'm hoping that it's a software issue that can > be fixed, though I'm not ruling out hardware.. The processors felt very > warm in both cases, as if some part of the kernel was stuck in an infinite > loop or race condition. In some of my previous kernels, I've commented out > the printk() line that writes the APIC error to the log so that my hard > drives could actually pin down from less disk activity, but of course this > isn't an ideal solution ;-). > > I've read that the infamous Abit BP6 dual Celeron motherboards were > notorious for APIC errors and had a lot of crashing problems, but I'm > hoping that my motherboard, a GigaByte dual-Pentium (GA-5DX or something > like that), doesn't have the same problems :-) (it came out well before the > BP6, back in 1997). > > Crashes are quite rare, but do happen a couple times each month (unusual > for a Linux system). X on my Matrox Millennium II crashes sometimes, too > (about once a month), but it doesn't bring the system down with it. My > router, by contrast, has been up since I last turned it on in March 2001 > (old RedHat 6.2 system on a single Pentium 133 and 300MB hard drive). > > > --Steven Lawrance-- > [EMAIL PROTECTED] -- --Steven Lawrance-- [EMAIL PROTECTED]
