Thanks Johnny -- and everyone. It'll take me a while to work through it here. I'll try to take a stab at everything everyone suggested, though it might be the weekend before I get time (and, to be truthful, the courage).
Meantime, I can still use the computer. (Just to clarify: The crashes only began after I replaced XP with Red Hat 8.0 and Windows 2000 Professional. (XP is not going back on this machine as long as I own it.) And my Dell is a desktop (rather than a laptop). Again, thanks everyone! Pat On Mon, 2003-02-03 at 08:08, Johnny Stork wrote: > Looking at your syslog output Pat, I would gfuess that these lines could help to >point out some things to check, related to IRQ's > > > Feb 2 08:53:17 alic6e0y37ff kernel: oprofile: can't get RTC I/O Ports > > This one looks very suspicious but I dont know where to look for "Real Time >Controller" (RTC) issues > > > Feb 2 08:53:18 alic6e0y37ff kernel: ohci1394: pci_module_init failed > > Hmmm, maybe you can disable Firewire support in the BIOS and see what happens? > > > Feb 2 12:13:01 alic6e0y37ff kernel: PCI: Found IRQ 11 for device 01:04.0 > > > Feb 2 12:13:01 alic6e0y37ff kernel: PCI: Sharing IRQ 11 with 00:1f.1 > > These two look the most suspicious.....related to the previous APIC suggestions >possibly. IRQ 11 can be used for video, pcmcia, scsi controllers etc but I would >guess that it is trying to share that IRQ for the two video cards. > > Combining all the suggestions so far... > > 1: Open the case, re-seat all cables and ensure they are snug and not twisted >sharply. > > 2: Check the BIOS and MB manual for any way to disable the on-board video (maybe a >jumper as suggested). If your looking at the mb, you will need a flashlight and >depending on your eyes, a magnifying glass or very careful search near jumpers. I did >take a quick look but dont recall anything other than the standard clear-bios jumper. > > 3: Check the BIOS and MB manual for an ability to disable Firewire support. > > 4: One thing you could try as a process of elimination, is remove the new video card >and try both Windows and Linux just long enough to determine that the behaviour has >changed. In other words, if the crashing suddenly stops, then we can be fairly >certain that it is a conflict between the two video controllers. > > If you are uncomfortable trying these things, then I can come by one night this week >and we can go through all these together. Just let me know when. > > > > > <hr> > <b><font color=blue size=4>Open Enterprise Solutions</font></b> > <b><font color=red>Linux & Open Source Solutions for Business</font></b> > > Johnny Stork, BA > Calgary, AB > Canada > > <a href="http://www.openenterprise.ca";> > www.openenterprise.ca</a> > > ------ original message ------ > From: Pat Roche <[EMAIL PROTECTED]> > To: CLUG <[EMAIL PROTECTED]> > Sent: Sun Feb 02 18:51:45 PST 2003 > Subject: Re: (clug-talk) (RH8-Dell D. 2350) crashes > > Thanks, again Szemir. I went through the log file and copied some > <br>samples of what may (or may not) be error-related messages. Some of > <br>these lines also have references to IRQ (which Garth mentioned -- thanks > <br>for your very detailed reply, Garth.) By the way, whenever I boot up, > <br>the line reading "Bringing up interface eth0:" seems to display for > <br>quite a while (probably 15-20 seconds). I don't know if that indicates > <br>something is amiss with the network card. Anyway, here are some samples > <br>from the log files: > <br> > <br>Feb 2 08:53:17 alic6e0y37ff kernel: oprofile: can't get RTC I/O Ports > <br>Feb 2 08:53:17 alic6e0y37ff kernel: block: 992 slots per queue, > <br>batch=248 > <br>Feb 2 08:53:17 alic6e0y37ff kernel: Uniform Multi-Platform E-IDE driver > <br>Revision: 6.31 > <br>Feb 2 08:53:17 alic6e0y37ff kernel: ide: Assuming 33MHz system bus > <br>speed for PIO modes; override with idebus=xx > <br>Feb 2 08:53:17 alic6e0y37ff kernel: PIIX4: IDE controller on PCI bus 00 > <br>dev f9 > <br>Feb 2 08:53:17 alic6e0y37ff kernel: PCI: Found IRQ 11 for device > <br>00:1f.1 > <br>Feb 2 08:53:17 alic6e0y37ff kernel: PCI: Sharing IRQ 11 with 01:04.0 > <br>Feb 2 08:53:17 alic6e0y37ff kernel: PIIX4: chipset revision 2 > <br>Feb 2 08:53:17 alic6e0y37ff kernel: PIIX4: not 100%% native mode: will > <br>probe irqs later > <br> > <br>Feb 2 08:53:18 alic6e0y37ff kernel: EXT3-fs: mounted filesystem with > <br>ordered data mode. > <br>Feb 2 08:53:18 alic6e0y37ff kernel: IA-32 Microcode Update Driver: > <br>v1.11 <[EMAIL PROTECTED]> > <br>Feb 2 08:53:18 alic6e0y37ff kernel: microcode: CPU0 no microcode found! > <br>(sig=f27, pflags=4) > <br>Feb 2 08:53:18 alic6e0y37ff kernel: parport0: PC-style at 0x378 > <br>[PCSPP,TRISTATE,EPP] > <br>Feb 2 08:53:18 alic6e0y37ff kernel: ohci1394: pci_module_init failed > <br> > <br>Feb 2 12:13:01 alic6e0y37ff kernel: IA-32 Microcode Update Driver: > <br>v1.11 <[EMAIL PROTECTED]> > <br>Feb 2 12:13:01 alic6e0y37ff kernel: microcode: CPU0 no microcode found! > <br>(sig=f27, pflags=4) > <br>Feb 2 12:13:01 alic6e0y37ff kernel: parport0: PC-style at 0x378 > <br>[PCSPP,TRISTATE,EPP] > <br>Feb 2 12:13:01 alic6e0y37ff kernel: ohci1394: pci_module_init failed > <br>Feb 2 12:13:01 alic6e0y37ff kernel: ip_tables: (C) 2000-2002 Netfilter > <br>core team > <br>Feb 2 12:13:01 alic6e0y37ff kernel: via-rhine.c:v1.10-LK1.1.14 > <br>May-3-2002 Written by Donald Becker > <br>Feb 2 12:13:01 alic6e0y37ff kernel: > <br><a >href="http://www.scyld.com/network/via-rhine.html"target="top";>http://www.scyld.com/network/via-rhine.html</a> > <br>Feb 2 12:13:01 alic6e0y37ff kernel: PCI: Found IRQ 11 for device > <br>01:04.0 > <br>Feb 2 12:13:01 alic6e0y37ff kernel: PCI: Sharing IRQ 11 with 00:1f.1 > <br> > <br>--Pat > <br> > <br> > <br>On Sun, 2003-02-02 at 18:02, Bogi wrote: > <br>> Hi Pat. > <br>> > <br>> You seem to have done fairly well with the df command. Not Knowing is not a > <br>> shame, not wanting to know IS. There is nothing wreong about not knowing as > <br>> long at you are trying to know, and you are. > <br>> > <br>> the command line should look along the lines of: > <br>> > <br>> cat /var/log/messages | less > <br>> > <br>> then you can scroll through the output using the cursor keys including pgup > <br>> and pgdn etc... > <br>> > <br>> when satisfied / bored reading the logfile, exiting is possible with q > <br>> > <br>> And you will, i am almost sure seee the error messages if there are any. > <br>> The good once are usually very close to a reboot :-) / crash. > <br>> > <br>> Cheers > <br>> Szemir > <br>> > <br>> On Sunday 02 February 2003 17:49, you wrote: > <br>> > Thanks, Szemir. Here's what I get from df: > <br>> > > <br>> > Filesystem 1K-blocks Used Available Use% Mounted on > <br>> > /dev/hda9 21164916 4287636 15802132 22% / > <br>> > /dev/hda8 101089 17303 78567 19% /boot > <br>> > none 256456 0 256456 0% /dev/shm > <br>> > /dev/hda6 10231392 106896 10124496 2% /home/pat/Data > <br>> > > <br>> > I apologize for my lack of Linux knowledge (ie, command-line ignorance) > <br>> > -- but I don't even know how to check var/log/messages. > <br>> > > <br>> > For what it's worth, I don't get any error messages booting up when RH8 > <br>> > goes through a long checklist. > <br>> > > <br>> > Pat > <br>> > > <br>> > Pat > <br>> > > <br>> > On Sun, 2003-02-02 at 17:08, Bogi wrote: > <br>> > > Hi Pat. > <br>> > > The crashing rate is above avarage for sure. > <br>> > > I would suggest the following barring faulty ram and or faulty hdd. > <br>> > > See if you have enough diskspace, browsers tend to use up (large) amounts > <br>> > > of diskspace to cache webstuff you access. > <br>> > > df command should do the trick, If your home partition , if you have one > <br>> > > seperate, and you should, must have atleast 100 megs free, mozilla for > <br>> > > starters defaults to 50 megs for chache size, the others are not mutch > <br>> > > different. > <br>> > > If you have a single partition for your home and root (/) , the exhustion > <br>> > > of disk space could lead to unpredictabel behaviour (crash) when (/) gets > <br>> > > full. also check /var/log/messages . This is the (main) logfile, see if > <br>> > > there are any error messages, they are usually very helpful in situations > <br>> > > like yours. Changing your hardware would most definately improve on the > <br>> > > situation. Most of the (built-in) cards are slightly different then their > <br>> > > stand-alone counterparts, and not being able to turn-off / diable yours > <br>> > > is cairtinly a very strong indication of your problems. > <br>> > > So my initial suggestion is a log reading, find some error messagegs > <br>> > > relating to graphics_display_adapter or anything else. I have seen really > <br>> > > strange behaviour from motherboards like yours. > <br>> > > Cheers > <br>> > > Szemir > <br>> > > > <br>> > > On Sunday 02 February 2003 16:42, you wrote: > <br>> > > > I've been using Red Hat 8.0 on a Dell Dimension 2350 (2GH Celeron / > <br>> > > > 512MB RAM) since early January. Bottom-line impression: I love Linux. > <br>> > > > > <br>> > > > But there's one problem I'd love to resolve. Linux has crashed while > <br>> > > > surfing the Web, on average, at least every two hours (during at least > <br>> > > > 50 hours total browsing). > <br>> > > > > <br>> > > > Konqueror seems more stable than Galeon and Opera, and all seem > <br>> > > > somewhat more stable than Mozilla. There's little consistency. Mozilla, > <br>> > > > for example, crashed on the third page I accessed today (Yahoo News); > <br>> > > > but when I rebooted I surfed that page and many others for more than an > <br>> > > > hour without hanging. > <br>> > > > > <br>> > > > Meanwhile, I've worked more than 40 hours in OpenOffice Writer and can > <br>> > > > only recall it hanging once. Evolution rarely hangs and the CD player > <br>> > > > (kscd) hasn't crashed yet > <br>> > > > > <br>> > > > To compare OSes on the same machine, I've spent about 15 hours surfing > <br>> > > > in Windows 2000 (using Opera, Internet Explorer and Phoenix) and had > <br>> > > > only two crashes -- a much better average than Linux. In each OS, > <br>> > > > control-alt-delete won't force a reboot after a crash. > <br>> > > > > <br>> > > > I usually work in KDE, but didn't document any improvement during about > <br>> > > > four hours in GNOME (on Mozilla, Opera, Galeon and Konqueror). > <br>> > > > > <br>> > > > Another quirk in Linux: the hard drive sometimes runs for three minutes > <br>> > > > for no reason. > <br>> > > > > <br>> > > > Also, Evolution is very slow (maybe 10 seconds) opening HTML e-mails. > <br>> > > > Is this normal, or possibly related to my Web-browsing problem? > <br>> > > > > <br>> > > > BACKGROUND: Inside the Dell box, there are NO CARDS. Everything -- > <br>> > > > video controller, network controller, etc. -- is embedded in the > <br>> > > > motherboard. (Luckily there were three empty slots.) The video > <br>> > > > controller -- which wouldn't work with Linux -- could NEITHER BE > <br>> > > > REMOVED NOR DISABLED. > <br>> > > > > <br>> > > > As a non-technical person, I was in over my head long before this > <br>> > > > point. I'm much obliged to CLUG member Johnny Stork for coming up with > <br>> > > > the various workarounds that enabled me to use Linux my new computer. > <br>> > > > (I just wish I'd talked to him before buying the Dell.) > <br>> > > > > <br>> > > > The workaround Johnny came up for the video controller was to set up a > <br>> > > > dual-monitor system and to designate the new video card (HIS TNT2 M64 > <br>> > > > PCI) as primary and the built-in video controller as secondary. (So the > <br>> > > > original controller which couldn't be disabled is now outputting to a > <br>> > > > non-existent monitor.) > <br>> > > > > <br>> > > > Does anyone have any suggestions? Do I have any options beyond selling > <br>> > > > the new $800 Dell at a discount to my brother-in-law (who prefers XP) > <br>> > > > and starting from scratch on another new computer?? > <br>> > > > > <br>> > > > Any comments or anecdotes (e.g., your own experience with bundled, > <br>> > > > brand-name systems, both positive and negative) will be much > <br>> > > > appreciated. > <br>> > > > > <br>> > > > Thanks for reading. Sorry it's so long. > <br>> > > > > <br>> > > > Pat > <br> > <br> > <br> > <br> > > --------------------------------------------------- > The Webtop Personal Web-based Email System > http://www.yourwebtop.com >
