Fedora 9 constant crashing.

cdb931 · 06-12-2008, 03:12 PM

My distro of fedora 9 is constantly crashing. It seems completely random. When it crashes, I can't get on to the internet unless I reset the router, restart my computer a few times, rinse and repeat until finally it randomly works. When it crashes, I can't move the mouse, and if I press num lock on my keyboard the light stays on. It is like a complete freeze. The programs I have running :
amsn, azureus, and firefox.
The patterns I've seen involve the internet. That is if I open firefox the PC might crash. There is no risk while writing this, for it seems it only crashes when I try to do something. It also crashes if I try updating steam via wine. It gets to 27 percent and crashes. I tried uninstalling CS source cause it would never load, but going yum remove wine did nothing. Also in apps>wine>programs>valve>steam> uninstall steam does nothing. It says it's deleting it but nothing happens.
I don't know why this is going on. I wish I could get a bug report but the thing completely locks up and I have to power off. This didn't happen with fedora 8. It's also only been going on since I tried getting rid of WINE and cs source.

sycamorex · 06-12-2008, 04:47 PM

Can it be connected with your X driver? Which gfx driver are you using? Perhaps, disable desktop effects (3d, compiz etc) and we'll see what happens.

cdb931 · 06-13-2008, 01:33 AM

Nvidia livna one.. Yeah I have desktop effects disabled. Actually, last time it crashed it was while waiting for the tooltip to show for some random application I have installed(yeah, the one that only takes a second while holding the cursor over it). After it crashes my net was always screwed, but it seems that if I leave my PC off for a minute then turn it back on my net works fine.

The_Trooper · 06-13-2008, 02:39 PM

You might have the same problem I did, My computer case didn't provide sufficient cooling.

I had random crashes occurring, at first I thought it was KDE's fault so I switched to gnome but the crashes continued on until I got a cooler case.

DteK · 06-24-2008, 06:30 PM

I am having the same issue here. Was working fine up until this morning, now it just keeps freezing. Was running azureus and avidemux but I have stopped running those while I troubleshoot and the issue persist.

Linux tek420.dtek360 2.6.25.6-55.fc9.i686 #1 SMP Tue Jun 10 16:27:49 EDT 2008 i686 i686 i386 GNU/Linux

for me it seems to happen when my disk is reading or writing, as my hard drive light is lit everytime during the crash. but not during every read write action, just randomly.

all my logs are clean, showing no errors, they just stop at the time of the lockup and resume after I do a hard reboot.

I have been using the same case for some time so I dont think its a cooling issue although it has been hotter than normal. running gkrellm to see what the temp is at crash time.

EDIT:
Ok so I am getting errors in /var/log/messages before every freeze.

Jun 24 19:17:26 tek420 kernel: ata5.00: exception Emask 0x10 SAct 0x0 SErr 0x380100 action 0x2
Jun 24 19:17:26 tek420 kernel: ata5.00: port_status 0x20400000
Jun 24 19:17:26 tek420 kernel: ata5: SError: { UnrecovData 10B8B Dispar BadCRC }
Jun 24 19:17:26 tek420 kernel: ata5.00: cmd c8/00:00:07:5b:1b/00:00:00:00:00/eb tag 0 dma 131072 in
Jun 24 19:17:26 tek420 kernel: res d0/00:00:07:5b:1b/00:00:00:00:00/eb Emask 0x32 (host bus error)
Jun 24 19:17:26 tek420 kernel: ata5.00: status: { Busy }
Jun 24 19:17:31 tek420 kernel: ata5: port is slow to respond, please be patient (Status 0xd0)
Jun 24 19:17:36 tek420 kernel: ata5: device not ready (errno=-16), forcing hardreset
Jun 24 19:17:36 tek420 kernel: ata5: hard resetting link
Jun 24 19:17:36 tek420 kernel: ata5: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
Jun 24 19:17:36 tek420 kernel: ata5.00: configured for UDMA/133
Jun 24 19:17:36 tek420 kernel: ata5: EH complete
Jun 24 19:18:36 tek420 kernel: ata5.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x2 frozen
Jun 24 19:18:36 tek420 kernel: ata5.00: cmd c8/00:00:07:5b:1b/00:00:00:00:00/eb tag 0 dma 131072 in
Jun 24 19:18:36 tek420 kernel: res 40/00:00:07:5b:1b/00:00:00:00:00/eb Emask 0x4 (timeout)
Jun 24 19:18:36 tek420 kernel: ata5.00: status: { DRDY }

Ideas any one?
I have 1 250GB WD and 3 500GB WD hard drives all under 3 months old.

DteK · 06-26-2008, 08:24 AM

I think I have traced my issue to a shady sata cable. Since I was able to replicate the freezing issue by encoding a video with avidemux, I was using that to get to the root of what seemed to by my hard drive issues. I have switched out the cable and now my issues have gone away so far.

shafty023 · 06-26-2008, 09:49 AM

I'd recommend to anyone having these issues to run a hard drive test.

Download and burn the iso, pop in your cdrom and reboot. Go through all the prompts and choose advanced test. Let it test your hard drive.
This utility will test ALL hard drives, not just IBM ones. I've used this utility for over 8 years now.
http://www.hitachigst.com/hdd/suppor...2_v413_b00.iso

DteK · 06-28-2008, 08:17 PM

confirming sata cable was at the root of my problems.
Thanks for that tool shafty

DteK · 07-07-2008, 08:41 PM

OK confirming the cable was not the root cause.

It worked for some time after changing the cable, but the issue
arose again after a day. I eventually decided it had to be the last hard drive I added
as all was good before then.

I was running 3 with no issue, when I added the 4th is when the problems started. I guess my power supply
could not handle it, was running a 500W powtek. I switched it out to a 600W Aspire
and my PC is good as new and has been since.

Looking on upgrading to a more robust and better quality power supply now as this
has taught me a valuable lesson about using quality hardware for the less glamorous
parts of my PC.

ragnorok · 09-12-2008, 10:15 AM

Quote:

Originally Posted by DteK

for me it seems to happen when my disk is reading or writing, as my hard drive light is lit everytime during the crash. but not during every read write action, just randomly.

- I'm seeing this as well ... random crashes during disk operations, mainly via samba, but I'm not convinced samba is the cause. The messages log appears to have a line like this around the time of the crash:

Code:

Sep 11 16:01:37 localhost console-kit-daemon[1886]:
 WARNING: Couldn't read /proc/6662/environ:
 Error reading file '/proc/6662/environ': No such process

- The number between proc/nnn/environ changes, but a line like this appears consistently.
- This is a file server running FC9 that I'm using for backup. It gets some variable amount into the backup, then the whole system freezes tight, with the backup drive light on. I haven't tried the numlock key, but it's utterly unresponsive otherwise. Can't even putty in to have a look around. It's very unlike Linux to crash like this (at least in the last 15 years of using it), much less crash so hard I can't even putty in.
- The drive is in a separate enclosure and accessed via USB. It's not a drive that was bought as an USB unit; I had this in another computer that tossed its cookies, so I got an "iRocks" external drive enclosure and put the hard drive in there, then plugged it into this machine. Since it was already formatted ext3 this box recognized and accesses it without any apparent trouble.
- This FC9 install is new as of last week, built specifically for fiddling with embedded Linux development, but put into service as a file server because the other one died. The install and configuration of this system was a low priority until it was needed as a temporary file server. (grin)
- Any ideas? I'm not a newbie to Linux, but I am a newbie to Fedora Core Linux of any variant. The server that died was running RH9 and I've had not need to "fix" it; it's been fire and forget for years and years.
- Thanks for your time...