Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Lockup on heavy IO
View unanswered posts
View posts from last 24 hours

 
Reply to topic    Gentoo Forums Forum Index Kernel & Hardware
View previous topic :: View next topic  
Author Message
ikke
Apprentice
Apprentice


Joined: 14 Jan 2005
Posts: 225
Location: Belgium

PostPosted: Thu Sep 29, 2005 2:15 pm    Post subject: Lockup on heavy IO Reply with quote

Hija,

I got a terribly annoying problem here with my machine. It locks up whenever I do a lot of hard-drive IO.

It is not a "new" problem, happens since I got this machine from my dad and removed his Windows in favor of Gentoo. First I used the ATA driver for my chipset (piix), so I got DMA on my hard-drive's IDE channel. When I started to notice the hangups, I dropped the specific driver and only started using the generic IDE driver. This one hangs my machine too, but after a longer period (as if it hangs after a specific number of bytes is read/written).

It happens at irregular intervals: sometimes I can untar a 100Mb tar.gz and do even more, sometimes it hangs while emerging something as small as, say, gaim.
It is really annoying as when it hangs, I can only use the magic SysRq, although sometimes that doesnt help/work either.

I never saw any kernel oops, as normally I work in X, and when the machine hangs, I cant switch to tty12 (where syslog-ng dumps its messages).
At first I suspected inotify, which I used long before it was in -mm. Removing that patch from my kernel tree doesnt help though.

I really don't know what's causing this. Is it a software issue? My IDE controller? My hard drive? RAM? The machine does pass an extensive memtest86+ testrun.
I hate these lockups, not because of the time I loose, but mostly because when I reboot, my gaim buddies.xml is screwed, my firefox cookies,...

I'm running this setup:
- P4 1.9Ghz, 256Mb RAMBUS RAM
- hda: MAXTOR 6L040J2
- I'm running an LVM2 system (everything except /boot and /), using XFS on all LVM volumes. I don't think it is XFS related though, as it also happens when I create a test ext3 volume and untar something in there.
- IDE interface: Intel Corporation 82801BA IDE U100 (rev 04)

Is there anything I can use to figure out what's causing this, some test tool,... Or is it possible to capture kernel oops messages somehow? (if at least I get an oops).

I'd really like to get rid of this. If it's just my harddrive, I can replace that one. Software problems can be fixed too, but I can hardly replace my IDE controller ;)

Thanks!
_________________
Working day and night to enhance your Linux Desktop experience :)
Blog
Back to top
View user's profile Send private message
potatoface
Guru
Guru


Joined: 18 Feb 2004
Posts: 542
Location: ::7F00:1

PostPosted: Thu Sep 29, 2005 3:05 pm    Post subject: Reply with quote

the only thing that comes to my mind is that your /tmp partition is too small, so that it gets full when you untar a file.
i hope i could help you.
post again if it didnt solve the problem.
_________________
The generation of random numbers is too important to be left to chance.
Adopt an unanswered post.
Back to top
View user's profile Send private message
ikke
Apprentice
Apprentice


Joined: 14 Jan 2005
Posts: 225
Location: Belgium

PostPosted: Thu Sep 29, 2005 4:01 pm    Post subject: Reply with quote

This doe not only happen when untarring something, sorry for the confusion, but also eg if I open a really large audio file (liveset) or so.
Even if /tmp is too small, the machine shouldnt hardlock, right?

Code:
/dev/mapper/nvg-tmp    1043776    881744    162032  85% /tmp

_________________
Working day and night to enhance your Linux Desktop experience :)
Blog
Back to top
View user's profile Send private message
ikke
Apprentice
Apprentice


Joined: 14 Jan 2005
Posts: 225
Location: Belgium

PostPosted: Sun Oct 09, 2005 4:14 pm    Post subject: Reply with quote

I was able to untar and remove the Xorg 6.8 tree 7 times in one session using the CLI version of 7zip under Windows, without any lockup, so this *must* be a software problem :(
I'll try to untar it under Linux using several kernels (IO shedulers), on real partitions and LVM ones, using diffirent filesystems (used fat32 under windows),... to try to figure out what's causing this. Guess I'll better go back to LKML if I get some more info?
I ran memtest86+ again too, no errors after 2 hours (10 test rounds)

Thanks for all input,

Ikke
_________________
Working day and night to enhance your Linux Desktop experience :)
Blog
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Kernel & Hardware All times are GMT
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum