<0>Kernel Panic - not syncing: Fatal exception in interrupt

stillnoname · n00b Joined: 01 Jan 2006 Posts: 8

Hello,

I have bootable RAID 1 setup (intel ich6). I'm new to gentoo.

i've installed dmraid, dm, etc and my setup boots fine from the RAID 1 array.

I've successfully installed samba and am planning to use the setup as mostly as a fileserver.

I get a kernel panic when I try to copy many large files (movies mostly) to the setup. the target share is a FAT32 partition on the raid array.

I've created a share and am using windows to copy files. there is no problem if i copy a file individually - even the one on which there was a kernel panic. so it is not specific to data.

i seem to get a panic after a few gigs of data transfer. i've got it twice now. i was hoping to enable some kind of monitoring, etc to do something about it. but my knowledge is very limited in this area.

I've got a photo like a screenshot of the panic message (sorry - could not think of another way - unless this data has been saved). if someone is interested i can send the photo (<200kb). if there is a way to upload it here, please let me know.

i also noticed pdflush taking a lot of cpu - even more than smbd.

i've typed some information from the screen photo. This is as follows.

widan

stillnoname · n00b Joined: 01 Jan 2006 Posts: 8

thanks for the response.

these are the urls for the images
http://img499.imageshack.us/img499/512/dsc000141be.jpg
http://img499.imageshack.us/img499/7462/dsc000173ep.jpg

I can make this happen quite regualrly. I've made it happen atleast 5 times. sometimes there is no panic information printed, but the system just crashes - no response either. surprisingly the ping to the system works (when not panic). keyboard does not work. no local login or remote ssh login.

most times it panics. other things is that it seems to work without problems when i transferred 3 gigs of digital photos and mp3 files. so it seems to hold well with small files. it seems to break down when i transfer a few gigs of movie mpeg, avi, etc. - many of the files are full cd length. i finally transferred in small batches and it worked mostly - without too many crashes. it seems to hold better if i'm reading from the disks - that slows down transfers a bit, but seems to add stability (i may be wrong here). the panic seems random - even when i deleted all and did same transfer, it paniced at different place - so it should not be specific to data. i've also read of someone else having issue with 2.6 series kernel with this (and other) version of samba - when transferring a number of large files.

fyi, i'm using linux 2.6.12-gentoo-r10 kernel (genkernel) with patch for my network driver (sk98lin). samba version is 3.0.10. if you need more version or other information

i'm using fat32 cause i want the data to be accessible from windows as well. i intend to make this a dual boot system. i'm still not so comfortable with linux to remove windows completely. may sometime down the line - if i can run it without any panics :-)

looking forward to hearing more from people.

cheers.

widan · Posted: Mon Jan 02, 2006 5:34 pm Post subject: Re: images

stillnoname · n00b Joined: 01 Jan 2006 Posts: 8

hello,

thanks for the previous reply.

i recompiled the kernel with 4k stacks disabled.

very similar result - failed. However, on the root terminal i left "vmstat 1" running. It continued running even after the server was lost on the network. previous times i had top running on the it froze. not sure this time.

so i thought maybe samba failed. ran rc-update and that showed samba started. i then issued "/etc/init.d/samba stop". then it stuck there - nothing printed. so i think the script didn;t run.

i went to second console and tried logging in as root. it printed out:

widan · i tried getting 2.6.14 from world, but it looks like it has not been release.

stillnoname · n00b Joined: 01 Jan 2006 Posts: 8

hello,

after much trial and error, i managed to update the kernel as suggested (2.6.14-gentoo-r5).

it is giving me the same problem. this time "vmstat 2" stopped running. i'm guessing that sometimes its a kernel panic, other times something else.

anything that anyone wants to suggest? what can i do to help diagnose this further?

i'm currently emerging samba-3.0.14a-r2. I'll let the forum know if it is successul. i am not too sure.

i have noted the booting from such a setup is still experimental. How can i help progress this further? any help will be appreciated.

thanks in advance

stillnoname · n00b Joined: 01 Jan 2006 Posts: 8

hi,

just to let all know that the new samba version stuck exactly like the previous version.

If anyone can shed any light on what to do (logging, etc) that would be appreciated.

thanks again.

stillnoname · n00b Joined: 01 Jan 2006 Posts: 8

can anybody help?

thanks in advance

stillnoname · n00b Joined: 01 Jan 2006 Posts: 8

hi all,

i'll give this one more try, hoping someone can help.

thanks in advance

Doctor Corndog · Posted: Tue Jan 10, 2006 1:12 am Post subject:

I'm not sure if this will help you, but I have had similar errors. When I put my computer together, I loaded up Gentoo first, so I had nothing else to compare it's performace to. I began receiving Kernel Panics whenever I would try to compile large amounts of data (happened twice about 7/10 through emerge gnome). On my system, it looked like a hardware fault right away. Happened here and there, and always under stress. I asked around a many people told me it could be a memory problem. All my memtests ran great, so I was a little confused. In the end, I bumped up the voltage to my memory to 2.7V, and I've never had a problem since. Don't know if that will help ya, but it might be worth a shot. Make sure you find out the max voltage on your memory first though.

stillnoname · n00b Joined: 01 Jan 2006 Posts: 8

hello,

thanks for that. i'll look into this as well. i've have been compiling and emerging big packages. Plus, I've not overclocked or changed any voltage anywhere. So, I'm not sure this should be the case. But probably worth trying.

I've also notices that there was some traffic in "dm-devel" mailing list about file system io locks when doing large transfers. And this was in Apr 2005. I've not been able to track this down further. I'm hoping to update the device-mapper to a later version to see if this is still a problem. I'm very new to gentoo (and linux in general), so i'm not exactly sure how to go about it. Will read up on that. I'll probably need to apply a patch to the kernel. Or wait for a new stable ebuild of gentoo sources. as of yesterday, the version I used is the latest.