[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#482100: linux-image-2.6.25-3-686-bigmem : Complete system freeze



maximilian attems wrote :
that's quite sad but as aboves gives no hint on where we would perform
that badly nor why those freezes that i cant reproduce happen.
i'd have to close the bug. the only big difference i know about our
kernels is that ubuntu uses a pata driver instead of the old ide ones.
could that make the difference?

best regards



Don't get me wrong, I don't blame anyone for this problem, especially Debian which I love, but I can't really investigate anymore than I did this issue. As soon as I will get some time I will reinstall a stock Lenny Deb (as Etch won't install on those boxes) and try out a 2.6.26.* vanilla on it. I'd rather run Debian than Ubuntu but right now I have too many problems (can't upgrade kernel which is bad since it's handling my hardware better, hum, when it's not freezing I mean ;-) , had trouble with virtualbox package (both ose and sun) ending in segfault anytime I want to run a virtual machine, and finally Xorg that got crazy with my mouse : it's not responding in Opengl programs, and worst inside my virtual machines (kvm or virtualbox) ). So for now it'll be Ubuntu which is close enough in line from Debian so that I can feel at home and export most of my conf files.

Sorry I can't help better, if you want any particular output I still have an image from the Debian system on a spare drive, I can boot in and perform some tests. But I'm no developer or kernel hacker, just a rookie sysadmin. I have two identical boxes, on being live and the other a testbed for the first and backup for the first. It's running on Asus P5K-E motherboards, Intel Core2Duo 6750, 4GB of ram and sataII everything (hard drives and optical drives).

I understand there have been huge changes between 2.6.24 and 2.6.25 kernels, but still can't figure what's creating the problem, I tried a "make oldconfig" from my 2.6.24 config, thus disabling most of the new features in 2.6.25, and still I got the same problem...

I'm pretty much convince that it has something to do with the way the kernel handles the cpu, especially interrupts, but I don't have the skills to gather evidences.

Regards,

Tom



Hi, just for information, following your email yesterday I booted back the Lenny system which is having trouble with 2.6.25* kernels. I have compiled a vanilla 2.6.26-rc8 from kernel.org, using my actual config as a starter, and it's been running fine since then. I loaded all the modules I usually use (including the beta nvidia blob 177.13), two kvm machines running and it's stable. I guess there really is something special with the 2.6.25 family... Whatever it is I hope it won't show up again in the future kernels.

I cloned the same system again on a different machine (still C2duo on Asus motherboard), tried a 2.6.25-7 kernel from Debian which froze the computer after 2 hours. Then I tried the vanilla 2.6.26-rc8 and it's still running fine.

I am not the only one experiencing this problem, but I guess I am the only one who reported it...

Best regards,

Tom



Reply to: