|
From: <mar...@es...> - 2007-09-15 02:15:49
|
> Hi, > > On 14 Sep 2007, at 21:27, mar...@es... wrote: >>> On 14 Sep 2007, at 20:12, mar...@es... wrote: >>>>>> escabot CAD # ls -ltrh >>>>>> total 49G >>>>>> -r-------- 1 root root 2.0G Oct 7 2004 RTC26001.GHS >>>>>> -r-------- 1 root root 2.0G Oct 7 2004 RTC26002.GHS >>>>>> -r-------- 1 root root 2.0G Oct 7 2004 RTC2604.GHO >>>>>> -r-------- 1 root root 620M Oct 7 2004 RTC26003.GHS >>>>>> -r-------- 1 root root 4 Sep 14 09:32 asdf.txt >>>>>> -r-------- 1 root root 8 Sep 14 09:33 rty.txt >>>>>> -r-------- 1 root root 42G Sep 14 10:36 bigfile.gho >>>>>> escabot CAD # cp bigfile.gho /dev/null >>>> >>>> Whoops! Sorry, it still causes the kernel to panic when I copy to / >>>> dev/null. >>> >>> Hm. There goes my theory. Either it must be NTFS itself that is at >>> fault or it is a bug in the VFS/VM. >>> >>> Could you now try the following two things and let me know whether >>> they work/panic? >>> >>> 1) Reboot and place "mem=2G" on the kernel command line (this makes >>> the kernel use only 2G of your RAM instead of all 12G). Then try the >>> copy of bigfile.gho to /dev/null again. Does it now work or panic >>> again? >> >> It worked setting 'mem=2G' -- it copied all 42GB to /dev/null. >> Should I >> try recompiling a kernel with high memory support disabled now, or >> was the >> first test sufficient? > > Great! Thanks! The first test was sufficient. > > I think this must be a kernel memory management bug and NTFS just > happens to trigger it. Now at least you have a workaround (you can > try higher values than 2G btw - I have no idea where the magic number > will be that makes it to run out of memory). Also, given the nature > of the bug I would expect that if you upgrade to a 64-bit kernel (you > will probably need an at least partially 64-bit userspace to match) > this bug will no longer exist. Okay, I'll experiment and see what works. Thanks for your help. > > If you are happy with that, I will repost your original report with > your testing results on the kernel mailing list and CC: you so you > can answer any questions people there might come up with. Let me > know if you are happy for me to do this and I will do it tomorrow. > Or alternatively just send it yourself... (I am off to bed...) Yeah, I would appreciate that. Thanks again for all your help Anton. --Marc > > Best regards, > > Anton > -- > Anton Altaparmakov <aia21 at cam.ac.uk> (replace at with @) > Unix Support, Computing Service, University of Cambridge, CB2 3QH, UK > Linux NTFS maintainer, http://www.linux-ntfs.org/ > > |
