Red Hat Bugzilla – Bug 92013
kernel BUG at vmscan.c
Last modified: 2007-04-18 12:54:10 EDT
From Bugzilla Helper: User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.0.2) Gecko/20030208 Netscape/7.02 Description of problem: This morning I couldn't log into my server (ssh), open xterm-windows and samba didn't work. The machine replied to ping, but nothing else. I was forced to boot. After that I looked into /var/log/messages (timestamp information removed): May 31 04:02:39 pullasorsa kernel: ------------[ cut here ]------------ kernel: kernel BUG at vmscan.c:118! kernel: invalid operand: 0000 kernel: loop ip_nat_irc ip_conntrack_irc ip_nat_ftp ip_conntrack_ftp autofs 3c59x ipt_REJECT ipt_limit ipt_LOG ipt_state iptable_nat ip_conntrack iptable_filter ip_ta kernel: CPU: 0 kernel: EIP: 0060:[<c0135fc0>] Not tainted kernel: EFLAGS: 00010206 kernel: kernel: EIP is at reclaim_page [kernel] 0x2a0 (2.4.20-9) kernel: eax: 010c020c ebx: c1292a48 ecx: 00000100 edx: 0000059d kernel: esi: c02ec860 edi: c1292a2c ebp: 0000052a esp: cf6cbe68 kernel: ds: 0068 es: 0068 ss: 0068 kernel: Process rpmq (pid: 6516, stackpage=cf6cb000) kernel: Stack: c02eca44 00000000 c02ec860 c02ecdc4 00000003 00000001 c0138da0 c02ecdcc kernel: 0000031f 000001d2 00000000 c0138e61 c02ecdc0 00000000 00000003 00000001 kernel: c124968c c7859148 c124968c c124968c 00000001 c02ecdc0 c7859148 0000190d kernel: Call Trace: [<c0138da0>] __alloc_pages_limit [kernel] 0x70 (0xcf6cbe80)) kernel: [<c0138e61>] __alloc_pages [kernel] 0xa1 (0xcf6cbe94)) kernel: [<c012e5b0>] page_cache_read [kernel] 0x80 (0xcf6cbed0)) kernel: [<c012edcb>] generic_file_readahead [kernel] 0xdb (0xcf6cbee8)) kernel: [<c012f253>] do_generic_file_read [kernel] 0x353 (0xcf6cbf10)) kernel: [<c012f610>] file_read_actor [kernel] 0x0 (0xcf6cbf3c)) kernel: [<c012f732>] generic_file_read [kernel] 0x82 (0xcf6cbf5c)) kernel: [<c012f610>] file_read_actor [kernel] 0x0 (0xcf6cbf6c)) kernel: [<c013ff92>] sys_pread [kernel] 0xa2 (0xcf6cbf94)) kernel: [<c0109103>] system_call [kernel] 0x33 (0xcf6cbfc0)) kernel: kernel: kernel: Code: 0f 0b 76 00 e1 f9 23 c0 e9 9f fd ff ff 8d 76 00 57 56 ba 01 kernel: ------------[ cut here ]------------ As I look into other messages file, I can see that this did happen once before, May 16. The time, messages and process are exactly the same. There is a cron job in cron.daily which caused this to happen: [root@pullasorsa cron.daily]# cat rpm #!/bin/sh rpm -qa --qf '%{name}-%{version}-%{release}.%{arch}.rpm\n' 2>&1 \ | sort > /var/log/rpmpkgs The rpm package version is: [root@pullasorsa cron.daily]# rpm -q rpm rpm-4.2-0.69 Version-Release number of selected component (if applicable): 2.4.20-9 How reproducible: Couldn't Reproduce Additional info: I had some crashes before this bug (reported as bug # 90990) so I run memtest86 for about 48 hours (all test and only standard set of tests) and it couldn't find anything. I'm not sure if this is related to it. Atleast these crashes seem to happen when there are lot of disk activity (see bug 90990: backup script, samba, rpm, updatedb). I did shutdown the machine with -F to force fsck on reboot. It didn't find anything wrong with hard drives.
Created attachment 95428 [details] vmscan.c:118 kernel panic with 2.4.20-20.9 + bootup info The very same bug did hit my friends server last night. This is what he got. It seems that there's something fishy going on with reclaim_page() ... that's my guess, anyway. My friends server has Red Hat 9.0 with latest kernel errata installed. The same system used to work without a hitch even under heavier load, but couple of days after he installed the new kernel this oops did hit. And during the last few days he even didn't stress the server.
Oh dear. I just heard that my friends server is quite a bit overclocked. The box used to run Windows and overclock settings were tuned for it ("Does this thing still boot? Good, let's O/C a bit more! Does it still boot? No... let's go down a step."), but since Linux actually uses the hardware :-), overclocking made the box very crashy. He will restore it to default settings and see if that helps to this problem. Personally I think that calming down the server will resolve this case for him.
Thanks for the bug report. However, Red Hat no longer maintains this version of the product. Please upgrade to the latest version and open a new bug if the problem persists. The Fedora Legacy project (http://fedoralegacy.org/) maintains some older releases, and if you believe this bug is interesting to them, please report the problem in the bug tracker at: http://bugzilla.fedora.us/