Subject: [netatalk-admins] strange errors
From: Colin Postel (cpostel@primo.com)
Date: Thu Jan 21 1999 - 14:43:28 EST
I am sure there is a better mailing list to post this to, maybe somebody
has an idea which one?
I run netatalk+asun-latest on a Dell PII, internal IDE, 45GB raid array,
linux 2.0.36 (based on an old RH distro). At seemingly random intervals,
maybe twice a day, the machine spits out an error like this:
Jan 19 10:00:02 mayhem kernel: swap_duplicate: weirdness, entry f0000000
Jan 19 10:00:02 mayhem kernel: swap_free: weirdness
These errors get progressively more frequent, and after about a week I get
this:
Dec 29 15:30:34 mayhem kernel: general protection: 0000
Dec 29 15:30:34 mayhem kernel: CPU: 0
Dec 29 15:30:34 mayhem kernel: EIP: 0010:[mark_buffer_uptodate+20/76]
Dec 29 15:30:34 mayhem kernel: EFLAGS: 00010282
Dec 29 15:30:34 mayhem kernel: eax: 00000000 ebx: 0739cd98 ecx:
0739cd98 e
dx: 9739cd18
Dec 29 15:30:34 mayhem kernel: esi: 02334800 edi: 02334c00 ebp:
001208da e
sp: 03153e1c
Dec 29 15:30:34 mayhem kernel: ds: 0018 es: 0018 fs: 002b gs: 002b
ss: 0
018
Dec 29 15:30:34 mayhem kernel: Process afpd (pid: 20819, process nr: 18,
stackpa
ge=03153000)
Dec 29 15:30:34 mayhem kernel: Stack: 0015d5f8 0739cd98 00000001 00000000
001208
da 00000001 062e3700 00000400
Dec 29 15:30:34 mayhem kernel: 03282a18 00000400 0015dbec 062e3700
001208
da 03153ef4 00000008 000000aa
Dec 29 15:30:34 mayhem kernel: 00000001 062e3700 00000002 06c57aa8
0015de
75 062e3700 07459f18 000000aa
Dec 29 15:30:34 mayhem kernel: Call Trace: [ext2_alloc_block+236/412]
[rw_swap_p
age+386/736] [block_getblk+348/612] [rw_swap_page+386/736]
[ext2_getblk+385/556]
[ext2_file_write+389/1116] [kfree_skbmem+67/80]
Dec 29 15:30:34 mayhem kernel: [kfree_skb+235/244] [dev_kfree_skb+62/76]
[3c59x+8114/16384] [do_aic7xxx_isr+98/116] [sys_write+339/396]
[IRQ10_interrupt+
95/132] [system_call+85/124]
Dec 29 15:30:34 mayhem kernel: Code: f6 42 14 01 74 31 8b 52 10 85 d2 74 04
39 c
a 75 ef 8b 41 24
If I let it slide for a day it starts to corrupt the filesystem and I get
errors along these lines:
Jan 8 10:24:02 mayhem kernel: attempt to access beyond end of device
Jan 8 10:24:02 mayhem kernel: 08:04: rw=1, want=280748618, limit=26780355
Jan 8 10:24:06 mayhem kernel: EXT2-fs error (device 08:01): ext2_readdir:
bad e
ntry in directory #6145: rec_len is too small for name_len - offset=340,
inode=2
74506, rec_len=28, name_len=49169
I can unmount the filesystems and run fsck, and reboot the system, and
everything is okay for another day or another week, but then it starts all
over. Another note: one time when I was rebooting a few times in a row, the
system halted in the middle of loading the kernel, just before mounting
filesystems. The error just before this was something about the internal
IDE drive, then it said "reset successful" and froze. I can't duplicate
this one.
I doubt this has anything to do with atalk. In fact it looks like a
hardware problems.. bad disk, RAM, mainboard, something? The system is
about a year old. Odd? I think.
Thanks for any help you may be able to offer..
Colin Postel, System Administrator
Primo Angeli, Inc.
cpostel@primo.com
415 551 1900 ext. 218
This archive was generated by hypermail 2b28 : Sat Dec 18 1999 - 16:16:14 EST