Assertion failure in journal_start() (ext3/quota related?) killssystem
From: James Renken
Date: Thu Oct 28 2004 - 22:07:56 EST
My SMP production machine, which has been running for about a year, has
wedged twice within the past week. It doesn't die completely, but new
processes run impossibly slowly and the system is unusable until a
hardware reboot. I don't have any details of the first wedge, but the
second one returned the trace and associated ksymoops below.
This is running stock 2.4.27 plus the 2.4.27-ow1 patch from
http://www.openwall.com/linux/. My drives are all on a 3ware RAID card,
using the 3w-xxxx 1.02.00.037 driver. I have five partitions, all ext3,
and three have quotas enabled; these three are using LVM. The quota
version reported by the kernel is vdquot_6.5.1, and my quota utils are
version 3.09 via backports.org (Debian). I have CONFIG_QUOTA and
CONFIG_QFMT_V2 set to Y; EXT3_FS is also Y.
This is a newish SuperMicro dual-Xeon server with ECC memory, so I am
fairly sure that this is not a random hardware problem nor a massive I/O
overload.
-----
Oct 28 16:20:33 herring kernel: Assertion failure in journal_start() at transaction.c:251: "handle->h_transaction->t_journal == journal"
Oct 28 16:21:15 herring kernel: kernel BUG at transaction.c:251!
Oct 28 16:21:40 herring kernel: invalid operand: 0000
Oct 28 16:21:42 herring kernel: CPU: 0
Oct 28 16:21:55 herring kernel: EIP: 0010:[journal_start+80/196] Not tainted
Oct 28 16:21:55 herring kernel: EFLAGS: 00010286
Oct 28 16:21:55 herring kernel: eax: 0000006c ebx: c9859a80 ecx: ffffffff edx: 00000002
Oct 28 16:21:55 herring kernel: esi: f6672c00 edi: d9874000 ebp: 00000003 esp: d9875bb0
Oct 28 16:21:55 herring kernel: ds: 0018 es: 0018 ss: 0018
Oct 28 16:21:55 herring kernel: Process ipop3d (pid: 26852, stackpage=d9875000)
Oct 28 16:21:55 herring kernel: Stack: c0280e00 c0280eec c0280de0 000000fb c0280ec0 f6619888 00000004 f65eb900
Oct 28 16:21:55 herring kernel: c016f597 f6672c00 0000001b debe4380 00000001 d42f8080 00000003 c01598e2
Oct 28 16:21:55 herring kernel: debe4380 00000000 00000001 c015a4bc debe4380 d42f8080 d42f8080 d9875c4c
Oct 28 16:21:55 herring kernel: Call Trace: [ext3_write_dquot+211/460] [dqput+90/176] [dquot_drop+56/72] [clear_inode+143/284] [dispose_list+82/196]
Oct 28 16:21:55 herring kernel: [prune_icache+189/528] [shrink_icache_memory+27/48] [shrink_cache+778/1028] [shrink_caches+60/72] [try_to_free_pages_zone+100/240] [balance_classzone+72/456]
Oct 28 16:21:58 herring kernel: [getblk+28/76] [__alloc_pages+396/652] [_alloc_pages+22/24] [do_generic_file_read+829/1168] [generic_file_read+139/396] [file_read_actor+0/224]
Oct 28 16:21:59 herring kernel: [read_blk+84/100] [find_tree_dqentry+63/204] [v2_read_dquot+70/312] [kmem_cache_alloc+116/316] [get_empty_dquot+23/172] [read_dqblk+70/120]
Oct 28 16:21:59 herring kernel: [dqget+330/464] [dquot_initialize+168/332] [ext3_new_inode+1966/2516] [ext3_new_inode+2138/2516] [journal_start+149/196] [ext3_create+130/256]
Oct 28 16:21:59 herring kernel: [vfs_create+323/412] [open_namei+324/1960] [filp_open+59/92] [sys_open+54/156] [system_call+51/56]
Oct 28 16:21:59 herring kernel:
Oct 28 16:21:59 herring kernel: Code: 0f 0b fb 00 e0 0d 28 c0 83 c4 14 ff 43 08 89 d8 eb 5b 8d b6
Using defaults from ksymoops -t elf32-i386 -a i386
>>ebx; c9859a80 <_end+94ec900/385cce80>
>>ecx; ffffffff <END_OF_CODE+758f8bc/????>
>>esi; f6672c00 <_end+36305a80/385cce80>
>>edi; d9874000 <_end+19506e80/385cce80>
>>esp; d9875bb0 <_end+19508a30/385cce80>
Code; 00000000 Before first symbol
00000000 <_EIP>:
Code; 00000000 Before first symbol
0: 0f 0b ud2a
Code; 00000002 Before first symbol
2: fb sti
Code; 00000003 Before first symbol
3: 00 e0 add %ah,%al
Code; 00000005 Before first symbol
5: 0d 28 c0 83 c4 or $0xc483c028,%eax
Code; 0000000a Before first symbol
a: 14 ff adc $0xff,%al
Code; 0000000c Before first symbol
c: 43 inc %ebx
Code; 0000000d Before first symbol
d: 08 89 d8 eb 5b 8d or %cl,0x8d5bebd8(%ecx)
Code; 00000013 Before first symbol
13: b6 00 mov $0x0,%dh
-----
Any ideas? This is mostly over my head, but I'd be happy to do any
additional debugging. Thanks in advance!
(P.S.: This is my first post to linux-kernel; my apologies if I've missed
anything important.)
--
James Renken, System Administrator jrenken@xxxxxxxxxxxx
Sandwich.Net Internet Services http://www.sandwich.net/ 1-877-HUBWICH
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/