kern/135690: [panic] [ata] ufs_dirbad: /backuphd: bad dir ino 22259126 at offset 0: mangled entry

Thu Jun 18 21:00:14 UTC 2009

The following reply was made to PR kern/135690; it has been noted by GNATS.

From: =?utf-8?Q?Derek_Kuli=C5=84ski?= <takeda at takeda.tk>
To: bug-followup at FreeBSD.org, takeda at takeda.tk
Cc: Volker Werth <vwe at FreeBSD.org>
Subject: Re: kern/135690: [panic] [ata] ufs_dirbad: /backuphd: bad dir ino 22259126 at offset 0: mangled entry
Date: Thu, 18 Jun 2009 13:57:53 -0700

 Looks like my original e-mail with the extensive tests reached the
 destination.

 I haven't noticed anything new in the SMART output, except those
 additional tests which passed successfully.
 Raw_Read_Error_Rate, Reallocated_Sector_Ct, Seek_Error_Rate,
 Spin_Retry_Count, Calibration_Retry_Count, Reallocated_Event_Count,
 Current_Pending_Sector, Offline_Uncorrectable, UDMA_CRC_Error_Count
 and Multi_Zone_Error_Rate are all 0.

 If the controler would be broken do you think it would show in SMART?
 I'm asking since the controller is on the disk (it's an IDE disk).

 Is there a good way to check the cable besides replacing it with a new
 one and seeing if the problem persists?

 I haven't notice this in my logfiles:

 First crash:
 Jun 17 08:43:28 chinatsu sshguard[90671]: Blocking 202.104.3.83: 4 failures over 0 seconds.
 Jun 17 09:30:48 chinatsu kernel: ad2: TIMEOUT - READ_DMA48 retrying (1 retry left) LBA=344820288
 Jun 17 09:30:54 chinatsu kernel: ad2: TIMEOUT - READ_DMA48 retrying (0 retries left) LBA=344820288
 Jun 17 09:35:05 chinatsu syslogd: kernel boot file is /boot/kernel/kernel
 Jun 17 09:35:05 chinatsu kernel: Copyright (c) 1992-2009 The FreeBSD Project.
 Jun 17 09:35:05 chinatsu kernel: Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994
 [...]
 Jun 17 09:35:05 chinatsu kernel: GEOM_JOURNAL: Journal 56801142: ad2 contains data.
 Jun 17 09:35:05 chinatsu kernel: GEOM_JOURNAL: Journal 56801142: ad2 contains journal.
 Jun 17 09:35:05 chinatsu kernel: GEOM_JOURNAL: Journal ad2 consistent.
 Jun 17 09:35:05 chinatsu kernel: Trying to mount root from ufs:/dev/ad0s1a
 Jun 17 09:35:05 chinatsu kernel: WARNING: / was not properly dismounted
 Jun 17 09:35:05 chinatsu savecore: reboot after panic: ufs_dirbad: /backuphd: bad dir ino 21573632 at offset 0: mangled entry
 Jun 17 09:35:05 chinatsu savecore: no dump, not enough free space on device (110564 available, need 118358)
 Jun 17 09:35:05 chinatsu savecore: unsaved dumps found but not saved

 I didn't have enough space on /var

 Second one (no signs of DMA errors or anything):
 Jun 17 09:40:38 chinatsu named[833]: transfer of 'pckrzyz.pl/IN' from 83.14.32.234#53: failed to connect: timed out
 Jun 17 09:43:24 chinatsu named[833]: transfer of 'pckrzyz.pl/IN' from 83.14.32.234#53: failed to connect: timed out
 Jun 17 09:46:29 chinatsu named[833]: transfer of 'pckrzyz.pl/IN' from 83.14.32.234#53: failed to connect: timed out
 Jun 17 09:52:31 chinatsu named[833]: transfer of 'pckrzyz.pl/IN' from 83.14.32.234#53: failed to connect: timed out
 Jun 17 10:06:28 chinatsu named[833]: transfer of 'pckrzyz.pl/IN' from 83.14.32.234#53: failed to connect: timed out
 Jun 17 10:35:55 chinatsu named[833]: transfer of 'pckrzyz.pl/IN' from 83.14.32.234#53: failed to connect: timed out
 Jun 17 10:59:51 chinatsu syslogd: kernel boot file is /boot/kernel/kernel
 [...]
 Jun 17 10:59:51 chinatsu kernel: GEOM_JOURNAL: Journal 56801142: ad2 contains data.
 Jun 17 10:59:51 chinatsu kernel: GEOM_JOURNAL: Journal 56801142: ad2 contains journal.
 Jun 17 10:59:51 chinatsu kernel: GEOM_JOURNAL: Journal ad2 consistent.
 Jun 17 10:59:51 chinatsu kernel: Trying to mount root from ufs:/dev/ad0s1a
 Jun 17 10:59:51 chinatsu kernel: WARNING: / was not properly dismounted
 Jun 17 10:59:51 chinatsu savecore: reboot after panic: ufs_dirbad: /backuphd: bad dir ino 22259126 at offset 0: mangled entry
 Jun 17 10:59:51 chinatsu savecore: writing core to vmcore.0
 Jun 17 11:00:07 chinatsu named[829]: starting BIND 9.4.2-P2 -t /var/named -u bind
 Jun 17 11:00:08 chinatsu named[829]: command channel listening on 127.0.0.1#953
 Jun 17 11:00:08 chinatsu named[829]: command channel listening on ::1#953

 Third one:
 Jun 17 11:37:51 chinatsu sshguard[1145]: Blocking 59.124.109.227: 4 failures over 740 seconds.
 Jun 17 12:03:16 chinatsu named[829]: transfer of 'pckrzyz.pl/IN' from 83.14.32.234#53: failed to connect: timed out
 Jun 17 12:41:26 chinatsu kernel: ad2: TIMEOUT - READ_DMA48 retrying (1 retry left) LBA=344820316
 Jun 17 12:41:31 chinatsu kernel: ad2: TIMEOUT - READ_DMA48 retrying (0 retries left) LBA=344820316
 Jun 17 12:45:26 chinatsu syslogd: kernel boot file is /boot/kernel/kernel
 Jun 17 12:45:26 chinatsu kernel: ad2: FAILURE - READ_DMA48 timed out LBA=344820316
 Jun 17 12:45:26 chinatsu kernel: GEOM_JOURNAL: Error while reading data from ad2 (error=5).
 Jun 17 12:45:26 chinatsu kernel: panic: ufs_dirbad: /backuphd: bad dir ino 21573639 at offset 512: mangled entry
 Jun 17 12:45:26 chinatsu kernel: KDB: stack backtrace:
 Jun 17 12:45:26 chinatsu kernel: db_trace_self_wrapper(c0809ee8,c0882620,c081def3,d1372930,d1372930,...) at db_trace_self_wrapper+0x26
 Jun 17 12:45:26 chinatsu kernel: panic(c081def3,c271e4dc,1493007,200,c081dfb1,...) at panic+0xf8
 Jun 17 12:45:26 chinatsu kernel: ufs_dirbad(c2d20180,200,c081dfb1,0,d13729cc,...) at ufs_dirbad+0x73
 Jun 17 12:45:26 chinatsu kernel: ufs_lookup(d13729f8,d13729f8,d1372bcc,d1372bb8,c2b91d00,...) at ufs_lookup+0x4bd
 Jun 17 12:45:26 chinatsu kernel: vfs_cache_lookup(d1372a84,c05f5271,2,c2d1f8a0,d1372aa4,...) at vfs_cache_lookup+0xf2
 Jun 17 12:45:26 chinatsu kernel: VOP_LOOKUP_APV(c085a940,d1372a84,d1372bcc,c080fb81,2a9,...) at VOP_LOOKUP_APV+0x3d
 Jun 17 12:45:26 chinatsu kernel: lookup(d1372ba4,c2b3f000,0,d1372bc0,c2d22088,...) at lookup+0x50f
 Jun 17 12:45:26 chinatsu kernel: namei(d1372ba4,d1372b44,60,0,c288a8c0,...) at namei+0x3a8
 Jun 17 12:45:26 chinatsu kernel: kern_stat(c288a8c0,bfbfe514,0,d1372c14,52,...) at kern_stat+0x3d
 Jun 17 12:45:26 chinatsu kernel: stat(c288a8c0,d1372cf8,8,c05a18b5,c288a8c0,...) at stat+0x2f
 Jun 17 12:45:26 chinatsu kernel: syscall(d1372d38) at syscall+0x208
 Jun 17 12:45:26 chinatsu kernel: Xint0x80_syscall() at Xint0x80_syscall+0x20
 Jun 17 12:45:26 chinatsu kernel: --- syscall (188, FreeBSD ELF32, stat), eip = 0x206fa46b, esp = 0xbfbfca7c, ebp = 0xbfbfca98 ---
 [...]
 Jun 17 12:45:26 chinatsu kernel: Trying to mount root from ufs:/dev/ad0s1a
 Jun 17 12:45:26 chinatsu kernel: WARNING: / was not properly dismounted
 Jun 17 12:45:27 chinatsu savecore: reboot after panic: ufs_dirbad: /backuphd: bad dir ino 21573639 at offset 512: mangled entry
 Jun 17 12:45:27 chinatsu savecore: writing core to vmcore.1
 Jun 17 12:45:45 chinatsu named[829]: starting BIND 9.4.2-P2 -t /var/named -u bind
 Jun 17 12:45:45 chinatsu named[829]: command channel listening on 127.0.0.1#953
 Jun 17 12:45:45 chinatsu named[829]: command channel listening on ::1#953

 As for installing 7.2 I can start doing it. Though the system was
 stable for quite some time until this begin to happen.

 Oh one more thing that might be important but I forgot about it.
 Before I start copying the data, I got an error while using a ls.
 ls was reporting that one directory entry is invalid or something like
 that (it was two days ago, I don't rememebr exactly) I run fsck and it
 found few errors I fixed all of them. It was kind of weird that they
 were there in the first place since journaling was used. That was
 before I experienced the crashes. Is it possible that the filesystem
 is somehow damaged there and fsck wasn't able to fix it correctly?

 As for the suggestion of changing UDMA, how can I do it?