Re: [Jfs-discussion] Journal Corrupt, Logredo Failed.

[Dave K. <dav...@or...>]

On 09/17/2013 04:08 PM, Adam Crane wrote:
> hey,
> I'm not entirely sure if this is the correct place to ask for help, but
> I've been looking around and can't find anywhere else. I'm really at a
> loss and could use some help. The system crashed and upon reboot I can't
> get the drive to mount. The system is running off of a Debian based
> operating system, OpenMediaVault.

Have you tried mounting the volume read-only (mount -oro)? jfs will try
to mount a dirty volume read-only, but will not allow it to be mounted
read-write until fsck succeeds against it. If a read-only mount works,
it may be best to back up what you can before trying to fix it again.

> Linux Kernel: 2.6.32-5-amd64
> Problem array: 8 x 3tb sata raid5
> Array is a single Partition using the entire volume. Currently there is
> about 7TBs of data.
> 
> jfs_tune -l /dev/sdb1
> jfs_tune version 1.1.13, 17-Jul-2008
> 
> JFS filesystem superblock:
> 
> JFS magic number:    'JFS1'
> JFS version:        1
> JFS state:        dirty
> JFS flags:        JFS_LINUX  JFS_COMMIT  JFS_GROUPCOMMIT  JFS_INLINELOG 
> Aggregate block size:    4096 bytes
> Aggregate size:        40956038120 blocks
> Physical block size:    512 bytes
> Allocation group size:    67108864 aggregate blocks
> Log device number:    0x811
> Filesystem creation:    Sat Oct  6 15:45:18 2012
> Volume label:        'satisfactio'
> 
>  jfs_logdump -a /dev/sdb1
> jfs_logdump version 1.1.13, 17-Jul-2008
> Device Name: /dev/sdb1
> LOGREDO:  The Journal Log has wrapped. [logredo.c:1339]
> LOGREDO:   logRead: Log wrapped over itself (lognumread = (d) 8191).
> [log_read.c:377]
> log read failed 0x4c0c9c
> JFS_LOGDUMP: The current JFS log has been dumped into ./jfslog.dmp
> root@akakios:/home/ftp# cat ./jfslog.dmp
> JOURNAL SUPERBLOCK:
> ------------------------------------------------------
>    magic number: x 87654321
>    version     : x 1
>    serial      : x e
>    size        : t 8192 pages (4096 bytes/page)
>    bsize       : t 4096 bytes/block
>    l2bsize     : t 12
>    flag        : x 10200900
>    state       : x 0
>    end         : x 18277a0
> 
> I have the full log available if need be.
> 
> When Running fsck I get:
> 
>  fsck.jfs -v /dev/sdb1
> fsck.jfs version 1.1.13, 17-Jul-2008
> processing started: 9/12/2013 12.25.19
> Using default parameter: -p
> The current device is:  /dev/sdb1
> Open(...READ/WRITE EXCLUSIVE...) returned rc = 0
> Primary superblock is valid.
> The type of file system for the device is JFS.
> Block size in bytes:  4096
> Filesystem size in blocks:  5119669248
> **Phase 0 - Replay Journal Log
> LOGREDO:   doAfter: updatePage failed.  (logaddr = 0x04c1234, rc = (d) 274)
> LOGREDO:  Invalid RedoPage record at 0x04c1234.
> logredo failed (rc=-274).  fsck continuing.
> **Phase 1 - Check Blocks, Files/Directories, and  Directory Entries
> File system object DF1702687 has corrupt data (9).
> File system object DF1702721 has corrupt data (9).
> File system object DF1702728 has corrupt data (9).
> File system object DF1702730 has corrupt data (9).
> File system object DF2393002 has corrupt data (9).
> .......|
> Then it just continues. 22 data corrupt alerts, then just the | moving
> back and forth. It has so far been running for 2 days and is still on
> Phase 1 and no more corrupt data messages, since the first 30 minutes of
> fsck.

That doesn't seem right. You have an old version of jfsutils, not that
it has been updated in a while, but version 1.15 might be worth trying.
You can find the latest at http://jfs.sourceforge.net/

> 
> I was looking into moving the journal to an external journal but wasn't
> sure if that would work and what the affects would be if it didn't work
> and how much data I would loose.
> I do not mind loosing a few hours to a days worth of data. I can recover
> that.
> 
> If someone could help provide some steps or solutions I would very grateful.
> Thank You.

If jfsutils 1.15 fails in the same way, I can try to figure out what's
wrong. I'm not sure I'd find a way to get it to fix the volume or not.

Good luck,
Shaggy