Hi,
I have a JFS filesystem image that corrupted. Since this is a question
about data recovery I expect a lot "non-answers" in the line of "where's
your backup?" or "do it on block level image copies". Thanks, I know
that, you don't have to lecture me on that. See the notes at the end for
what and why. Thank you.
*jfs_fsck* refuses to work on it, bailing out with the following error
message:
~ % sudo jfs_fsck /dev/loop0
jfs_fsck version 1.1.15, 04-Mar-2011
processing started: 3/1/2017 13:08:53
Using default parameter: -p
The current device is: /dev/loop0
Superblock is corrupt and cannot be repaired
since both primary and secondary copies are corrupt.
CANNOT CONTINUE.
Using *jfs_debugfs* commands `su` and `s2p` the folloing information was
obtained:
~ % sudo jfs_debug /dev/loop0
jfs_debugfs version 1.1.15, 04-Mar-2011
Aggregate Block Size: 4096
**Output of `su p`:**
[1] s_magic: 'JFS1' [15] s_ait2.addr1: 0x00
[2] s_version: 1 [16] s_ait2.addr2: 0x0000e92f
[3] s_size: 0x000000015d4d4ec0 s_ait2.address: 59695
[4] s_bsize: 4096 [17] s_logdev: 0x00000900
[5] s_l2bsize: 12 [18] s_logserial: 0x0009afb1
[6] s_l2bfactor: 3 [19] s_logpxd.len: 8192
[7] s_pbsize: 512 [20] s_logpxd.addr1: 0x00
[8] s_l2pbsize: 9 [21] s_logpxd.addr2: 0x2baa0160
[9] pad: Not Displayed s_logpxd.address: 732561760
[10] s_agsize: 0x00800000 [22] s_fsckpxd.len: 22408
[11] s_flag: 0x10200900 [23] s_fsckpxd.addr1: 0x00
JFS_LINUX [24] s_fsckpxd.addr2: 0x2ba9a9d8
JFS_COMMIT JFS_GROUPCOMMIT s_fsckpxd.address: 32539352
JFS_INLINELOG [25] s_time.tv_sec: 0x4902c28b
[26] s_time.tv_nsec: 0x00000000
[27] s_fpack: 'thor_storag'
[12] s_state: 0x00000001 FM_MOUNT
[13] s_compress: 0
[14] s_ait2.len: 4
**Output of `su s`:**
[1] s_magic: ' ' [15] s_ait2.addr1: 0x00
[2] s_version: 0 [16] s_ait2.addr2: 0x00000000
[3] s_size: 0x0000000000000000 s_ait2.address: 0
[4] s_bsize: 0 [17] s_logdev: 0x00000000
[5] s_l2bsize: 0 [18] s_logserial: 0x00000000
[6] s_l2bfactor: 0 [19] s_logpxd.len: 0
[7] s_pbsize: 0 [20] s_logpxd.addr1: 0x00
[8] s_l2pbsize: 0 [21] s_logpxd.addr2: 0x00000000
[9] pad: Not Displayed s_logpxd.address: 0
[10] s_agsize: 0x00000000 [22] s_fsckpxd.len: 0
[11] s_flag: 0x00000000 [23] s_fsckpxd.addr1: 0x00
[24] s_fsckpxd.addr2: 0x00000000
s_fsckpxd.address: 0
[25] s_time.tv_sec: 0x00000000
[26] s_time.tv_nsec: 0x00000000
[27] s_fpack: ''
[12] s_state: 0x00000000
FM_CLEAN
[13] s_compress: 0
[14] s_ait2.len: 0
**Output of `s2p p`:**
[1] s_magic: 'JFS1' [16] s_aim2.len: 2
[2] s_version: 1 [17] s_aim2.addr1: 0x00
[3] s_size: 0x000000015d4d4ec0 [18] s_aim2.addr2: 0x0000e92d
[4] s_bsize: 4096 s_aim2.address: 59693
[5] s_l2bsize: 12 [19] s_logdev: 0x00000900
[6] s_l2bfactor: 3 [20] s_logserial: 0x0009afb1
[7] s_pbsize: 512 [21] s_logpxd.len: 8192
[8] s_l2pbsize: 9 [22] s_logpxd.addr1: 0x00
[9] s_agsize: 0x00800000 [23] s_logpxd.addr2: 0x2baa0160
[10] s_flag: 0x10200900 s_logpxd.address: 732561760
LINUX [24] s_fsckpxd.len: 22408
GROUPCOMMIT [25] s_fsckpxd.addr1: 0x00
INLINELOG [26] s_fsckpxd.addr2: 0x2ba9a9d8
s_fsckpxd.address: 732539352
[11] s_state: 0x00000001 [27] s_fsckloglen: 50
MOUNT [28] s_fscklog: 2
[12] s_compress: 0 [29] s_fpack: 'thor_storagة�+'
[13] s_ait2.len: 4
[14] s_ait2.addr1: 0x00
[15] s_ait2.addr2: 0x0000e92f
s_ait2.address: 59695
**Output of `s2p s`:**
[1] s_magic: ' ' [16] s_aim2.len: 0
[2] s_version: 0 [17] s_aim2.addr1: 0x00
[3] s_size: 0x0000000000000000 [18] s_aim2.addr2: 0x00000000
[4] s_bsize: 0 s_aim2.address: 0
[5] s_l2bsize: 0 [19] s_logdev: 0x00000000
[6] s_l2bfactor: 0 [20] s_logserial: 0x00000000
[7] s_pbsize: 0 [21] s_logpxd.len: 0
[8] s_l2pbsize: 0 [22] s_logpxd.addr1: 0x00
[9] s_agsize: 0x00000000 [23] s_logpxd.addr2: 0x00000000
[10] s_flag: 0x00000000 s_logpxd.address: 0
[24] s_fsckpxd.len: 0
[25] s_fsckpxd.addr1: 0x00
[26] s_fsckpxd.addr2: 0x00000000
s_fsckpxd.address: 0
[11] s_state: 0x00000000 [27] s_fsckloglen: 0
CLEAN [28] s_fscklog: 0
[12] s_compress: 0 [29] s_fpack: ' '
[13] s_ait2.len: 0
[14] s_ait2.addr1: 0x00
[15] s_ait2.addr2: 0x00000000
s_ait2.address: 0
Now the question is, in what do I have to manipulate in the
superblock(s) in which values, so that either *jfs_fsck* can be
convinced to go to work, or that I can mount it (using the Linux jfs
kernel filesystem implementation)?
It'd be also perfectly acceptable to have some tool that chews through
the filesystem image and spits out everything it considers to be a file
therein, think `lost+found`.
----
###End Notes
Due to events somewhat under my control a Linux mdadm RAID-5 with a JFS
filesystem on it got corrupted in a way, which failure mode I still not
fully understand to this day. This FS was part of a NAS I shared with
another dude and I repeatedly reminded him, that RAIDs don't replace
backups, yet despite these warnings other dude has some non-backuped
data on that RAID (*and is now in a stage of being semi-angry… at me I
have the impression*). The immediate action was to remove the disks from
the system and create block level copies of them. And any kind of
manipulation I'm going to do happens on *snapshots* of these copies, so
whatever fuckup happens in the recovery process, I can rollback at any time.
Recently I was finally able to put the mdadm RAID into a (hopefully
consistent) state, that allows for further steps.
Basically I want to recover what's there, call it a day and write off
the irrecoverable losses.
Regards,
Wolfgang
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Jfs-discussion mailing list
Jfs-discussion@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/jfs-discussion
