Re: journal size==0 [reiserfs 3.6, reiserutils 3.x.1b]

[Hans Reiser <reiser@namesys.com>]

Thomas Wiedemann wrote:

>hi reiserfs-developers,
>
>i cannot mount my reiserfs partition any more, because of that:
>
>super-459: read_super_block: super found at block 16 is within its own log. It must not be of this format type.
>
>
>so, here's the story how i "created" this error and some details:
>
>due a hardware failure (the harddisk?), i got some weird error messages:
>
>Apr 13 17:32:13 hq kernel: hdb: dma_intr: status=0x51 { DriveReady SeekComplete Error }
>Apr 13 17:32:13 hq kernel: hdb: dma_intr: error=0x84 { DriveStatusError BadCRC }
>(...)
>
>followed by some filesystem errors:
>
>Apr 13 17:33:46 hq kernel: is_tree_node: node level 432 does not match to the expected one 1
>Apr 13 17:33:46 hq kernel: vs-5150: search_by_key: invalid format found in block 19232. Fsck?
>Apr 13 17:33:46 hq kernel: vs-13070: reiserfs_read_inode2: i/o failure occurred trying to find stat data of [1584 1590 0x0
> SD]
>...and so on....
>
>i had to reboot my pc, the disk was ok, but the partition was dead.
>so i downloaded the newest reiserfs-utils (3.x.1b) and tried to repair it.
>the superblock seemed to be screwed up, so...
>
># reiserfsck /dev/hdb1 --rebuild-sb --no-journal-available
># reiserfsck --rebuild-tree
>
>...did some work on the filesystem, but i'm still not able to mount it, because while rebuilding the superblock, fsck did not make a journal on the device:
>
># debugreiserfs /dev/hdb1
>
><-------------debugreiserfs, 2002------------->
>reiserfsprogs 3.x.1b
>
>
>Filesystem state: consistent
>
>Reiserfs super block in block 16 on 0x341 of format 3.6 with standard journal
>Count of blocks on the device: 11273605
>Number of bitmaps: 345
>Blocksize: 4096
>Free blocks (count of blocks - used [journal, bitmaps, data, reserved] blocks): 1611835
>Root block: 431
>Filesystem is cleanly umounted
>Tree height: 5
>Hash function used to sort names: "r5"
>Objectid map size 972, max 972
>Journal parameters:
>        Device [0x0]
>        Magic [0x0]
>        Size 1 blocks (including 1 for journal header) (first block 0)
>        Max transaction length 0 blocks
>        Max batch size 0 blocks
>        Max commit age 0
>Blocks reserved by journal: 0
>Fs state field: 0x0
>sb_version: 2
>inode generation number: 0
>UUID: c0ce572a-1ed0-4827-b72a-8aef5cdb2015
>LABEL: 
>Set flags in SB:
>
>
>now, that the journal's size is only one block, mount fails, while fsck claims everything to be alright (mount -o nolog also fails).
>i don't know if this is a bug in reiserfsck....and maybe there is a way to insert a journal into between the rootblock and the superblock (where the journal was before the crash...)?
>
>thanks, thomas.
>
>
>
bad/misconfigured hardware consequences (and recovering from them) are 
handled by www.namesys.com/support.html, or by users on the mailing 
list.   The more people who go to www.namesys.com/support.html, the more 
I can afford to have Vitaly improve fsck....

Hans