reiserfsck question - session results

reiserfsck question - session results

[David R. Bergstein]

[-- Attachment #1: Type: text/plain, Size: 9096 bytes --]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello,

I recently encountered some issues on my boot partition on a system running 
the linux 2.4.19-rc3 kernel.  There were some corruptions that were reported 
by reiserfsck version 3.6.2 that could be repaired using the fix-fixable 
option.  To ensure that this could be done with the partition completely 
unmounted I built a rescure CD with the appropriate kernel, modules, and 
statically linked copies of the reiserfsprogs 3.6.2 utilities.

I had to back the partition up to an image file, so I was able to replay what 
happened next.  Apparently, when I boot up the system in normal mode. I get a 
message indicating the boot partition was NOT cleanly unmounted, which is in 
conflict with checking the partition offline (says it was cleanly unmounted), 
i.e.: here is a copy of the on-line message I see in /var/log/boot.log:

Jul 28 22:49:56 rebel fsck: Reiserfs super block in block 16 on 0x815 of 
format 3.6 with standard journal
Jul 28 22:49:56 rebel fsck: Blocks (total/free): 126496/75295 by 4096 bytes
Jul 28 22:49:56 rebel fsck: Filesystem is NOT cleanly umounted

Below is a replay of what I did offline using a test copy of the partition 
dump (also attached are copies of the check.log and rebuild.log files in case 
they are any help):

[root@rebel tmp]# cp -v sdb5.img sdb5-test.img
`sdb5.img' -> `sdb5-test.img'
[root@rebel tmp]# v
total 1011969
drwxr-xr-x    2 root     root          104 Jul 28 22:53 ./
drwxr-xr-x   17 root     root          432 Jul 28 17:10 ../
- -rw-r--r--    1 root     root     518127616 Jul 28 17:11 sdb5.img
- -rw-r--r--    1 root     root     518127616 Jul 28 22:53 sdb5-test.img
[root@rebel tmp]# reiserfsck sdb5-test.img --logfile check.log

<-------------reiserfsck, 2002------------->
reiserfsprogs 3.6.2

Will read-only check consistency of the filesystem on sdb5-test.img
Will put log info to 'check.log'

Do you want to run this program?[N/Yes] (note need to type Yes):Yes
###########
reiserfsck --check started at Sun Jul 28 22:54:21 2002
###########
Replaying journal..
No transactions found
Checking S+tree..ok
Comparing bitmaps..ok
Checking Semantic tree...
ok
There were found 2 corruptions which can be fixed with --fix-fixable
###########
reiserfsck finished at Sun Jul 28 22:54:26 2002
###########
[root@rebel tmp]# ls
check.log  sdb5.img  sdb5-test.img
[root@rebel tmp]# cat check.log
bad_stat_data: block 8278, [68 86 0x0 SD (0)], SD has bad nlink number
bad_stat_data: block 8305, [68 95 0x0 SD (0)], SD has bad nlink number
[root@rebel tmp]# reiserfsck sdb5-test.img --fix-fixable --logfile fix.log

<-------------reiserfsck, 2002------------->
reiserfsprogs 3.6.2

Will check consistency of the filesystem on sdb5-test.img
Will fix what can be fixed w/o --rebuild-tree
Will put log info to 'fix.log'

Do you want to run this program?[N/Yes] (note need to type Yes):Yes
###########
reiserfsck --fix-fixable started at Sun Jul 28 22:55:06 2002
###########
Replaying journal..
No transactions found
Checking S+tree..ok
Comparing bitmaps..ok
Checking Semantic tree...
ok
No corruptions found
There are on the filesystem:
        Leaves 1062
        Internal nodes 8
        Directories 1101
        Other files 21788
        Data block pointers 42347 (5 of them are zero)
        Safe links 0
###########
reiserfsck finished at Sun Jul 28 22:55:09 2002
###########
[root@rebel tmp]# ls
check.log  fix.log  sdb5.img  sdb5-test.img
[root@rebel tmp]# v
total 1011973
drwxr-xr-x    2 root     root          160 Jul 28 22:55 ./
drwxr-xr-x   17 root     root          432 Jul 28 17:10 ../
- -rw-r--r--    1 root     root          142 Jul 28 22:54 check.log
- -rw-r--r--    1 root     root            0 Jul 28 22:55 fix.log
- -rw-r--r--    1 root     root     518127616 Jul 28 17:11 sdb5.img
- -rw-r--r--    1 root     root     518127616 Jul 28 22:55 sdb5-test.img
[root@rebel tmp]# view check.log
[root@rebel tmp]# reiserfsck sdb5-test.img --rebuild-tree --logfile 
rebuild.log
<-------------reiserfsck, 2002------------->
reiserfsprogs 3.6.2

  **********************************************************
  ** This  is  an  experimental  version  of  reiserfsck, **
  **              !! MAKE A BACKUP FIRST !!               **
  ** Don't run this program unless something  is  broken. **
  ** Some types of random FS damage can be recovered from **
  ** by  this  program,   which  basically   throws  away **
  ** the internal nodes of the tree and then reconstructs **
  ** them. This program is for use only by the desperate, **
  ** and is  of only beta quality.  If you are using  the **
  ** latest  reiserfsprogs  and  it  fails  please  email **
  ** bug reports to reiserfs-list@namesys.com.            **
  **********************************************************

Will rebuild the filesystem (sdb5-test.img) tree
Will put log info to 'rebuild.log'

Do you want to run this program?[N/Yes] (note need to type Yes):Yes
Replaying journal..
No transactions found
###########
reiserfsck --rebuild-tree started at Sun Jul 28 22:56:02 2002
###########

Pass 0:
Loading on-disk bitmap .. ok, 51626 blocks marked used
Skipping 8214 blocks (super block, journal, bitmaps) 43412 blocks will be read
0%....20%....40%....60%....80%....100%                        left 0, 6201 
/sec         "r5" hash is selected
Flushing..done
        Read blocks (but not data blocks) 43412
                Leaves among those 1062
                Objectids found 22885

Pass 1 (will try to insert 1062 leaves):
Looking for allocable blocks .. ok
0%....20%....40%....60%....80%....100%                        left 0, 1062 
/sec Flushing..done
        1062 leaves read
                1059 inserted
                3 not inserted

Pass2:
0%....20%....40%....60%....80%....100%                           left 0, 0 
/sec Flushing..done
        Leaves inserted item by item 3
Pass 3 (semantic):
Flushing..done
        Files found: 6342
        Directories found: 1102
        Symlinks found: 659
        Others: 14780
Pass 3a (looking for lost dir/files):
Looking for lost directories:
Looking for lost files:0 /sec
Flushing..done done 0, 0 /sec
        Objects without names 2
        Empty lost dirs removed 2
        Files linked to /lost+found 2
Pass 4 - done           done 0, 0 /sec
Flushing..done
Syncing..done
###########
reiserfsck finished at Sun Jul 28 22:56:12 2002
###########
[root@rebel tmp]# ls
check.log  fix.log  rebuild.log  sdb5.img  sdb5-test.img
[root@rebel tmp]# v
total 1011977
drwxr-xr-x    2 root     root          192 Jul 28 22:56 ./
drwxr-xr-x   17 root     root          432 Jul 28 17:10 ../
- -rw-r--r--    1 root     root          142 Jul 28 22:54 check.log
- -rw-r--r--    1 root     root            0 Jul 28 22:55 fix.log
- -rw-r--r--    1 root     root          458 Jul 28 22:56 rebuild.log
- -rw-r--r--    1 root     root     518127616 Jul 28 17:11 sdb5.img
- -rw-r--r--    1 root     root     518127616 Jul 28 22:56 sdb5-test.img
[root@rebel tmp]# view rebuild.log
[root@rebel tmp]# reiserfsck sdb5-test.img --logfile check2.log

<-------------reiserfsck, 2002------------->
reiserfsprogs 3.6.2

Will read-only check consistency of the filesystem on sdb5-test.img
Will put log info to 'check2.log'

Do you want to run this program?[N/Yes] (note need to type Yes):Yes
###########
reiserfsck --check started at Sun Jul 28 22:57:19 2002
###########
Replaying journal..
No transactions found
Checking S+tree..ok
Comparing bitmaps..ok
Checking Semantic tree...
ok
No corruptions found
There are on the filesystem:
        Leaves 1061
        Internal nodes 9
        Directories 1102
        Other files 21790
        Data block pointers 42347 (5 of them are zero)
        Safe links 0
###########
reiserfsck finished at Sun Jul 28 22:57:22 2002
###########
[root@rebel tmp]# ls
check2.log  check.log  fix.log  rebuild.log  sdb5.img  sdb5-test.img
[root@rebel tmp]# v
total 1011977
drwxr-xr-x    2 root     root          224 Jul 28 22:57 ./
drwxr-xr-x   17 root     root          432 Jul 28 17:10 ../
- -rw-r--r--    1 root     root            0 Jul 28 22:57 check2.log
- -rw-r--r--    1 root     root          142 Jul 28 22:54 check.log
- -rw-r--r--    1 root     root            0 Jul 28 22:55 fix.log
- -rw-r--r--    1 root     root          458 Jul 28 22:56 rebuild.log
- -rw-r--r--    1 root     root     518127616 Jul 28 17:11 sdb5.img
- -rw-r--r--    1 root     root     518127616 Jul 28 22:57 sdb5-test.img
 
- -- 

David R. Bergstein
Systems Engineer and Blues Musician - 
http://mywebpages.comcast.net/dbergstein/
Heart of Blue - bookings on-line at http://www.heartofblue.com
OpenPGP Public Key 0xE1F138CA - For info see http://www.gnupg.org
Key fingerprint = C86E CA2A 4171 AC73 91D7  3DCE 8832 D764 E1F1 38CA
________________________________________________________________________

I am an optimist.  It does not seem too much use being anything else.
		-- Winston Churchill
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)

iD8DBQE9RLLQiDLXZOHxOMoRAsBXAJ43cCmYnpm0Q8JT6cJEigtppdr38QCgoV4g
XUjJKoXWfp5jFToKixYJNQo=
=ecds
-----END PGP SIGNATURE-----

[-- Attachment #2: rebuild.log --]
[-- Type: text/plain, Size: 458 bytes --]

####### Pass 0 #######
"r5" got 22888 hits
####### Pass 1 #######
####### Pass 2 #######
####### Pass 3 #########
are_file_items_correct: convert the indirect item back to direct 486 17077
are_file_items_correct: convert the indirect item back to direct 486 17078
are_file_items_correct: convert the indirect item back to direct 486 17076
are_file_items_correct: convert the indirect item back to direct 619 17027
####### Pass 3a (lost+found pass) #########

[-- Attachment #3: check.log --]
[-- Type: text/plain, Size: 142 bytes --]

bad_stat_data: block 8278, [68 86 0x0 SD (0)], SD has bad nlink number
bad_stat_data: block 8305, [68 95 0x0 SD (0)], SD has bad nlink number

Re: reiserfsck question - session results

[David R. Bergstein]

It turned out there was no problem after I ran reiserfsck, it was an 
initscript issue, specifically in /etc/rc.d/rc.sysinit on Mandrake 8.1.  The 
script assumes that the root filesystem is first mounted read-only, however I 
had previously taken steps to ensure that this reiserfs partition is mounted 
read-write so journal replays work correctly.  This resulted in the boot time 
message about the root filesystem not being cleanly unmounted.

For the benefit of others in similar situations, I changed the following 
lines to check for reiserfs in /proc/mounts:

[dbergst@rebel rc.d]$ diff -C 4 rc.sysinit rc.sysinit.original
*** rc.sysinit  Mon Jul 29 00:28:13 2002
--- rc.sysinit.original Tue Oct 16 14:25:51 2001
***************
*** 378,387 ****
  }

  _RUN_QUOTACHECK=0
  ROOTFSTYPE=`grep " / " /proc/mounts | awk '{ print $3 }'`
! if [ -z "$fastboot" -a "$ROOTFSTYPE" != "nfs" ] && \
!    [ -z "$fastboot" -a "$ROOTFSTYPE" == "reiserfs" ]; then

          gprintf "Checking root filesystem\n"
        Fsck -T -a $fsckoptions /

--- 378,386 ----
  }

  _RUN_QUOTACHECK=0
  ROOTFSTYPE=`grep " / " /proc/mounts | awk '{ print $3 }'`
! if [ -z "$fastboot" -a "$ROOTFSTYPE" != "nfs" ]; then

          gprintf "Checking root filesystem\n"
        Fsck -T -a $fsckoptions /

On Sunday 28 July 2002 11:13 pm, David R. Bergstein wrote:
> Hello,
>
> I recently encountered some issues on my boot partition on a system running
> the linux 2.4.19-rc3 kernel.  There were some corruptions that were
> reported by reiserfsck version 3.6.2 that could be repaired using the
> fix-fixable option.  To ensure that this could be done with the partition
> completely unmounted I built a rescure CD with the appropriate kernel,
> modules, and statically linked copies of the reiserfsprogs 3.6.2 utilities.
>
> I had to back the partition up to an image file, so I was able to replay
> what happened next.  Apparently, when I boot up the system in normal mode.
> I get a message indicating the boot partition was NOT cleanly unmounted,
> which is in conflict with checking the partition offline (says it was
> cleanly unmounted), i.e.: here is a copy of the on-line message I see in
> /var/log/boot.log:
>
> Jul 28 22:49:56 rebel fsck: Reiserfs super block in block 16 on 0x815 of
> format 3.6 with standard journal
> Jul 28 22:49:56 rebel fsck: Blocks (total/free): 126496/75295 by 4096 bytes
> Jul 28 22:49:56 rebel fsck: Filesystem is NOT cleanly umounted

-- 

David R. Bergstein
Systems Engineer and Blues Musician - 
http://mywebpages.comcast.net/dbergstein/
Heart of Blue - bookings on-line at http://www.heartofblue.com
OpenPGP Public Key 0xE1F138CA - For info see http://www.gnupg.org
Key fingerprint = C86E CA2A 4171 AC73 91D7  3DCE 8832 D764 E1F1 38CA
________________________________________________________________________

There's another way to survive.  Mutual trust -- and help.
		-- Kirk, "Day of the Dove", stardate unknown

Re: reiserfsck question - session results

[Oleg Drokin]

Hello!

On Mon, Jul 29, 2002 at 12:43:08AM -0400, David R. Bergstein wrote:

> It turned out there was no problem after I ran reiserfsck, it was an 
> initscript issue, specifically in /etc/rc.d/rc.sysinit on Mandrake 8.1.  The 
> script assumes that the root filesystem is first mounted read-only, however I 
> had previously taken steps to ensure that this reiserfs partition is mounted 
> read-write so journal replays work correctly.  This resulted in the boot time 

In fact journal replays work ok even with readonly mounts. So mounting rootfs
readonly is not required in reiserfs case.

Bye,
    Oleg

RE: reiserfsck question - session results

[David R. Bergstein]

Oleg,

Thanks very much for this information.  I sounds like what I did overall
suppressed several harnless messages, i.e., by mounting the root reiserfs
read-write, and skipping fsck in the initscripts.  From what you are saying,
journal replay can also work safely when even a reiserfs fs is mounted
read-only.  Is the warning message about journal replay on read-only
filesystems outdated and if so, will it be removed sometime in the near
future?

Regards,

- David

David R. Bergstein
Systems Engineer and Blues Musician -
http://mywebpages.comcast.net/dbergstein/
Heart of Blue - bookings on-line at http://www.heartofblue.com
OpenPGP Public Key 0xE1F138CA - For info see http://www.gnupg.org
Key fingerprint = C86E CA2A 4171 AC73 91D7  3DCE 8832 D764 E1F1 38CA

> -----Original Message-----
> From: Oleg Drokin [mailto:green@namesys.com]
> Sent: Monday, July 29, 2002 2:01 AM
> To: David R. Bergstein
> Cc: reiserfs-list
> Subject: Re: [reiserfs-list] reiserfsck question - session results
>
>
> Hello!
>
> On Mon, Jul 29, 2002 at 12:43:08AM -0400, David R. Bergstein wrote:
>
> > It turned out there was no problem after I ran reiserfsck, it was an
> > initscript issue, specifically in /etc/rc.d/rc.sysinit on
> Mandrake 8.1.  The
> > script assumes that the root filesystem is first mounted
> read-only, however I
> > had previously taken steps to ensure that this reiserfs
> partition is mounted
> > read-write so journal replays work correctly.  This resulted in
> the boot time
>
> In fact journal replays work ok even with readonly mounts. So
> mounting rootfs
> readonly is not required in reiserfs case.
>
> Bye,
>     Oleg
>

Re: reiserfsck question - session results

[Oleg Drokin]

Hello!

On Mon, Jul 29, 2002 at 07:20:45AM -0400, David R. Bergstein wrote:

> Thanks very much for this information.  I sounds like what I did overall
> suppressed several harnless messages, i.e., by mounting the root reiserfs
> read-write, and skipping fsck in the initscripts.  From what you are saying,
> journal replay can also work safely when even a reiserfs fs is mounted
> read-only.  Is the warning message about journal replay on read-only
> filesystems outdated and if so, will it be removed sometime in the near
> future?

Yes, the warning will be removed. Actually it is just informational message,
just badly phrased.

Bye,
    Oleg