All of lore.kernel.org
 help / color / mirror / Atom feed
From: Laurynas Biveinis <Laurynas.Biveinis@mif.vu.lt>
To: Vitaly Fertman <vitaly@namesys.com>
Cc: reiserfs-list@namesys.com
Subject: Re: Partial data loss with reiserfs3
Date: Wed, 16 Feb 2005 18:10:05 +0200	[thread overview]
Message-ID: <1108570205.4213705d41816@webmail.vu.lt> (raw)
In-Reply-To: <200502161622.06930.vitaly@namesys.com>

Hello,

Cituojant Vitaly Fertman <vitaly@namesys.com>:

> (1) copy NTFS -> ext3
> (2) shrink NTFS
> (3) reboot

> > At this point the disk has, in order of physical layout: /dev/hda5
> > (NTFS), 4GB /dev/hda6 (no FS), 2GB /dev/hda7 (reiser).
> 
> was hda7 reiserfs or ext3 ??

Following history as you have written it, it was ext3. There were
following extra steps, that I have skipped, because at the end of them
I didn't have any problems with my data.
(3.1) create /dev/hda6 as ext3
(3.2) copy data from /dev/hda7 to /dev/hda6
(3.3) delete /dev/hda7
(3.4) find out that ext3 move/resize is unsupported
(3.5) create /dev/hda7 reiser
(3.6) copy data from /dev/hda6 to /dev/hda7

At this point I had hda7 reiser.

> (4) create reiserfs on former NTFS, between NTFS and ext3
> (5) copy ext3 -> reiserfs

> > It is possible that I had created, copied, destroyed partitions
> > since last boot - but no repartitioning.
> 
> this is not clear: created, destroyed partitions means
> repartitioning,
> changing parittion table means repartitioning. Do I understand
> correctly
> that you mean create, copy, destroy filesystems here?

Right, I meant filesystems, not partitions. Sorry.

> (6) fdisk: remove hda7, remove hda6, create large hda6;
> (7) reboot

> > [root@laurynas-pc ~]# resizefs.reiserfs -f /dev/hda6 +3G
> 
> 3G just looks too large, there was only 2G hda7 as you have
> mentioned.
> right?

Right. I have specified a value knowingly too big together with '-f'
option so that fs is resized to the maximum possible size (which is
about, but not exactly, 2G). 

> (8) rezise.reiserfs
> (9) reboot
> (10) login failed
> (11) lots of fs corruptions.
>  -- some lost file StatDatas 
>  -- some wrong delimiting keys
>  -- some unformatted blocks are used more then once
>  -- 1 formatted block is used more then once 
>  -- many broken bitmap blocks
> 
> when you run resizer, has it ever asked for confirmation?
> like 'Do you want to continue? [y/N]:'

I do not remember well, but I think there wasn't any serious warning.
If
there was a warning like "You specified +3G, I can resize only to +2G,
continue?" or similar, then I had accepted it. I would have stopped on
any more serious warning or warning that I do not understand well.

> are you sure you have the relable memory? would you test it?

The system generally is stable and GCC, being the notorious RAM
checker,
compilations do not SIGSEGV. I think I will arrange memcheck86 run to
be sure.

> the reason of the problem is not clear. resizer on expand writes only
> 
> to _new_ bitmap blocks and overwrites only super block, and it does 
> not touch the fs storage tree.
> 
> so you can try to recover from the whole partition on the copy. Have
> 
> you made fs/partition backup before rebuild-tree (probably fdisk)? 
> If no, make a copy of the hda6 partition and run
> 	reiserfsck --rebuild-tree -S <hda6-copy> -l logfile
> if yes, copy it back to hda6 and run on hda6.

I don't partition backup before rebuild-tree. I'll make one now, but it
will take some time, as free space is not readily available. I'll reply
later with results.

--
Thanks,
Laurynas


  reply	other threads:[~2005-02-16 16:10 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-02-16 13:22 Partial data loss with reiserfs3 Vitaly Fertman
2005-02-16 16:10 ` Laurynas Biveinis [this message]
2005-02-17  9:22 ` Laurynas Biveinis
  -- strict thread matches above, loose matches on Subject: below --
2005-02-13 17:29 Laurynas Biveinis
2005-02-14 12:44 ` Vitaly Fertman
2005-02-14 14:40   ` Laurynas Biveinis
2005-02-14 16:21     ` Vitaly Fertman
2005-02-14 16:38       ` Laurynas Biveinis
2005-02-14 17:08         ` Vitaly Fertman
2005-02-15  9:33           ` Laurynas Biveinis
2005-02-15 12:48         ` Esben Stien
2005-02-14 15:16   ` Laurynas Biveinis
2005-02-15 12:50   ` Esben Stien
2005-02-15 14:36     ` _evil
2005-02-15 23:03       ` Simon "Sturmflut" Raffeiner

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1108570205.4213705d41816@webmail.vu.lt \
    --to=laurynas.biveinis@mif.vu.lt \
    --cc=reiserfs-list@namesys.com \
    --cc=vitaly@namesys.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.