All of lore.kernel.org
 help / color / mirror / Atom feed
From: jmerkey <jmerkey@soleranetworks.com>
To: jmerkey <jmerkey@soleranetworks.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>,
	linux-kernel <linux-kernel@vger.kernel.org>
Subject: Re: Swap Bug Massive EXT3 Corruption on FC4 with 2.6.14 update
Date: Fri, 18 Nov 2005 13:13:07 -0700	[thread overview]
Message-ID: <437E35D3.1050105@soleranetworks.com> (raw)
In-Reply-To: <437E33E5.2060603@soleranetworks.com>

jmerkey wrote:

> Alan Cox wrote:
>
>> On Iau, 2005-11-17 at 11:05 -0700, jmerkey wrote:
>>  
>>
>>> To reproduce, install FC2 on an /dev/hda device with defaults, then 
>>> install FC4 on a /dev/hdb device, build the 2.6.14 update for
>>> FC4 and watch your data disappear.
>>>   
>>
>>
>> Should be reported in the FC bugzilla although I've not been able to
>> reproduce it.
>>
>>
>>  
>>
>
> Alan,
>
> I'll report over there.  I reproduced it with an install of Suse 10.0 
> and FC4 and got to the bottom of it.  During install of FC4, anaconda 
> allocates
> the swap partitions assigned to Suse 10.0 on /dev/hda (or any swap 
> partitions on the primary drive) for use during the install.  After 
> the install
> completes, FC4 uses this LABEL-SWAP-hda2 (etc.) method for determining 
> which partitions to use for swap.  What happened here it turned
> out was not related to swap extents, but misidentifcation of which 
> partition was assigned this LABEL-XXX tag.  Upon first boot of FC4,
> it allocated /dev/hda6 (the / partitition) as swap and started 
> swapping to the / partition for Suse 10.0.  I first saw it when I 
> installed FC4 on a system
> with FC2.  After FC2 / partition got trashed, I reinstalled with Suse 
> 10.0 (since I am porting DSFS to all of these distributions) and then 
> reinstalled
> FC4 on /dev/hdb -- same thing happened again.
> I just finished reinstalling Suse 10.0 and tried with FC2 on 
> /dev/hdb.  FC2 does the same thing and gets mixed on on Swap on the 
> /dev/hda device, but this time, it did not corrupt the Suse 10.0 on 
> /dev/hda.  This appears to be a bug in anaconda and the setup for the 
> FCX distributions.  ES and AS probably do the same thing since they 
> use anaconda, so I would have someone look into this.
>
> Jeff
>
NOTE;  One this that's unique in this case is that after I instal DSFS 
rpms, I remove these LABEL-XXX constructs from the grub.conf and 
/etc/fstab files and use the actual device names /dev/hdX.  It seems 
related to removing these labels in a running distribution.

Jeff

      reply	other threads:[~2005-11-18 20:39 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-11-17 18:05 Swap Bug Massive EXT3 Corruption on FC4 with 2.6.14 update jmerkey
2005-11-18 20:35 ` Alan Cox
2005-11-18 20:04   ` jmerkey
2005-11-18 20:13     ` jmerkey [this message]

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=437E35D3.1050105@soleranetworks.com \
    --to=jmerkey@soleranetworks.com \
    --cc=alan@lxorguk.ukuu.org.uk \
    --cc=linux-kernel@vger.kernel.org \
    /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.