All of lore.kernel.org
 help / color / mirror / Atom feed
From: Nathan Kidd <nathan@spicycrypto.ca>
To: linux-btrfs@vger.kernel.org
Subject: Recovering from persistent  kernel oops on  'btrfs balance'
Date: Sat, 08 Feb 2014 13:29:00 -0500	[thread overview]
Message-ID: <52F6776C.4000109@spicycrypto.ca> (raw)

Hi,

I added a 2nd device and 'btrfs balance' crashed (kernel oops) half way 
through, now I can only read the fs from a rawhide livedvd, but even 
that can't fix the fs (finish balance, or remove 2nd device to try 
again).  I'd be grateful for any advice on getting back to a working 
btrfs filesystem.

Details
=======

Hardware: Asus P5G41T-M with Pentium dual core E2140,4GB ram, OS on 
ext4 drive, two 4TB Segate "NAS" SATA drives.

On Ubuntu 13.04 x86_64 (3.8 kernel, btrfs-tools 0.19+20130117)

1. Install new 4TB drive (/dev/sdb), use gparted to create full-disk 
btrfs partition, mount on /ark copy ~500GB data, everything working well 
for a couple weeks

2. Install additional identical 4TB drive,

Following 
https://btrfs.wiki.kernel.org/index.php/Using_Btrfs_with_Multiple_Devices#Adding_new_devices

3. btrfs device add /dev/sdc /ark

4. btrfs balance start -dconvert=raid1 -mconvert=raid1 /ark

5. After ~1 hour, at about 50% (according to 'btrfs balance status', the 
system locks up with this displayed (sorry, JPEG): 
http://i.imgur.com/Ds9pnZV.jpg

6. System repeat same oops on startup

7. After removing /dev/sdc system boots but can't see anything on /ark

I guess using a 3.8 kernel wasn't the smartest idea. Let's update.

8. Update to Ubuntu 13.11 x86_64 (3.11 kernel, btrfs-tools 
0.19+20130705-1) 	

9. Now system boots with /dev/sdc plugged in but still can't see data on 
/ark, IIRC the balance command gave similar kernel oops.

10。 Fine I'll try Rawhide. From Jan 30, 2014, kernel 
3.14.0-0.rc0.git17.1.fc21.x86_64

11. I can see data on /ark!

12. If I try to 'btrfs balance resume' or 'btrfs balance cancel' I get 
roughly the same kernel oops: http://pastebin.ca/2634583

13. 'btrfs device delete /dev/sdc /ark' says it cannot be done while 
balance is underway

14. Help!  Any suggestion on how to recover the btrfs fs?

My last resort idea is pull /dev/sdb (which seems to have actual data 
that rawhide can see), format /dev/sdc ext4, plug both drives in again 
and copy from btrfs /dev/sdb to ext4 /dev/sdc, then wipe the btrfs fs on 
/dev/sdb and try again with the 3.11 kernel (or just with rawhide?).
But that is a whole lot of copying it would be nice to avoid.

Thanks,

-Nathan



                 reply	other threads:[~2014-02-08 18:29 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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=52F6776C.4000109@spicycrypto.ca \
    --to=nathan@spicycrypto.ca \
    --cc=linux-btrfs@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.