Especially broken btrfs

Especially broken btrfs

[sepero111]

Hello all. I submit bugs to different foss projects regularly, but I don't 
really have a bug report this time. I have a broken filesystem to report. And I 
have no idea how to reproduce it.

I am including a link to the filesystem itself, because it appears to be 
unrepairable and unrestorable. I have no personal information on the disk image. 
The filesystem is almost 512MB uncompressed. I was using it on an old usb drive 
with 512MB size limitation. I only used (abused?) it about 2 days before this 
corruption.

My goal was to use the usb as a bootable rescue system. I decided to try Btrfs 
instead of Ext4, because it supports filesystem compression.

BTRFS IMAGE LINK (please pardon my file hosting service)
http://www.mediafire.com/download/gdaydt3mz8uwtmm/sdb1.btrfs.xz


These are some things that may have helped to cause the corruption.

+Created btrfs with -M flag
+Installed Debian testing/unstable
+When mounting, I always used at least these options: 
ssd_spread,noatime,compression=zlib,autodefrag
+Occasionally force powering off computer.
+While booted into usb system, I was constantly running out of space while 
trying to install new packages.

It is my hope that this image might be used to improve the btrfs restore and 
btrfsck tools. Please let me know if I can provide any further information. Big 
thanks to everyone helping to further development of Btrfs.

Sepero

Re: Especially broken btrfs

[Marc MERLIN]

On Thu, Mar 20, 2014 at 11:21:27PM -0400, sepero111@gmx.com wrote:
> Hello all. I submit bugs to different foss projects regularly, but I
> don't really have a bug report this time. I have a broken filesystem
> to report. And I have no idea how to reproduce it.
> 
> I am including a link to the filesystem itself, because it appears
> to be unrepairable and unrestorable. I have no personal information
> on the disk image. The filesystem is almost 512MB uncompressed. I
> was using it on an old usb drive with 512MB size limitation. I only
> used (abused?) it about 2 days before this corruption.
> 
> My goal was to use the usb as a bootable rescue system. I decided to
> try Btrfs instead of Ext4, because it supports filesystem
> compression.
> 
> BTRFS IMAGE LINK (please pardon my file hosting service)
> http://www.mediafire.com/download/gdaydt3mz8uwtmm/sdb1.btrfs.xz
 
It looks like you got no answer, and I'm not a dev so I can't help you
either.

The btrfs devs are pretty busy and can't always get back to everyone.
Hopefully they'll be able to look at this, but sorry if not.

Marc
 
> These are some things that may have helped to cause the corruption.
> 
> +Created btrfs with -M flag
> +Installed Debian testing/unstable
> +When mounting, I always used at least these options:
> ssd_spread,noatime,compression=zlib,autodefrag
> +Occasionally force powering off computer.
> +While booted into usb system, I was constantly running out of space
> while trying to install new packages.
> 
> It is my hope that this image might be used to improve the btrfs
> restore and btrfsck tools. Please let me know if I can provide any
> further information. Big thanks to everyone helping to further
> development of Btrfs.
> 
> Sepero
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

-- 
"A mouse is a device used to point at the xterm you want to type in" - A.S.R.
Microsoft is to operating systems ....
                                      .... what McDonalds is to gourmet cooking
Home page: http://marc.merlins.org/                         | PGP 1024R/763BE901

Re: Especially broken btrfs

[sepero111]

Hi, I probably should have used a better subject title. Also, I submitted this 
without knowing if it would be helpful or not. If it can be used in a good way 
Great! If not, then no problem. I appreciate you getting back with me, Marc. 
Thanks. :)


On 03/30/2014 12:50 AM, Marc MERLIN wrote:
> On Thu, Mar 20, 2014 at 11:21:27PM -0400, sepero111@gmx.com wrote:
>> Hello all. I submit bugs to different foss projects regularly, but I
>> don't really have a bug report this time. I have a broken filesystem
>> to report. And I have no idea how to reproduce it.
>>
>> I am including a link to the filesystem itself, because it appears
>> to be unrepairable and unrestorable. I have no personal information
>> on the disk image. The filesystem is almost 512MB uncompressed. I
>> was using it on an old usb drive with 512MB size limitation. I only
>> used (abused?) it about 2 days before this corruption.
>>
>> My goal was to use the usb as a bootable rescue system. I decided to
>> try Btrfs instead of Ext4, because it supports filesystem
>> compression.
>>
>> BTRFS IMAGE LINK (please pardon my file hosting service)
>> http://www.mediafire.com/download/gdaydt3mz8uwtmm/sdb1.btrfs.xz
>   
> It looks like you got no answer, and I'm not a dev so I can't help you
> either.
>
> The btrfs devs are pretty busy and can't always get back to everyone.
> Hopefully they'll be able to look at this, but sorry if not.
>
> Marc
>   
>> These are some things that may have helped to cause the corruption.
>>
>> +Created btrfs with -M flag
>> +Installed Debian testing/unstable
>> +When mounting, I always used at least these options:
>> ssd_spread,noatime,compression=zlib,autodefrag
>> +Occasionally force powering off computer.
>> +While booted into usb system, I was constantly running out of space
>> while trying to install new packages.
>>
>> It is my hope that this image might be used to improve the btrfs
>> restore and btrfsck tools. Please let me know if I can provide any
>> further information. Big thanks to everyone helping to further
>> development of Btrfs.
>>
>> Sepero
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>

Re: Especially broken btrfs

[Bob Marley]

Hi, I hadn't noticed this post,
I think I know the reason this time : you have used USB you bad guy!
I think USB does not support flush / barrier , which is mandatory for 
BTRFS to work correctly in case of power loss.
For most filesystems actually, but the damages suffered by COW 
filesystems such as btrfs are much more severe than for static 
filesystems such as ext4 .

Please check if when you connect the USB drive you see in dmesg 
something like:

|[  .... . .....] sd ...:0:0:0: [sdf] Write cache: ....., read cache: ....., doesn't support DPO or FUA
|

Regards
BM


On 21/03/2014 04:21, sepero111@gmx.com wrote:
> Hello all. I submit bugs to different foss projects regularly, but I 
> don't really have a bug report this time. I have a broken filesystem 
> to report. And I have no idea how to reproduce it.
>
> I am including a link to the filesystem itself, because it appears to 
> be unrepairable and unrestorable. I have no personal information on 
> the disk image. The filesystem is almost 512MB uncompressed. I was 
> using it on an old usb drive with 512MB size limitation. I only used 
> (abused?) it about 2 days before this corruption.
>
> My goal was to use the usb as a bootable rescue system. I decided to 
> try Btrfs instead of Ext4, because it supports filesystem compression.
>
> BTRFS IMAGE LINK (please pardon my file hosting service)
> http://www.mediafire.com/download/gdaydt3mz8uwtmm/sdb1.btrfs.xz
>
>
> These are some things that may have helped to cause the corruption.
>
> +Created btrfs with -M flag
> +Installed Debian testing/unstable
> +When mounting, I always used at least these options: 
> ssd_spread,noatime,compression=zlib,autodefrag
> +Occasionally force powering off computer.
> +While booted into usb system, I was constantly running out of space 
> while trying to install new packages.
>
> It is my hope that this image might be used to improve the btrfs 
> restore and btrfsck tools. Please let me know if I can provide any 
> further information. Big thanks to everyone helping to further 
> development of Btrfs.
>
> Sepero
>
> -- 
> To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

Re: Especially broken btrfs

[Duncan]

Bob Marley posted on Mon, 31 Mar 2014 19:04:38 +0200 as excerpted:

> Hi, I hadn't noticed this post,
> I think I know the reason this time : you have used USB you bad guy!
> I think USB does not support flush / barrier , which is mandatory for
> BTRFS to work correctly in case of power loss.
> For most filesystems actually, but the damages suffered by COW
> filesystems such as btrfs are much more severe than for static
> filesystems such as ext4 .
> 
> Please check if when you connect the USB drive you see in dmesg
> something like:
> 
> |[  .... . .....] sd ...:0:0:0: [sdf] Write cache: ....., read cache:
> ....., doesn't support DPO or FUA
> |

Umm... Most direct-connect SATA drives are apparently missing DPO/FUA 
also, or at least it's not normally enabled by the driver.  Certainly 
that's what dmesg says for my SATA drives here, tho for the pair of SSDs 
(but not the spinning rust) hdparm -I says WRITE__{DMA|MULTIPLE}_FUA_EXT 
is enabled (star in the enabled column), while  WRITE_DMA_QUEUED_FUA_EXT 
is not (no star).

Google says only enterprise-grade SAS/SCSI drives are likely to have it.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman