[gentoo-user] Re: ext4 - grow_buffers: requested out-of-range block <BLOCKID>

[gentoo-user] Re: ext4 - grow_buffers: requested out-of-range block <BLOCKID>

[James Broadhead]

On 12 December 2011 14:14, James Broadhead <jamesbroadhead@gmail.com> wrote:
> ext4: "fill_buffer on unknown block <BLOCKID> out of range"

Apologies; the correct message is:
grow_buffers: requested out-of-range block 18446744072382021139 for device sdb1

This appears 42 times immediately following mount.

Running picasa today, it informed me that one of the files I was
working with was corrupted (but put the message in a box too small to
read the full path).

This makes me think that perhaps the disk is bad.  Any advice, aside
from the usual "get your data off asap"?

Re: [gentoo-user] Re: ext4 - grow_buffers: requested out-of-range block <BLOCKID>

[Paul Hartman]

On Mon, Dec 12, 2011 at 2:33 PM, James Broadhead
<jamesbroadhead@gmail.com> wrote:
> On 12 December 2011 14:14, James Broadhead <jamesbroadhead@gmail.com> wrote:
>> ext4: "fill_buffer on unknown block <BLOCKID> out of range"
>
> Apologies; the correct message is:
> grow_buffers: requested out-of-range block 18446744072382021139 for device sdb1
>
> This appears 42 times immediately following mount.
>
> Running picasa today, it informed me that one of the files I was
> working with was corrupted (but put the message in a box too small to
> read the full path).
>
> This makes me think that perhaps the disk is bad.  Any advice, aside
> from the usual "get your data off asap"?

Does it happen to be a >2TB USB drive? I remember reading about
problems with some of those. It works in Windows with the factory
partition/FAT tables because of tricks they do to the addressing that
works in Windows, but once you reformat it you can't access the >2TB
areas. Something like that... As far as I recall, you could
repartition to create a 2TB or smaller partition and that would work,
but then the rest of the drive was inaccessible.

Re: [gentoo-user] Re: ext4 - grow_buffers: requested out-of-range block <BLOCKID>

[James Broadhead]

On 12 December 2011 20:55, Paul Hartman <paul.hartman+gentoo@gmail.com> wrote:
> On Mon, Dec 12, 2011 at 2:33 PM, James Broadhead
> <jamesbroadhead@gmail.com> wrote:
>>
>> Apologies; the correct message is:
>> grow_buffers: requested out-of-range block 18446744072382021139 for device sdb1
>>
>> This appears 42 times immediately following mount.
>>
>> Running picasa today, it informed me that one of the files I was
>> working with was corrupted (but put the message in a box too small to
>> read the full path).
>>
>> This makes me think that perhaps the disk is bad.  Any advice, aside
>> from the usual "get your data off asap"?
>
> Does it happen to be a >2TB USB drive? I remember reading about
> problems with some of those. It works in Windows with the factory
> partition/FAT tables because of tricks they do to the addressing that
> works in Windows, but once you reformat it you can't access the >2TB
> areas. Something like that... As far as I recall, you could
> repartition to create a 2TB or smaller partition and that would work,
> but then the rest of the drive was inaccessible.

So on returning to this machine, I see that another USB disk that I
have connected to it is also having those messages printed about it.
This leads me to suspect that it's either an ext4 bug or the situation
that you mentioned above.

Both are Western Digital 2TB disks;
1058:1130 Western Digital Technologies, Inc.
1058:1021 Western Digital Technologies, Inc. Elements 2TB

There are 42 messages in quick succession for each disk, appearing to
cycle through the same list of blocks twice. I'll attach the messages.

I'm inclined towards the bad-usb-firmware idea - do you have a link to
where you read about the 2TB partition problem ?

I don't have much time to deal with this at the moment, so I think
that I'll just power them down and wait until I do.


On 12 December 2011 21:52, Florian Philipp <lists@binarywings.net> wrote:
>
> I've looked at the kernel code that causes the error message. It
> verifies that this is most likely a dead disk:
>

Now that I have 2 from the same manufacturer, of similar vintage
causing the same errors, it's probably not simultaneous failure
(unless I'm super-unlucky!). It's also entirely possible that it's an
ext4 bug, so I'll try with a different kernel;
Linux broadhej-D830 3.1.2-gentoo #2 SMP PREEMPT Sun Nov 27 17:41:32
GMT 2011 i686 Intel(R) Core(TM)2 Duo CPU T7500 @ 2.20GHz GenuineIntel
GNU/Linux
... although I didn't see this problem until recently. (Or maybe I
just didn't notice it ... )




On 12 December 2011 22:44, Adam Carter <adamcarter3@gmail.com> wrote:
>> I've looked at the kernel code that causes the error message. It
>> verifies that this is most likely a dead disk:
>
> It would be worth running smartctl from smartmontools to see what it
> knows of the disks status.

Having suffered with a faulty power supply for a while, I'm pretty
good with smartmontools - if you read the Google paper though, you'll
see that it only predicts failure in ~50% of cases. Thanks though!

James

Re: [gentoo-user] Re: ext4 - grow_buffers: requested out-of-range block <BLOCKID>

[Paul Hartman]

On Mon, Dec 12, 2011 at 6:03 PM, James Broadhead
<jamesbroadhead@gmail.com> wrote:
>> Does it happen to be a >2TB USB drive? I remember reading about
>> problems with some of those. It works in Windows with the factory
>> partition/FAT tables because of tricks they do to the addressing that
>> works in Windows, but once you reformat it you can't access the >2TB
>> areas. Something like that... As far as I recall, you could
>> repartition to create a 2TB or smaller partition and that would work,
>> but then the rest of the drive was inaccessible.
>
> So on returning to this machine, I see that another USB disk that I
> have connected to it is also having those messages printed about it.
> This leads me to suspect that it's either an ext4 bug or the situation
> that you mentioned above.
>
> Both are Western Digital 2TB disks;
> 1058:1130 Western Digital Technologies, Inc.
> 1058:1021 Western Digital Technologies, Inc. Elements 2TB
>
> There are 42 messages in quick succession for each disk, appearing to
> cycle through the same list of blocks twice. I'll attach the messages.
>
> I'm inclined towards the bad-usb-firmware idea - do you have a link to
> where you read about the 2TB partition problem ?
>
> I don't have much time to deal with this at the moment, so I think
> that I'll just power them down and wait until I do.

The problem I was referring to was for drives LARGER than 2TB (I think
the real limit hits somewhere around 2.1 or 2.2TB).  So if your drives
are 2TB then I don't think it's that problem.

I use a 2TB external USB drive myself (LaCie brand, with a pair of
spanned 1TB Samsung disks inside), formatted as ext4, and it works
fine. However, that was not always the case. I had to replace the USB
cable after I suffered a lot of corruption and random USB disconnects.
Later on, the drive started going offline and making the click of
death, and eventually failed to start up. It turned out to be a faulty
power supply. They sent me a replacement free of charge, despite the
drive being out of warranty, and it worked perfectly fine with the new
power supply. And it has worked fine ever since.

Re: [gentoo-user] Re: ext4 - grow_buffers: requested out-of-range block <BLOCKID>

[James Broadhead]

[-- Attachment #1: Type: text/plain, Size: 2600 bytes --]

On Dec 13, 2011 12:25 a.m., "Paul Hartman" <paul.hartman+gentoo@gmail.com>
wrote:
>
> On Mon, Dec 12, 2011 at 6:03 PM, James Broadhead
> <jamesbroadhead@gmail.com> wrote:
> >> Does it happen to be a >2TB USB drive? I remember reading about
> >> problems with some of those. It works in Windows with the factory
> >> partition/FAT tables because of tricks they do to the addressing that
> >> works in Windows, but once you reformat it you can't access the >2TB
> >> areas. Something like that... As far as I recall, you could
> >> repartition to create a 2TB or smaller partition and that would work,
> >> but then the rest of the drive was inaccessible.
> >
> > So on returning to this machine, I see that another USB disk that I
> > have connected to it is also having those messages printed about it.
> > This leads me to suspect that it's either an ext4 bug or the situation
> > that you mentioned above.
> >
> > Both are Western Digital 2TB disks;
> > 1058:1130 Western Digital Technologies, Inc.
> > 1058:1021 Western Digital Technologies, Inc. Elements 2TB
> >
> > There are 42 messages in quick succession for each disk, appearing to
> > cycle through the same list of blocks twice. I'll attach the messages.
> >
> > I'm inclined towards the bad-usb-firmware idea - do you have a link to
> > where you read about the 2TB partition problem ?
> >
> > I don't have much time to deal with this at the moment, so I think
> > that I'll just power them down and wait until I do.
>
> The problem I was referring to was for drives LARGER than 2TB (I think
> the real limit hits somewhere around 2.1 or 2.2TB).  So if your drives
> are 2TB then I don't think it's that problem.
>
> I use a 2TB external USB drive myself (LaCie brand, with a pair of
> spanned 1TB Samsung disks inside), formatted as ext4, and it works
> fine. However, that was not always the case. I had to replace the USB
> cable after I suffered a lot of corruption and random USB disconnects.
> Later on, the drive started going offline and making the click of
> death, and eventually failed to start up. It turned out to be a faulty
> power supply. They sent me a replacement free of charge, despite the
> drive being out of warranty, and it worked perfectly fine with the new
> power supply. And it has worked fine ever since.
>

Actually, a bit more triage shows that these errors are triggered on mount,
and only when using pmount. Mounting manually as root doesn't trigger them.

I'll have another look through my logs to see if they've happened at other
times, but for the moment I'll stop using pmount and see if it reoccurs.

[-- Attachment #2: Type: text/html, Size: 3245 bytes --]

Re: [gentoo-user] Re: ext4 - grow_buffers: requested out-of-range block <BLOCKID>

[Walter Dnes]

On Tue, Dec 13, 2011 at 12:03:32AM +0000, James Broadhead wrote

> So on returning to this machine, I see that another USB disk that I
> have connected to it is also having those messages printed about it.
> This leads me to suspect that it's either an ext4 bug or the situation
> that you mentioned above.
> 
> Both are Western Digital 2TB disks;
> 1058:1130 Western Digital Technologies, Inc.
> 1058:1021 Western Digital Technologies, Inc. Elements 2TB

  Could these disks be slightly larger than 2 TB?  If you you use
"make menuconfig", the setting is...

-*- Enable the block layer  --->
    [*]   Support for large (2TB+) block devices and files

  If you edit .config manually, set...

CONFIG_LBDAF=y

-- 
Walter Dnes <waltdnes@waltdnes.org>

Re: [gentoo-user] Re: ext4 - grow_buffers: requested out-of-range block <BLOCKID>

[James Broadhead]

On 2011-12-13, Walter Dnes <waltdnes@waltdnes.org> wrote:
> On Tue, Dec 13, 2011 at 12:03:32AM +0000, James Broadhead wrote
>
>> So on returning to this machine, I see that another USB disk that I
>> have connected to it is also having those messages printed about it.
>> This leads me to suspect that it's either an ext4 bug or the situation
>> that you mentioned above.
>>
>> Both are Western Digital 2TB disks;
>> 1058:1130 Western Digital Technologies, Inc.
>> 1058:1021 Western Digital Technologies, Inc. Elements 2TB
>
>   Could these disks be slightly larger than 2 TB?  If you you use
> "make menuconfig", the setting is...
>
> -*- Enable the block layer  --->
>     [*]   Support for large (2TB+) block devices and files
>
>   If you edit .config manually, set...
>
> CONFIG_LBDAF=y
>
> --
> Walter Dnes <waltdnes@waltdnes.org>
>
>

No! It's the usual SI measurement.
1.8TiB in powers-of-two.

I'll enable it anyway.

Re: [gentoo-user] Re: ext4 - grow_buffers: requested out-of-range block <BLOCKID>

[Florian Philipp]

[-- Attachment #1: Type: text/plain, Size: 1444 bytes --]

Am 12.12.2011 21:33, schrieb James Broadhead:
> On 12 December 2011 14:14, James Broadhead <jamesbroadhead@gmail.com> wrote:
>> ext4: "fill_buffer on unknown block <BLOCKID> out of range"
> 
> Apologies; the correct message is:
> grow_buffers: requested out-of-range block 18446744072382021139 for device sdb1
> 
> This appears 42 times immediately following mount.
> 
> Running picasa today, it informed me that one of the files I was
> working with was corrupted (but put the message in a box too small to
> read the full path).
> 
> This makes me think that perhaps the disk is bad.  Any advice, aside
> from the usual "get your data off asap"?
> 

I've looked at the kernel code that causes the error message. It
verifies that this is most likely a dead disk:

<quote>

commit e5657933863f43cc6bb76a54d659303dafaa9e58
Author: Andrew Morton <akpm@osdl.org>
Date:   Wed Oct 11 01:21:46 2006 -0700

[PATCH] grow_buffers() infinite loop fix

If grow_buffers() is for some reason passed a block number which wants
to lie outside the maximum-addressable pagecache range (PAGE_SIZE * 4G
bytes) then it will accidentally truncate `index' and will then
instnatiate[sic] a page at the wrong pagecache offset.  This causes
__getblk_slow() to go into an infinite loop.

This can happen with corrupted disks, or with software errors elsewhere.

Detect that, and handle it.

</quote>

Regards,
Florian Philipp


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 262 bytes --]

Re: [gentoo-user] Re: ext4 - grow_buffers: requested out-of-range block <BLOCKID>

[Adam Carter]

> I've looked at the kernel code that causes the error message. It
> verifies that this is most likely a dead disk:

It would be worth running smartctl from smartmontools to see what it
knows of the disks status.