NBD: exported files over something around 1 TiB get an insane device size on the client side and are actually empty

NBD: exported files over something around 1 TiB get an insane device size on the client side and are actually empty

[Christoph Anton Mitterer]

[-- Attachment #1: Type: text/plain, Size: 870 bytes --]

Hi.

On advice from Alex Bligh I'd like to ping linux-block and nbd-general
about the issue described here:
https://github.com/NetworkBlockDevice/nbd/issues/44

What basically happens is, that with a recent kernel (Linux heisenberg
4.9.0-1-amd64 #1 SMP Debian 4.9.2-2 (2017-01-12) x86_64 GNU/Linux),
when a device larger than something between 1 TiB and 2 TiB is exported
via nbd-server and connected to via nbd-client (which uses the kernel
driver, AFAIU) than the device size gets insanely large while reading
from /dev/nbd0 gives actually nothing.


The bug does not seem to happen when the kernel is not involved (we
tried using nbd-server and qemu-img as the client).

Don't think it makes all too much sense to copy & paste everything from
what was tried already for testing here, so please have a look at the
issue on github.


Best,
Chris.

[-- Attachment #2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 5930 bytes --]

Re: NBD: exported files over something around 1 TiB get an insane device size on the client side and are actually empty

[Josef Bacik]

On Sat, Jan 14, 2017 at 6:31 PM, Christoph Anton Mitterer 
<calestyo@scientia.net> wrote:
> Hi.
> 
> On advice from Alex Bligh I'd like to ping linux-block and nbd-general
> about the issue described here:
> https://github.com/NetworkBlockDevice/nbd/issues/44
> 
> What basically happens is, that with a recent kernel (Linux heisenberg
> 4.9.0-1-amd64 #1 SMP Debian 4.9.2-2 (2017-01-12) x86_64 GNU/Linux),
> when a device larger than something between 1 TiB and 2 TiB is 
> exported
> via nbd-server and connected to via nbd-client (which uses the kernel
> driver, AFAIU) than the device size gets insanely large while reading
> from /dev/nbd0 gives actually nothing.
> 
> 
> The bug does not seem to happen when the kernel is not involved (we
> tried using nbd-server and qemu-img as the client).
> 
> Don't think it makes all too much sense to copy & paste everything 
> from
> what was tried already for testing here, so please have a look at the
> issue on github.

Yeah I noticed this in testing, there's a bug with NBD since it's 
inception where it uses a 32 bit number for keeping track of the size 
of the device, I fixed it with

nbd: use loff_t for blocksize and nbd_set_size args

It'll be in 4.10.  Thanks,

Josef

Re: NBD: exported files over something around 1 TiB get an insane device size on the client side and are actually empty

[Christoph Anton Mitterer]

[-- Attachment #1: Type: text/plain, Size: 259 bytes --]

On Sat, 2017-01-14 at 20:03 -0500, Josef Bacik wrote:
> nbd: use loff_t for blocksize and nbd_set_size args

I'm just trying your patch, which does however not apply to 4.9.2 (but
I've adapted it)... will tell later if it worked for me.

Cheers,
Chris.

[-- Attachment #2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 5930 bytes --]

Re: NBD: exported files over something around 1 TiB get an insane device size on the client side and are actually empty

[Christoph Anton Mitterer]

[-- Attachment #1: Type: text/plain, Size: 177 bytes --]

On Sat, 2017-01-14 at 20:03 -0500, Josef Bacik wrote:
> nbd: use loff_t for blocksize and nbd_set_size args

Tested it, and it seems to fix the issue :-)


Cheers,
Chris.

[-- Attachment #2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 5930 bytes --]

Re: [Nbd] NBD: exported files over something around 1 TiB get an insane device size on the client side and are actually empty

[Alex Bligh]

> On 15 Jan 2017, at 01:03, Josef Bacik <jbacik@fb.com> wrote:
>=20
> Yeah I noticed this in testing, there's a bug with NBD since it's=20
> inception where it uses a 32 bit number for keeping track of the size=20=

> of the device, I fixed it with
>=20
> nbd: use loff_t for blocksize and nbd_set_size args
>=20
> It'll be in 4.10.  Thanks,

Interestingly enough, this is a regression, as the random 3.x kernel I =
tried did not show the issue.

Thanks for the fix Josef.

--=20
Alex Bligh