Re: XFS Calltraces by using XFS with Ceph

[list@jonas-server.de]

Am 2017-06-16 19:46, schrieb Darrick J. Wong:
> On Fri, Jun 16, 2017 at 07:37:11AM +0200, list@jonas-server.de wrote:
>> Am 2017-06-15 19:20, schrieb Emmanuel Florac:
>> >Le Thu, 15 Jun 2017 07:55:12 +0200
>> >list@jonas-server.de écrivait:
>> >
>> >>> Looking at the new Elixir[1], it looks like we're trying to deref
>> >>> ((*bpp)->b_addr)->magic, so that might explain the crash you see.
>> >>>
>> >>> --D
>> >>>
>> >>> [1]
>> >>> http://elixir.free-electrons.com/linux/v4.4.72/source/fs/xfs/libxfs/xfs_da_btree.c#L270
>> >>>
>> >>>>
>> >
>> >Indeed that matches exactly the crash Darrick mentioned. Looks like you
>> >need either a special hand-patched kernel... or wait for an official
>> >fix for this one :/
>> 
>> When will a fix normally be in the ubuntu kernel? I need to have a 
>> stable
>> distribution kernel, I can not install a patched version.
> 
> Like Emmanuel said, first a fix has to go upstream, then it can go to 
> the
> 4.4 stable kernel, and then it'll probably end up in the Ubuntu kernel.
> 
> Can you post a metadump of the xfs filesystem somewhere for further
> analysis?  I'm pretty sure it's the same problem reported by David Shaw
> on this list a couple of weeks ago.  Since you can't run one-off test
> kernels, send us a metadump for recreation in a lab is the only other
> way to figure out if the fix works.

So it is hard for me to run tests with untested kernels as the system is 
already
used in production. I've created a metadump from the xfs filesystem and
uploaded it:
https://transfer.sh/XdlZ5/xfs_metadump.gz

I hope it helps you trigger the bug during your tests.

> 
> (And yes, I /do/ have a lightly tested fix which I'm about to send 
> out.)
> 
> --D
> 
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-xfs" 
>> in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html