new btrfs-progs needs testing and karma

[Chris Murphy]

On Nov 22, 2013, at 1:14 PM, Gene Czarcinski <gene at czarc.net> wrote:

> On 11/21/2013 10:25 PM, Chris Murphy wrote:
>> On Nov 21, 2013, at 8:04 PM, Adam Williamson <awilliam at redhat.com> wrote:
>> 
>>> On Thu, 2013-11-21 at 19:52 -0700, Chris Murphy wrote:
>>>> On Nov 21, 2013, at 5:08 PM, Adam Williamson <awilliam at redhat.com> wrote:
>>>> 
>>>>> On Thu, 2013-11-21 at 14:53 -0700, Chris Murphy wrote:
>>>>>> btrfs-progs-0.20.rc1.20131114git9f0c53f-1.fc20 incorporates 16KB
>>>>>> leafsize/nodesize by default, which significantly reduces metadata
>>>>>> fragmentation and improves performance. Since this is set at the time
>>>>>> the file system is created, it would be nice to get this in before
>>>>>> freeze. Since it's the new default, it's considered safe.
>>>>>> 
>>>>>> There is one minor regression found, which is 'btrfs filesystem show'
>>>>>> reports duplicate volumes. A fix is committed upstream already and a
>>>>>> post-install update will fix this.
>>>>> Why not just get it in now? Final doesn't freeze for another week.
>>>> It needs 2 more karma points, doesn't it?
>>> What? I'm talking about the fix that sounds like it hasn't been put into
>>> Fedora at all yet.
>> Oh, got it. That fix was committed upstream a few days ago. It's a question for Eric Sandeen or Josef if it's possible.
>> 
>> https://bugzilla.redhat.com/show_bug.cgi?id=1031299
>> 
>> 
>> Chris Murphy
> Raher than just "btrfs fi show", give "btrfs fi show -m" a try.
> 
> Until the patch is in it makes this a little easier.  Another problem however is that this update seriously breaks system-storage-manager (which the package maintainer admitted is badly backlevel).

ssm was broken with btrfs with the previous btrfs-progs also, just not as seriously. And as ssm isn't a default tool, *shrug* I'm not overly worried about it. But otherwise I do like ssm. It's nifty.


Chris Murphy