[Gluster-users] Recommended filesystem for GlusterFS bricks.

[Kaleb KEITHLEY]

On 07/03/2013 11:25 AM, Kurian Thayil wrote:
> On Wed, 3 Jul 2013 11:12:25 -0400 (EDT)
> Ben Turner <bturner at redhat.com> wrote:
>
>> ----- Original Message -----
>>> From: "Kurian Thayil" <kurian at gnuhack.com>
>>> To: "Bobby Jacob" <bobby.jacob at alshaya.com>
>>> Cc: gluster-users at gluster.org
>>> Sent: Wednesday, July 3, 2013 10:08:04 AM
>>> Subject: Re: [Gluster-users] Recommended filesystem for GlusterFS
>>> bricks.
>>>
>>> On Wed, 3 Jul 2013 13:52:14 +0000
>>> Bobby Jacob <bobby.jacob at alshaya.com> wrote:
>>>
>>>> Hi,
>>>>
>>>> Which is the recommended filesystem to be used for the bricks in
>>>> glusterFS. ?? XFS/EXT3/EXT4 etc .????
>>>>
>>>> Thanks & Regards,
>>>>
>>>> Bobby Jacob
>>>> Senior Technical Systems Engineer | eGroup
>>>> P SAVE TREES. Please don't print this e-mail unless you really
>>>> need to.
>>>>
>>>
>>> XFS is recommended for bricks.
>>
>> Be sure to set the inode size to 512 to allow enough room for
>> glusterfs's xattrs:
>>
>> mkfs.xfs -i size=512 <my block device>
>>
>> -b
>>
>>> Regards,
>>> K.
>>> _______________________________________________
>>> Gluster-users mailing list
>>> Gluster-users at gluster.org
>>> http://supercolony.gluster.org/mailman/listinfo/gluster-users
>>>
>
> Hmm.. Something which I didn't do when builing the cluster :(
>
> Is there any other way other than mkfs.xfs paritition to
> change this attribute?
>
> currently xfs_bmap gives the floowing
>
> xfs_bmap -a index.html
> index.html:
> 	0: [0..7]: 616184..616191
>

No, but I wouldn't worry about it. The latest results from our 
performance engineering team show that the default values are generally 
the best choice.

OTOH, you should mount with inode64 mount option.

--

Kaleb