[Gluster-users] It necessary make backup the .glusterfs directory ?

Ravishankar N ravishankar at redhat.com
Sat Jan 27 03:57:14 UTC 2018



On 01/26/2018 06:17 PM, César E. Portela wrote:
> El 25-01-2018 22:27, Ravishankar N escribió:
>> On 01/25/2018 02:14 AM, César E. Portela wrote:
>>> Hi All,
>>>
>>> I have two glusterfs servers and doing the backup of these is very 
>>> slow, when it does not fail.
>>> I have thousand and thousand and thousand files...
>>>
>>> Apparently the directory .glusterfs has some responsibility for the 
>>> backup failure.
>>>
>>> Is necessary to make a backup of the .glusterfs directory?
>>
>> Not sure what kind of backup you're doing but .glusterfs contents are
>> essential for things to work properly. Have you tried using
>> replication/geo-replication/volume snapshot features that gluster
>> provides?
>
> Hi Ravishankar
>
> I make external full/incremental backup the applicattion's data. Only 
> one server, not all. I not using special features the gluster.
> I think if I need restore files, .glusterfs it is rebuilt when I copy 
> the files in the filesystem. Is this correct?
So you are copying off data from the mount point to your backup location 
and 'restore' it by again copying from the backup on to the volume via 
the mount point, then you need not worry about .glusterfs directory. I 
thought you were copying files directly from the backend bricks.

-Ravi



More information about the Gluster-users mailing list