[Gluster-users] gluster volume create

Andreas Hollaus andreas.hollaus at ericsson.com
Sat Apr 11 08:55:03 UTC 2015


Hi and thanks,

Well, I guess I didn't explain my problem particularly well. Sorry for that.

I guess that a normal gluster user would in the beginning add the 
replica bricks required and then, whenever the servers restarts, let 
gluster restart using the same configuration files as before the 
restart. Restarts would happen again and again, without any manual 
interaction changing the storage configuration. Every time you add a 
brick, you would need to sync the new brick for it to be a replica of 
the other bricks and that would take some time.

My case is worse. When setting up this replicated file system, I would 
do the same. However, at an early stage of every server restart I would 
need to make the local brick available to my clients on all servers. 
That means before the servers get in contact with each other.

At this stage, I would like to have a volume which is not replicated on 
every server. After a while, these single brick volumes would be dropped 
and replaced by the replicated volume I used before the restart.

All these stages (and volumes) I describe would make use of the same 
physical storage place (bricks). As long as no writing is involved with 
the single brick volumes, I wouldn't risk ending up in a split brain 
situation when I restore the replicated volume.

Restoring the volume would be done by keeping a backup of the 
configuration files which define the replicated volume. Restarting the 
gluster daemons using these configuration files would restore the 
replicated volume without having to add any bricks. The reason I don't 
want to add the brick at each restart, is that it would mean some 
penalty to sync the new brick. That new brick already contains all the 
necessary files from the time before the server restart.

Sorry for this complex and strange way of using glusterfs, but my task 
is to try to find a way to use it given a set of fixed use cases that we 
have. I know that this is not an easy thing to do, but I would at least 
want to find some way to implement this (good or bad). I think it's 
possible, but it may be too inefficient to be a good solution.

So, back to my original question: Would this 'gluster volume create 
myvol local_IP:/path/to/brick' affect the data or meta-data on the brick 
in any way? I think that is the most critical command in the chain, as I 
don't expect starting or mounting the volume to affect anything. I want 
to avoid affecting the (meta-)data as I intend to restore the replicated 
volume I used before the restart without gluster noticing that the brick 
was actually used in another configuration in between.


Regards
Andreas


On 04/11/2015 09:18 AM, Atin Mukherjee wrote:
>
> On 04/11/2015 01:21 AM, Andreas Hollaus wrote:
>> Hi,
>>
>> I wonder what happens when the command 'gluster volume create...' is
>> executed? How is the file system on the brick affected by that command,
>> data and meta-data (extended attributes)?
>>
>> The reason for this question is that I have a strange use case where my
>> 2 (mirrored) servers are restarted. At an early stage of the reboot
>> phase, I have to create a new gluster file system on each server so that
>> the same directory can be used for read access. Later on, I would like
>> to delete these single server volumes and replace them with the mirrored
>> gluster volume I used before the restart.
>>
>> I guess I can restore the previous volume definition from a backup of
>> the gluster configuration files, but I'm worried that the 'gluster
>> volume create...' command might have affected the brick so that it is in
>> a different state compared to before the restart, when the restored
>> gluster configuration was valid. I realize that for this to work, the
>> extended attributes can not change while the mirrored volume is stopped.
>>
>> Any idea if I can use glusterfs like this or am I violating some rules?
> If I've understood your requirement correctly you are trying to convert
> a distributed volume into a replicate one. In that case you would need
> to convert it through add brick mentioned the replica count. I don't
> think you would face any issues for restoring configurations if all the
> steps are performed correctly.
>
> ~Atin
>>
>> Regards
>> Andreas
>> _______________________________________________
>> Gluster-users mailing list
>> Gluster-users at gluster.org
>> http://www.gluster.org/mailman/listinfo/gluster-users
>>



More information about the Gluster-users mailing list