[Gluster-users] Replication logic

Zenon Panoussis oracle at provocation.net
Wed Jan 6 02:57:46 UTC 2021

>> https://gluster.readthedocs.io/en/latest/Administrator-Guide/Managing-Volumes/#triggering-self-heal-on-replicate

>> says "NUFA should be enabled before creating any data in the
>> volume". What happens if one tries to enable it with data in
>> the volume? Will it refuse, or will it corrupt the data, or
>> will it work OK under certain conditions?

> I don't see that option under Gluster v7, 

Now that you say it, I checked 'gluster volume get <volume> all
|grep cluster'. Indeed, cluster.nufa does not exist. I'm using 8.3.

But check the link above, it says /latest/ in the URL. This means
that the documentation is severely lagging behind development.

> but I got a 'cluster.choose-local':
> Option: cluster.choose-local
> Default Value: true
> Description: Choose a local subvolume (i.e. Brick) to read from if
> read-subvolume is not explicitly set.
> It's nice to verify that it is enabled for you, as you want to read
> from the local brick instead of the high-lattency one.

I checked and it is on. It is a very sensible default. Actually,
not even knowing about this option, I had taken it for granted
that mounting a volume as localhost:/vol would always read from
the local brick first.



More information about the Gluster-users mailing list