Okay, soooo a specific permutation of the replace-brick command does seem to do what i was looking for. In this case server1 is being replaced by server2. The command executes successfully and I then see the data on server 2.
gluster volume replace-brick gv0 server1:/data/glusterfs/gv0/brick1/brick server2:/data/glusterfs/gv0/brick1/brick commit force
However, the reason I had initially stopped trying to use replace-brick is due to this post in the list server saying that the command is going to be deprecated and that there was a better way. So what are peoples opinions? Is replace-brick the right way to do this or should be handled differently?
Thanks,
Joe
On Sun, Feb 5, 2017 at 3:40 PM, Joseph Lorenzini <jaloren@xxxxxxxxx> wrote:
All:I am quite new to gluster so this is likely my lack of knowledge. Here's my scenario: I have a distribute replicate volume with a replica count of 3. Each brick is on a different server and the total number of bricks in the volume is 3.Now lets say one server goes bad or down. Now i want to bring up a new server with a single brick on that, add that into my volume, and then replicate a copy of all the files from the one of the existing bricks to the new one.What's the procedure for doing that?At first I thought, I would just add a new brick in and then remove the other brick. However, that didn't seem to work. The add brick command wanted a new set of bricks that match the volume replica count (so in this case three new bricks). Moreover this was a different replica set, which of course defeats the purpose of what I am trying to do.Thanks,JoePS I did try using replace-brick but discovered in the list serve this was considered a deprecated command as far back as 2012.
_______________________________________________ Gluster-users mailing list Gluster-users@xxxxxxxxxxx http://lists.gluster.org/mailman/listinfo/gluster-users
