Hide Forgot
(In reply to comment #0) > Bricks: > Brick1: 10.168.100.242:/var/tmp/test0 oops forgot to clean up the host address & path, that should read Brick1: server:/brick/test0
The gluster CLI allows removal of a single brick from a replicate only volume. This can lead to invalid Replicate configurations such as "Number of Bricks: 1" or even worse, "Number of Bricks: 1 x 2 = 3" I've tested this on 3.1.x, 3.2.x, and git master 3.3.0qa9+, as far as I can tell it exists in all versions of gluster CLI. To reproduce the behavior... # gluster volume create test replica 2 server:/brick/test{0..1} # gluster volume start test # gluster volume info Volume Name: test Type: Replicate Status: Started Number of Bricks: 2 Transport-type: tcp Bricks: Brick1: server:/brick/test0 Brick2: server:/brick/test1 # gluster volume remove-brick test server:/brick/test1 # gluster volume info Volume Name: test Type: Replicate Status: Started Number of Bricks: 1 Transport-type: tcp Bricks: Brick1: 10.168.100.242:/var/tmp/test0 # gluster volume add-brick test server:/brick/test{2..3} # gluster volume info Volume Name: test Type: Distributed-Replicate Status: Started Number of Bricks: 1 x 2 = 3 Transport-type: tcp Bricks: Brick1: server:/brick/test0 Brick2: server:/brick/test2 Brick3: server:/brick/test3 Thank you.
With 'volume type' change 'bug 765037' (patch at http://review.gluster.com/158) this should not be a problem. I will keep it open till that patch gets merged.
bug 765037 is now closed. Not planing to backport it to release-3.2 branch, and hence if you test the git 'master' branch, you should see this problem going away. Please re-open if still exists.