[Gluster-users] remove-brick very slow for (distributed-)replicated volumes?

[B.K.Raghuram]

We have a gigabit ethernet lan on which there is no other traffic and
I am getting the following numbers when I do a remove-brick. The
sequence of steps is that I create a 2 way replicated volume, populate
it with 300 files totalling 100MB. I then add a pair of bricks to the
volume and then a remove-brick on the original two bricks.

Is this the expected speed for the operation or could there be an
issue? The only problem that I suspect is that the rebalance log for
the volume has the following kind of line for each file so I dont know
if this check is slowing down the process:

[2013-11-06 11:36:29.307612] I [dht-rebalance.c:860:dht_migrate_file]
0-v1-dht: completed migration of /file299 from subvolume
v1-replicate-0 to v1-replicate-1


[root at s5n9 ~]# gluster volume remove-brick v1
s5n9.testing.lan:/data/v1 s5n10.testing.lan:/data/v1 status
                                    Node Rebalanced-files
size       scanned      failures       skipped         status run-time
in secs
                               ---------      -----------
-----------   -----------   -----------   -----------   ------------
--------------
                               localhost              300
100.1MB           600             0      completed           600.00
                       s6n12.testing.lan                0
0Bytes             0             0    not started             0.00
                       s6n11.testing.lan                0
0Bytes             0             0    not started             0.00
                       s5n10.testing.lan                0
0Bytes           301             0      completed             2.00