[Gluster-users] Rebalance failures

[James Devine]

I am messing around with gluster management and I've added a couple bricks
and did a rebalance, first fix-layout and then migrate data.  When I do
this I seem to get a lot of failures:


gluster> volume rebalance MAIL status
                                    Node Rebalanced-files          size
  scanned      failures         status
                               ---------      -----------   -----------
-----------   -----------   ------------
                               localhost             6657     36611748
   71014        10458      completed
                            gs3                      3528     21167454
   51122         8491      completed
                            gs2                        0            0
 45079            0      completed
                            gs4                         0            0
   45069            0      completed


The logs for this show:

[2012-06-28 13:35:54.100842] I [dht-rebalance.c:639:dht_migrate_file]
0-MAIL-dht:
/testuser393/maildir/new/1340909913.V14Ia51f68d6a66824a5M364717.test-gluster-client1:
attempting to move from MAIL-replicate-1 to MAIL-replicate-0
[2012-06-28 13:35:54.111880] W [dht-rebalance.c:353:__dht_check_free_space]
0-MAIL-dht: data movement attempted from node (MAIL-replicate-1) with
higher disk space to a node (MAIL-replicate-0) with lesser disk space
(/testuser393/maildir/new/1340909913.V14Ia51f68d6a66824a5M364717.test-gluster-client1)
[2012-06-28 13:35:54.111947] E
[dht-rebalance.c:1194:gf_defrag_migrate_data] 0-MAIL-dht: migrate-data
failed for
/testuser393/maildir/new/1340909913.V14Ia51f68d6a66824a5M364717.test-gluster-client1



All the gluster servers have the same amount of disk space so I'm confused
by the failure.  Is this something to be worried about?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20120628/ff912c1a/attachment.html>