[Gluster-users] How does volume rebalance work?

[Olivier Nicole]

Hi,

I just did another test this morning.

After rebalancing, I get an error in the logs
/var/log/gluster/etc-glusterfs-glusterd.vol.log about file size:

[2011-09-29 09:36:39.703090] W [glusterd-rebalance.c:251:gf_glusterd_rebalance_move_data] 0-glusterfs: file sizes are not same : /etc/glusterd/mount/test-vol/ubuntu-9.10-server-i386.iso

But from the NFS client, I checked the original file and the temporary
copy, and they are the same:


banyan<root>: md5 ubuntu-9.10-server-i386.iso
MD5 (ubuntu-9.10-server-i386.iso) = 55618ad5f180692f9dac20cbff352634
banyan<root>: md5 .ubuntu-9.10-server-i386.iso.gfs671686656 
MD5 (.ubuntu-9.10-server-i386.iso.gfs671686656) = 55618ad5f180692f9dac20cbff352634
banyan<root>: ls -lart |grep ubuntu-9.10-server-i386.iso
-rw-r--r--  1 root  wheel  671686656 Sep 29 09:03 ubuntu-9.10-server-i386.iso
-rw-r--r--  1 root  wheel  671686656 Sep 29 09:36 .ubuntu-9.10-server-i386.iso.gfs671686656


What is happening?

TIA,

Olivier