Hi! I've a doubt about the Rebalancing system in Gluster. The structure of my Volume is: Server1 Brick0 Brick2 Server2 Brick1 Brick3 I've create 16 files in the mount folder and I expected to have 8 files in each brick, and two replicas (files of Brick0= files of Brick1, and, files of Brick2= files of Brick3). But in practice, gluster puts 6 files in one brick and 9 on the other brick. I've executed the command to rebalance and migrate files. Is this behavior correct? Can I configure that one brick has more files than other? root@NCSL008:/home# ls -la brick1 total 1732276 drwxr-xr-x 3 root root 4096 Oct 17 12:34 . drwxr-xr-x 5 root root 4096 Oct 17 11:41 .. -rw-r--r-- 2 root root 220514438 Oct 17 12:29 big_buck_bunny_480p_surround-fix2.avi -rw-r--r-- 2 root root 220514438 Oct 17 12:30 big_buck_bunny_480p_surround-fix3.avi -rw-r--r-- 2 root root 220514438 Oct 17 12:31 big_buck_bunny_480p_surround-fix5.avi -rw-r--r-- 2 root root 220514438 Oct 17 12:34 big_buck_bunny_480p_surround-fix9.avi -rw-r--r-- 2 root root 445866736 Oct 17 11:59 ED_1025.avi -rw-r--r-- 2 root root 445866736 Oct 17 12:00 ED_1026.avi drw------- 11 root root 4096 Oct 17 12:34 .glusterfs root@NCSL008:/home# ls -la brick3 total 2378332 drwxr-xr-x 3 root root 4096 Oct 17 12:34 . drwxr-xr-x 5 root root 4096 Oct 17 11:41 .. -rw-r--r-- 2 root root 220514438 Oct 17 12:34 big_buck_bunny_480p_surround-fix0.avi -rw-r--r-- 2 root root 220514438 Oct 17 12:01 big_buck_bunny_480p_surround-fix1.avi -rw-r--r-- 2 root root 220514438 Oct 17 12:30 big_buck_bunny_480p_surround-fix4.avi -rw-r--r-- 2 root root 220514438 Oct 17 12:31 big_buck_bunny_480p_surround-fix6.avi -rw-r--r-- 2 root root 220514438 Oct 17 12:31 big_buck_bunny_480p_surround-fix7.avi -rw-r--r-- 2 root root 220514438 Oct 17 12:31 big_buck_bunny_480p_surround-fix8.avi -rw-r--r-- 2 root root 220514438 May 31 2008 big_buck_bunny_480p_surround-fix.avi -rw-r--r-- 2 root root 445866736 Oct 17 11:55 ED_1024.avi -rw-r--r-- 2 root root 445866736 Oct 17 12:00 ED_1027.avi drw------- 13 root root 4096 Oct 17 12:34 .glusterfs root@NCSL008:/home# gluster volume status test-volume detail Status of volume: test-volume ------------------------------------------------------------------------------ Brick : Brick XXXX1:/home/brick0 Port : 24009 Online : Y Pid : 1270 File System : ext4 Device : /dev/mapper/vg0-root Mount Options : rw,errors=remount-ro,user_xattr Inode Size : 256 Disk Space Free : 142.5GB Total Disk Space : 153.9GB Inode Count : 10100736 Free Inodes : 10014569 ------------------------------------------------------------------------------ Brick : Brick XXXX2:/home/brick1 Port : 24014 Online : Y Pid : 1010 File System : ext4 Device : /dev/mapper/vg0-root Mount Options : rw,errors=remount-ro Inode Size : 256 Disk Space Free : 143.5GB Total Disk Space : 153.9GB Inode Count : 10100736 Free Inodes : 10028805 ------------------------------------------------------------------------------ Brick : Brick XXXX1:/home/brick2 Port : 24010 Online : Y Pid : 1276 File System : ext4 Device : /dev/mapper/vg0-root Mount Options : rw,errors=remount-ro,user_xattr Inode Size : 256 Disk Space Free : 142.5GB Total Disk Space : 153.9GB Inode Count : 10100736 Free Inodes : 10014569 ------------------------------------------------------------------------------ Brick : Brick XXXX2:/home/brick3 Port : 24015 Online : Y Pid : 1016 File System : ext4 Device : /dev/mapper/vg0-root Mount Options : rw,errors=remount-ro Inode Size : 256 Disk Space Free : 143.5GB Total Disk Space : 153.9GB Inode Count : 10100736 Free Inodes : 10028805
Please refer to http://hekafs.org/index.php/2012/03/glusterfs-algorithms-distribution/ for understanding how files are placed in glusterfs. It need not be equally distributed for the names that you chose. it all depends on 'file-names' you choose. Closing this bug as it is not a bug.