There has been discussion in the past about improvement in the way glusterfs
handles unequal components in a distribute, for instance unequal size brick
pairs in a distribute/replicate cluster. Has this happened and I missed it?
I don't see this listed as a feature in any of the new glusterfs releases
nor in the the proposes improvements in glusterfs 3.6. Perhaps it occured
under some other name? When one wants to add space to an old cluster or
replace dead disks the sweet spot in the size of disks usually will have
increased from the size used when the cluster was first built. I can see that
one would want to have the bricks making up a replicate to be the same size but
for distribute components it is inefficient to have to divide a big new disk
into a numerous small bricks.
Bill Sebok Computer Software Manager, Univ. of Maryland, Astronomy
Internet: [email protected] URL: http://furo.astro.umd.edu/
_______________________________________________
Gluster-users mailing list
[email protected]
http://supercolony.gluster.org/mailman/listinfo/gluster-users
