On 11/9/2012 7:51 AM, Anselm Strauss wrote: > ah, i just noticed i tested this on gluster 3.2 not 3.3. also in the > changelog to 3.3 it says that remove-brick now migrates data to > remaining bricks. If your 3.3 volume is near or over half full, the migration is likely to completely fill up your volume and then fail partway through. https://bugzilla.redhat.com/show_bug.cgi?id=862347 I have speculated that if I were to abort the remove-brick before the other bricks filled up completely, I could repeat that cycle and eventually it would finish with no errors. I have not actually tested this, though. Thanks, Shawn
