[Gluster-devel] Borking Gluster

[James]

Due to some weird automation things, I noticed the following:

Given a cluster of hosts A,B,C,D

It turns out that if you restart glusterd on host B while you are
running volume create on host A, this can cause host B to be borked.
This means: glusterd will refuse to start, and the only fix I found
was to delete the volume data from it, and re-create the volume. Not
sure if this is useful or not, but reproducing this is pretty easy in
case this uncovers a code path that isn't working properly.

HTH,
James