This is a problem for me too as I have provisioned a limited number of glusterfs volumes on our pilot platform at IG and I don't think that manually recycling failed volumes is a workable solution for me.
I've come up with an interim work-around whilst we wait for the plugin to be developed or the more exciting fully automatic provisioning to arrive. The temporary solution is to run a docker container that does the job of picking up gluster volumes that are in the failed state, wiping their files and returning them back to the pool of available volumes. I've used the glusterfs-centos image as a base and written a simple shell script to perform the recycling process. The container is intended to be deployed by cluster-admins and is available on dockerhub
https://hub.docker.com/r/davemccormickig/gluster-recycler/ and the project files on github https://github.com/davemccormickig/gluster-recycler.
I hope that this is useful to other admins trialling the use of glusterfs volumes with their openshift clusters.
On 24 April 2016 at 11:32, David McCormick <davidemccormick gmail com> wrote: