When deploying using automation tools, follow-up steps will fail as they follow too closely to glusterd daemonizing. This causes automated deployments to take multiple runs to create a GlusterFS cluster.
If we can add the ability to have daemonizing wait until everything is active and running this would solve this problem.
Thanks Joe. here's the relevant commit and hack that I use to get around this :P
I don't mind though.
Upstart, and probably systemd & other event driven init systems, would benefit from this, in cases where a client mounts from localhost during boot. The service started event should not fire until the service is really ready to serve requests. Currently if mounts are tried immediately after the service starts they will fail, because it's not really ready when it daemonizes.
Migrated to github:
Please follow the github issue for further updates on this bug report.