New: use GFS2/DLM without a cluster

New: use GFS2/DLM without a cluster

[David Teigland]

A new daemon "dlm_sand" allows GFS2/DLM to be used without a cluster.
dlm_sand is an alternative to dlm_controld.  Where dlm_controld uses
corosync to coordinate GFS2/DLM usage and recovery, dlm_sand does the
same coordination using the SAN storage that GFS2 is already using.

This makes GFS2/DLM usage simpler, more self contained, and suited to use
cases that are more data-oriented, rather than clustering/HA oriented.
The behavior and requirements of a traditional cluster can be intrusive
and limiting when the sole interest is just using the GFS2 file system.

At the link below you'll find an initial version of dlm_sand that should
work well enough for basic usage.  The README includes how to use it, how
it works, and current limitations.  I'd like to hear if you'd find it
useful, how it works for you, or any suggestions.  Thanks!

https://pagure.io/dlm/blob/main/f/dlm_sand