[Gluster-devel] Stateless Gluster 3.x series

[Ian Latter]

Hello,


  Until yesterday all of my experiments with Gluster have
been with a vi'd vol file and the Glusterfsd process.  
Michael and I have now done testing with glusterd under 
Gluster version 3.2.6 and this has raised some issues.  
The one I'd like to query in this thread is stateless 
environments.

  If I have a PXE distribution that boots, tftp's a config
file, mounts storage devices and then initialises gluster
then;

  Previously - I wrote out the appropriate vol file and 
started the necessary daemons.

  Now - I start the necessary daemon and issue a set
of commands to build the appropriate configuration.

  .. however, this results in a new gluster UUID each
time the node restarts.

 So;

  1) Is there a gluster command to configure gluster
      with a specific UUID?

  -or-

  2) Is there a means to provide gluster with UUID
      data outside of the gluster commands

  -or-

  3) Is there a means/API to supply gluster with a 
       complete cluster map (i.e. of all nodes)

  -and-

  4) Is there a gluster command to disable peer 
       probe responses from the local glusterd?

      

  I have a daemon that I developed to robustly 
replace an older daemon -- tyd -- that was developed
by Ty Miller, Rob Dartnell and I, for openMosix - to do
essentially the same thing.  But openMosix provided 
a /proc API for this capability.


Thanks,



--
Ian Latter
Late night coder ..
http://midnightcode.org/