[Date Prev] [Date Index] [Date Next] [Thread Prev] [Thread Index] [Thread Next]

Re: Conserver protocols.

Bryan Stansell bryan@conserver.com
Thu, 3 May 2007 11:55:49 -0700 (PDT)


On Wed, May 02, 2007 at 03:44:32PM -0700, Phillip Pacheco wrote:
> Could you please answer these questions to satisfy my boss'
> concerns/curiosity?

questions are no problem...i'll take a whack at 'em...

> 1.	Does config file need to be identical on each client server?  My
> boss would like to generate dynamic config files on each client for a
> decentralized configuration.  Is this possible?

no, they don't need to be identical.  the one reason for doing that is
to allow all the servers to know about eachother and redirect the client
to the appropriate server.  if each contain only their consoles, then
the user will need to direct the client to the appropriate server.  and,
of course, any other combination in between could be used.  multiple
servers in a particular datacenter or security zone could share console
definitions and redirect between them (which means users would just need
to know the datacenter or zone to point to), etc.

> 2.	What facility/protocol is used for communications between the
> client server and the master server?  To be clear, we currently use a
> master server, a few client servers, and several Cyclades console
> servers.  Is the communications protocol between the console port and
> the Cyclades device the same protocol as between the client server and
> the master?

by 'client server' and 'master server' i'm going to assume 'master
server' is just the default server a user connects to when using
'console'.

there is no communcation between any of the servers.  the client
('console') does all the work.  the only thing they can do is tell the
client where to go when they don't manage a console (by sharing the
conserver.cf file or at least populating the other server's
information).

as for the last question, there are multiple possibilities based on what
you mean by 'communications protocol between the console port and the
cyclades device'.  first possibility (which i'm guessing you're not
talking about):  many cyclades devices run conserver internally (they've
bundled the code in and some folks cross-compile and run it locally) and
the user can connect to it just like any other conserver host.  in this
case, it's just like another conserver instance...you can redirect users
to it as above, but there is no direct communication with other
conserver processes.  second possibility: you're using conserver and
connecting (via a tcp connection) to console ports on the cyclades.  in
this case the protocol is whatever the connection type you're
using...raw, telnet, ssh, etc.  either way, since there is no
direct communication between conserver instances, there's no way the
protocol can be the same.  ;-)

> Thanks for your help.  The Conserver software is a great product, and
> your support has been above and beyond.

no problem...and thanks!  hopefully the ramblings help.  if i just made
things more confusing or it brings up more questions, let me know!

Bryan