Mailing List Archive

The method for this is two-fold...

First, if each conserver is in a separate DNS domain, then each of the
console servers should have an alias of "conosle" in the 'local' domain.
(This way, the clients point to their 'local' conserver...)

Second, you need to add an argument for each port of the conserver.cf
file, which points to the hostname of the conserver which is doing the
logging for each port. That is, if ts-1 through ts-9 are logged on
cnsvr-1, and ts-20 through ts-29 are logged on cnsvr-2, then console
entries on ts-1 have the "at cnsvr-1" appended to each line. Similarly,
the console entries on ts-20 would have "at cnsvr-2" appended to each
line.

Then, when users at either site invoke the console client, they
connect by default to the hostname "console" (so, one client binary
works at all sites...). If someone in the cnsvr-1 domaion is trying to
get to a console that is logged on cnsvr-2, then;

1) the client connects to cnsvr-1, to get access to the port, gets
authenticated
2) cnsvr-1 sees that the requested console is hosted 'at cnsvr-2'
3) cnsvr-1 redirects the client to cnsvr-2...
4) the client contacts cnsvr-2, and goes through any authentication
needed
5) the client session, and the logging session are joined, and the
user has access to the console at the remote site.

-Z-

-----Original Message-----
From: users-bounces@conserver.com [mailto:users-bounces@conserver.com]
On Behalf Of Chris Riddoch
Sent: Monday, April 17, 2006 10:22 AM
To: users@conserver.com
Subject: Confusion on master/slave setup

Hi, folks.

I've been reading the manpages for a bit, but I can't say I'm fully
grokking what I need to do here...

I've got one system already set up, hosting several serial ports, and
everything's working smoothly there.

How would I take another system, with another two serial ports, and use
the first machine as the master server? Should I put the conserver.cf
file on our samba server and have all conserver processes read from
that, or do they each need to be different?

I'll help improve the documentation, if someone can help me figure this
out.

--
Chris Riddoch
epistemological humility

_______________________________________________
users mailing list
users@conserver.com
https://www.conserver.com/mailman/listinfo/users

there was a previous thread about this last month starting here (sorta):

https://www.conserver.com/pipermail/users/2006-March/msg00034.html

basically, you just define more consoles like you've done already, but
use a 'master' hostname that is your second console server (the one with
the two serial ports). that value determines whether or not conserver
tries to manage the port - otherwise it just knows about it and
redirects clients to the server that is the real master (based on that
'master' value).

then distribute the conserver.cf file to both conserver hosts (either
via a shared filesystem or any distribution command). when a client
connects to either conserver host, it'll either attach directly to the
console or be redirected to the appropriate server.

hope that helps.

Bryan

On Mon, Apr 17, 2006 at 11:21:35AM -0600, Chris Riddoch wrote:
> Hi, folks.
>
> I've been reading the manpages for a bit, but I can't say I'm fully
> grokking what I need to do here...
>
> I've got one system already set up, hosting several serial ports, and
> everything's working smoothly there.
>
> How would I take another system, with another two serial ports, and use
> the first machine as the master server? Should I put the conserver.cf
> file on our samba server and have all conserver processes read from
> that, or do they each need to be different?
>
> I'll help improve the documentation, if someone can help me figure this out.
_______________________________________________
users mailing list
users@conserver.com
https://www.conserver.com/mailman/listinfo/users

Bryan Stansell wrote:
<snip>
> basically, you just define more consoles like you've done already, but
> use a 'master' hostname that is your second console server (the one with
> the two serial ports). that value determines whether or not conserver
> tries to manage the port - otherwise it just knows about it and
> redirects clients to the server that is the real master (based on that
> 'master' value).
>
> then distribute the conserver.cf file to both conserver hosts (either
> via a shared filesystem or any distribution command). when a client
> connects to either conserver host, it'll either attach directly to the
> console or be redirected to the appropriate server.

Great! That's smarter than I expected, and I now have it working
exactly the way I wanted.

By the way, all I needed to do for our non-development cygwin
environments is to compile with CFLAGS=-static, and make sure that in
addition to --with-master and --with-port, I added
--with-cffile=//fileserver/path/to/conserver.cf so that no command-line
options should be required from anyone except the terminal name.

I now have conserver.exe and console.exe in a public directory, where my
co-workers can just copy it into their local environment (or run it out
of that directory over the network, if they like) and ask me to add
their terminals to the conserver.cf when new devices are plugged in.

I also have Ruby on Rails set up to provide a web-based front end to a
system of 'expect'-like scripts that run on the master server (in fact,
they're also Ruby) so that for an increasing number of tasks, people
won't even have to bring up 'console' at all.

Thanks for the suggestions!

--
Chris Riddoch
epistemological humility

_______________________________________________
users mailing list
users@conserver.com
https://www.conserver.com/mailman/listinfo/users