Mailing List Archive: Redundant master servers possible?

Redundant master servers possible?

Aug 9, 2007, 11:54 AM

Post #1 of 3 (3406 views)

Guys,

I'm running a minor conserver setup, the funky thing is that it's
spread out across four different sites. By that I mean I have the
master server running at one site (A) and it connects to all the serial
port concentrators at site A, and at remotes sites B, C & D.

But when I have to take down and reboot the master server, I loose all
console access, which is a bummer.

So aside from setting up multiple master servers, one for each site,
how can I setup things so that I have a master and a standby server,
so that if the master goes down, all the clients know to talk to the
standby box, which picks up the session(s) dropped by the master.

Sure, there could be problems here, but something to at least make DR
easier would be helpful. I don't care as much about the logs of
what's happened, but keeping the client side from having to make
changes is more key to me.

Thanks,
John
John Stoffel - Senior Staff Systems Administrator - System LSI Group
Toshiba America Electronic Components, Inc. - http://www.toshiba.com/taec
john.stoffel@taec.toshiba.com - 508-486-1087
_______________________________________________
users mailing list
users@conserver.com
https://www.conserver.com/mailman/listinfo/users

Re: Redundant master servers possible? [ In reply to ]

bryan at conserver

Aug 10, 2007, 10:58 AM

Post #2 of 3 (3242 views)

Permalink

On Thu, Aug 09, 2007 at 02:54:49PM -0400, John Stoffel wrote:
> So aside from setting up multiple master servers, one for each site,
> how can I setup things so that I have a master and a standby server,
> so that if the master goes down, all the clients know to talk to the
> standby box, which picks up the session(s) dropped by the master.

There isn't anything internally to support that setup. You can
certainly just have a copy of the conserver.cf file on another host
(which has access to everything) and turn up conserver there if/when
necessary. To make it even fancier, you could have a process test
connections to the main conserver host to auto-detect when it goes down
and then fire things up automatically. And I suppose you could even
make the startup sequence on the main server go to the backup and
shutdown any existing conserver processes, so that it can take back
control.

You've probably thought of all this...unfortunately I don't have any
other bright ideas.

As for the client issue, if you can swing a dns name quickly, you'd be
good. Or you create a wrapper for the console command that first tries
the master, and if it fails, tries the backup. Again, nothing internal
to help you out, but it seems doable.

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

Re: Redundant master servers possible? [ In reply to ]

cfowler at outpostsentinel

Aug 10, 2007, 11:10 AM

Post #3 of 3 (3241 views)

Permalink

Bryan,

I think you've got a good start there. If I were to do this this is
what I would do.

#1. Create some master conserver servers. These servers
will both connect to all the cosnoles in the organization.
Anyone could connect to either server.

#2. Create a replacement or wrapper for console. This
could be very easy. This program would then direct
the person to the correct server.

Example:

use Net::Ping;
use strict;

sub up {
my $ip = shift;
my $p = Net::Ping->new();
my $result = $p->ping($ip);
$p->close();
return $result;
}

# Main part
my $console_name = $ARGV[0];
my @servers = [ "192.168.1.1", "192.168.1.2" ];
foreach(@servers) {
if(up($_)) {
system "console .........";
exit 0;
}
}

die "All connections attempts to the console servers have failed!\n";

On Fri, 2007-08-10 at 10:58 -0700, Bryan Stansell wrote:
> On Thu, Aug 09, 2007 at 02:54:49PM -0400, John Stoffel wrote:
> > So aside from setting up multiple master servers, one for each site,
> > how can I setup things so that I have a master and a standby server,
> > so that if the master goes down, all the clients know to talk to the
> > standby box, which picks up the session(s) dropped by the master.
>
> There isn't anything internally to support that setup. You can
> certainly just have a copy of the conserver.cf file on another host
> (which has access to everything) and turn up conserver there if/when
> necessary. To make it even fancier, you could have a process test
> connections to the main conserver host to auto-detect when it goes down
> and then fire things up automatically. And I suppose you could even
> make the startup sequence on the main server go to the backup and
> shutdown any existing conserver processes, so that it can take back
> control.
>
> You've probably thought of all this...unfortunately I don't have any
> other bright ideas.
>
> As for the client issue, if you can swing a dns name quickly, you'd be
> good. Or you create a wrapper for the console command that first tries
> the master, and if it fails, tries the backup. Again, nothing internal
> to help you out, but it seems doable.
>
> Bryan
> _______________________________________________
> users mailing list
> users@conserver.com
> https://www.conserver.com/mailman/listinfo/users

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