then run Radius to connect to the cluster from a seperate box.
-----------------------
Denny Figuerres
InReach Internet
MIS Manager
On Tue, 22 Dec 1998, Brandon Bryant wrote:
> Maybe my idea of replication is not based in reality. I thought that it
> worked like mirrored drives, in that they both have the same information,
> when you change one, the other is automagically updated to match, and if
> one goes down, the other takes over.. But then, that would make too much
> sense for a Microsoft product. I think I'm going to have to re-evaluate
> the whole thing and put some more machines up..
>
> -----Original Message-----
> From: Kurt Schafer [SMTP:kahuna@cyberbeach.net]
> Sent: Monday, December 21, 1998 6:31 PM
> To: emerald@iea-software.com
> Subject: Re: [Emerald] Revelation
>
> Replication is somewhat straightforward in theory, but sometimes tricky in
> implementation. Basically you have 3 types of replication servers:
> Publishers, Subscribers, and Distributers.
>
> You can combine different types as well. (publishers can also be
> subscribers, as well as 'distribution publishers')
>
> Set up publication on your master server and have your secondary server
> subscribe to the published database.
>
> Works great for authentication but accounting is a little trickier because
> you have to merge the two databases together sometime prior to billing in
> the event of your radius clients (*grin*) switching away from the primary
> and logging calls onto your secondary.
>
> Also, in my experience, implementing concurrency control under the above
> scenario can cause some real heartache.
>
> Also, your life will be much much easier if you have a domain controller
> and
> both SQL servers are part of that domain.
>
> = K (my two canadian cents worth)
>
> -----Original Message-----
> From: Brandon Bryant <nailer@midlink.com>
> To: 'emerald@iea-software.com' <emerald@iea-software.com>
> Date: Monday, December 21, 1998 2:01 PM
> Subject: RE: [Emerald] Revelation
>
>
> >OK.. Then here is another question for you. How difficult is it to set
> up
> >replication for two SQL servers? I am reluctant to point both radius
> >servers to the same SQL server. Right now, the primary radius and the
> >primary SQL are the same machine, and the secondary radius and the
> >secondary SQL are the same machine. If the Primary radius server goes
> >down, chances are the db is down too. I guess we could have one db server
> >and point two other servers running radius to that one, and then have even
> >another machine running a backup DB, and then use the multiple DSN feature
> >on Radius to switch to the backup DB when there is a problem with the
> >first. Does (will) RadiusNT have the multi-DSN feature? I know it has
> >been discussed on the list before but I never really saw whether or not
> >this feature will do what I want it to here.
> >
> >I'm trying to keep my point of failure as far back as possible. If the
> >server goes down at 2AM, I don't want to have change anything to get
> >service back up.
> >
> >-----Original Message-----
> >From: Dale E. Reed Jr. [SMTP:daler@iea-software.com]
> >Sent: Monday, December 21, 1998 12:48 PM
> >To: emerald@iea-software.com
> >Subject: Re: [Emerald] Revelation
> >
> >Brandon Bryant wrote:
> >>
> >> I ran into an interesting situation that I was wondering if anyone else
> >has
> >> seen.
> >>
> >> We have two Ascend Maxes and two radius servers running RadiusNT. In
> the
> >> MAX setup, I set the primary auth server and the secondary auth server
> >> accordingly.
> >>
> >> Now, what is happening is when the primary goes down or is unavailable,
> >the
> >> MAX switches to the secondary. Then never switches back to the primary.
> >
> >This is the default behavior of the MAX. The RADIUS RFC does NOT
> >require
> >the NAS to always try first the primary server. In fact, the idea of a
> >primary/secondary server is a Livingston idea that most people copied.
> >Ascend, suprise, uses the one the is responded the best. Until the
> >second
> >RADIUS server has a problem, it will not switch back.
> >
> >> So, we replicate our db at night from primary to secondary. So any
> >change
> >> we make to the db during the day is null and void if the server was down
> >> before. It was driving us nuts until I figured that out over the
> >weekend.
> >> I'm going to start a ticket with Ascend today, but has anyone else run
> >> into this, and if so, how did you fix it?
> >
> >Point both of your RadiusNT servers at your primary DB. Thats the
> >easiest way to fix it. If your primary DB goes down, you can manually
> >switch them to the less recent, Secondary DB.
> >
> >--
> >Dale E. Reed Jr. (daler@iea-software.com)
> >_________________________________________________________________
> > IEA Software, Inc. | RadiusNT, Emerald, and NT FAQs
> > Internet Solutions for Today | http://www.iea-software.com
> >
> >For more information about this list, including removal,
> >please see http://www.iea-software.com/maillist.html
> >
> >
> >For more information about this list, including removal,
> >please see http://www.iea-software.com/maillist.html
> >
>
>
> For more information about this list, including removal,
> please see http://www.iea-software.com/maillist.html
>
>
> For more information about this list, including removal,
> please see http://www.iea-software.com/maillist.html
>
For more information about this list, including removal,
please see http://www.iea-software.com/maillist.html