RE: [RadiusNT] Radius Redundancy Solution

Troy Settle ( (no email) )
Sun, 13 Feb 2000 07:13:14 -0500

This is a multi-part message in MIME format.

------=_NextPart_000_001F_01BF75F1.CAE77DC0
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: 7bit

Radius Redundancy SolutionSean,

We're implementing the same thing. So far, I've successfully tested
transactional replication (everything but the calls table), and it's working
well. (we have well over 1000 ports here).

As for accounting, here's what I got from IEA:
> In order to get RadiusNT to do what you want, you'd have to define the
> accounting datasource separately in the admin and point it just at the
> primary DB. That way Radius will queue them up instead of trying to send
> the accounting to secondary read only replicas.

Currently, we're trying to figure out how to do a local calls table, and
push that back to the primary DB once a day. The issue that comes up, is if
we delete 100k records from the secondary, will that also cause the
replication mechinism to delete those records from the primary on the next
replication process?

-Troy

-----Original Message-----
From: radiusnt-request@iea-software.com
[mailto:radiusnt-request@iea-software.com]On Behalf Of Sean Herr
Sent: Friday, February 11, 2000 00:27
To: 'radiusnt@iea-software.com'
Subject: [RadiusNT] Radius Redundancy Solution

Has Iea-Software published any documentation / white paper on a good
solution for redundancy? We are talking about NT here!!

I have two SQL 7.0 boxes that I want to accomplish this with. With
efficiency / performance being a factor, I am curious how I can do the
following:

Primary Auth and Accounting, if it goes down, let backup pickup where it
left off and then when the primary is backup, update the primary, my only
concern is that my calls table is very active and I may lose transactions.

Doing transaction publishing every second may not be the best way to
accomplish this.

Any suggestions??

Thanks,

Sean Herr

------=_NextPart_000_001F_01BF75F1.CAE77DC0
Content-Type: text/html;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
Radius Redundancy Solution

Sean,
 
We're=20implementing the same thing.  So far, I've successfully tested=20transactional replication (everything but the calls table), and it's =working=20well. (we have well over 1000 ports here).
 
As for=20accounting, here's what I got from IEA:

> In order to get RadiusNT to do what =you want,=20you'd have to define the
> accounting datasource separately in =the admin=20and point it just at the
>=20primary=20DB. That way Radius will queue them up instead of trying to =send
> =the=20accounting to secondary read only =replicas.

Currently, we're trying to figure out how to =do a local=20calls table, and push that back to the primary DB once a day.  The =issue=20that comes up, is if we delete 100k records from the secondary, will =that also=20cause the replication mechinism to delete those records from the primary =on the=20next replication process?
 
-Troy
 
-----Original Message-----
From:=20 radiusnt-request@iea-software.com=20 [mailto:radiusnt-request@iea-software.com]On Behalf Of Sean=20 Herr
Sent: Friday, February 11, 2000 00:27
To:=20 'radiusnt@iea-software.com'
Subject: [RadiusNT] Radius =Redundancy=20 Solution

Has Iea-Software published any =documentation /=20 white paper on a good solution for redundancy? We are talking about NT = here!!

I have two SQL 7.0 boxes that I want to =accomplish=20 this with.  With efficiency / performance being a factor, I am =curious=20 how I can do the following:

Primary Auth and Accounting, if it goes =down, let=20 backup pickup where it left off and then when the primary is backup, =update=20 the primary, my only concern is that my calls table is very active and =I may=20 lose transactions.

Doing transaction publishing every =second may not=20 be the best way to accomplish this.

Any suggestions??

Thanks,

Sean Herr =

------=_NextPart_000_001F_01BF75F1.CAE77DC0--