Ascend Archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: (ASCEND) radiusdbm file and reading and writing



Hello,

I have made a GDBM patch for radius. Since then all problems with locking
went away. We are using our linux based radius server (based on
radius-970814 with some patches including GDBM) with about 10 NAS (Ascend
MAX and Linux PPP) without any problems.
I would make this patches available, but only if they will be include with
the Ascend standard radius distribution. Otherwise I would have to keep the
patches in sync with new releases. I have no time enough to do this...

Michael Seyfried
  
At 18:02 19.01.98 +0100, you wrote:
>On Sat, Jan 17, 1998 at 07:14:55PM -0500, reed wade wrote:
>> 
>> 
>> that's actually on my list of stuff to do "real soon now"
>> 
>> the key/content pairs are pretty straightforward:
>> 
>> username\0 -> password line\nRest of entry (no newlines)\0
>> 
>> so, it should be easy to write a perl script to poke around 
>> in the database
>
>Hmm, but would result in a DBM-only solution. DBM is weird enough to
>make problems on machines with real locking and the DB patches from
>this list make radiusd use a better lookup engine than DBM (GDBM may
>have worked as well but there never was a patch made available).
>
>IMHO the solution of choice is a set of m4 templates and local hacks
>and a source file with a simple syntax aka
>
>HACK(radius-outgoing-ip-fake-for-callback.m4)dnl
>BANNER(`banner.m4')dnl
>ROUTERCLID(`bla-gate', `123445566', `123.45.67.89', `255.255.255.240')dnl
>ROUTERCLID(`foo-gate', `33443322', `123.33.44.88', `255.255.255.0')dnl
>POOLUSER(`us1234', `secret', `1')dnl
>POOLUSER(`us2345', `nada', `1')dnl
>DEFAULT(`defunix.m4')dnl
>
>which is then exploded into real radius source by m4 with help of a number
>of template definitions. The definition of ROUTERCLID would create two
>profiles, the dial-in and the dial-out profile, it would enter the
>appropriate route into the route-n profile and split this one after
>maybe 10 routes. After the user file is built it is db[m]ified by 
>builddb[m] and ready for use. To change a parameter globally you just
>change it in the template. Finally make(1) can be used to keep everything
>up to date smoothly. Doesn't sound too complicated, does it ? And a concept
>like this has made sendmail easy to set up.
>
>-- 
>
>Kanther-Line: PGP SSH IDEA MD5 GOST RIPE-MD160 3DES RSA FEAL32 RC4
>
>+-o-+--------------------------------------------------------+-o-+
>| o |               \\\- Brain Inside -///                   | o |
>| o |                   ^^^^^^^^^^^^^^                       | o |
>| o | Andre' Beck (ABPSoft) beck@ibh-dd.de XLink PoP Dresden | o |
>+-o-+--------------------------------------------------------+-o-+
>++ Ascend Users Mailing List ++
>To unsubscribe:	send unsubscribe to ascend-users-request@bungi.com
>To get FAQ'd:	<http://www.nealis.net/ascend/faq>
>
>
--------------------------------------------------------------------
BLITZ Internet Service GmbH, Nebingerhof 5, D-96047 Bamberg, Germany
Tel: +49/951/9685163,   Fax: +49/951/9685164,   http://www.blitz.net
++ Ascend Users Mailing List ++
To unsubscribe:	send unsubscribe to ascend-users-request@bungi.com
To get FAQ'd:	<http://www.nealis.net/ascend/faq>


Follow-Ups: References: