|
Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com |
From: Keean Schupke (keean
fry-it.com)
Date: Mon Dec 03 2007 - 05:53:48 CST
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
On 03/12/2007, Wietse Venema <wietse
porcupine.org> wrote:
> Victor Duchovni:
> > On Sun, Dec 02, 2007 at 10:05:09PM -0500, Wietse Venema wrote:
> >
> > > > So the SASL code needs to use "proxymap" protocol directly, not
> > > > "proxy:..." tables, which always contacts the "real" proxymap.
> > >
> > > To access a single-writer auth_fail_map, use
> > > proxy:btree:/some/where/auth_fail_map a dedicated proxymap service,
> >
> > While the delivery agent concurrency to a single destination is ideally
> > just "1" in this case, the process limit for the delivery agent may
> > well be significantly higher. Is it acceptable for all the smtp(8)
>
> Use different master.cf entries for proxy read and write servers.
>
> Use the existing proxymap service name for the read-only multiserver.
>
> Use the new proxywrite service name for the single-updater service.
>
> This can be completely transparent to the application:
>
> dict_open proxy:foo O_RDONLY -> old proxymap (multiserver)
> dict_open proxy:bar O_RDWR|O_CREAT -> new proxywrite (single updater)
>
> I'll add a few lines of code for that today.
>
> Wietse
>
As the code will only work with proxywriter, should I prepend "proxy:"
to the map name in the code, validate the string (make sure it starts
with proxy: but not actually alter the string), or should I allow
invalid settings in the conf file to be accepted?
Regards,
Keean Schupke, Fry-IT Ltd.
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]