Geert Hendrickx wrote:
> On Tue, Jan 27, 2009 at 12:47:43PM -0800, J Sloan wrote:
>
>> I go a step farther: Instead of having each postfix server running an
>> instance of policyd, I use a single policyd server with the database on
>> the same machine, and all postfix boxes all talk to that same policyd
>> server.
>>
>
>
> The thing is policyd's behaviour when the database is unavailable is more
> configurable (either tempfail or bypass) than Postfix' behaviour when
> policyd is unavailable (always tempfail). That's why we run policyd
> locally on each Postfix server, with a central SQL database server.
>
Tha's a very good point. It's always a tradeoff. OTOH when we had
multiple policyd instances talking to mysql, we were seeing frequent
corruption of the policyd myisam tables whenever there was a network
glitch. Since moving to a single policyd on the same host as the mysql
database, we haven't had any table corruption.
Joe
_______________________________________________
Users mailing list
[email protected]
http://lists.policyd.org/mailman/listinfo/users
