> I'm setting this up for multiple servers. I would like to just run it on
> one server and have other postfix instances query it there. But it
> seemed to me to make the most sense to install policyd on each one and
> share a mysql database. That way if the main sever goes down mail can
> keep passing on the others. Does this work ok for consistency?
>
> Does this make the most sense or is there a way to tell postfix to skip
> policyd if it can't connect.
We run our policyd database on a dual-master mysql setup (with IP failover),
and each of our postfix servers runs its own local copy of policyd
connecting to the fail-over-IP. If the active mysql server dies, the IP
moves to the second mysql server, with a brief interupttion to policyd (this
has actually happened in production - and worked flawlessly).
The cleanup process runs from only one of the postfix/SMTP servers.
----------- -----------
| mysql A |====| mysql B |
----------- -----------
|
|----------------------------------
| | |
----------- ----------- -----------
| SMTP A |====| SMTP B |====| SMTP C |
----------- ----------- -----------
Note that this is with policyd 1.8.
Im not sure on policyd v2, but 1.8 allows you to specify failsafe mode which
would answer your database concerns:
# if the database or queries fail, continue accepting mail
# 1=on 0=off
FAILSAFE=1
You can also specify a policy service timeout within postfix, to limp around
policyd ever crashing (god forbid!):
# Set a sensible timeout for our policy daemons
smtpd_policy_service_timeout = 15s
Regards
__________________________________________
Dean Manners
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
policyd-users mailing list
policyd-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/policyd-users
