AW: [Mimedefang] Re: "Please try again later"

[Florian Meister]

Hi,

If you don't care if some messages are not checked, there is maybe a way to implement this:

There are some options in the INPUT_MAIL_FILTER statement in the sendmail configuration. There you can specify what should be done if the milter-socket is not reachable. The problem is, that the socket is reachable all the time, even mimedefang has no free slaves. So this will not help.

So how to modify mimedefang this way, that it don't send "please try again later" if it's busy. How to implement ? I don't know, but maybe someone in the list ;)

Cu, florian

-----Ursprüngliche Nachricht-----
Von: mimedefang-bounces at lists.roaringpenguin.com [mailto:mimedefang-bounces at lists.roaringpenguin.com] Im Auftrag von John Rudd
Gesendet: Samstag, 10. September 2005 20:02
An: mimedefang at lists.roaringpenguin.com
Betreff: Re: [Mimedefang] Re: "Please try again later"


On Sep 10, 2005, at 9:57 AM, John Rudd wrote:

>
> On Sep 10, 2005, at 9:33 AM, John Rudd wrote:
>
>>
>> This morning, I'm putting mimedefang into production, just for virus 
>> and attachment checks.  I'm seeing a lot of "Please try again later" 
>> responses from mimedefang.  The problem is, this same sever is used 
>> by my local users (not just remote MTAs), and we don't want them 
>> getting that type of response.
>>
>> What do I need to adjust to make this happen less?  (sendmail's load 
>> threshold is much higher than the load I'm seeing on teh system, so I 
>> don't think it's that)
>>
>> Specifically, I wouldn't mind having it do this more for remote 
>> senders and less for local senders.  I just don't want local senders 
>> to see this much (if at all).
>>
>
> If the answer is "adjust MX_MAXIMUM in the rc script", then I found it.
>

Hm.  Well, maybe not.  After adjusting that to 20 and 15, that's not 
really helping.  I assumed that maybe that number was causing me a 
problem because I was running out of slaves to process my incoming 
messages.  Increasing it just increased my system load, but didn't make 
the problem go away (and at 20, the load was starting to close in on 
what our sendmail load thresholds are at ... so I wasn't going to go 
higher).

Setting it lower seems like it would make the problem worse, maybe.

I'm thinking it might be that we run our front line mail servers on 
tiny/hold hardware (sunblade 150's and sunfire v100's, 2 of each, and 
so far I had only installed this on one of the sunblades).  Sophos 
Puremessage recommended that we not go with anything weaker than a 
Sunfire v210 ... so maybe it's the same issue here: if we're going to 
milter, we're going to have to upgrade.

(we plan to upgrade hardware soon, but this was supposed to be a 
middle-term upgrade of software only)

So, I guess no mimedefang in production at work yet.

(which is sad, because I REALLY love using it at home, and it was set 
up just fine on our test machines ... guess I have to wait til we get 
the new hardware)



_______________________________________________
Visit http://www.mimedefang.org and http://www.roaringpenguin.com
MIMEDefang mailing list
MIMEDefang at lists.roaringpenguin.com
http://lists.roaringpenguin.com/mailman/listinfo/mimedefang