Mail server backlogging - seems like a Bayes problem

[Julian Field]

On 25/10/04 4:24 pm, "Matt Kettler" <mkettler at EVI-INC.COM> wrote:

> At 11:37 PM 10/22/2004, Julian Field wrote:
>> Matt,
>>
>> Thanks for that. I originally kept the timeout short as genuine SA timeout
>> problems did cause difficulty when I first implemented. But SA has improved
>> enormously since I first supported it, and I suspect you are right that it
>> is now time to increase the timeout quite a bit. The last increase I did
>> was from 30 to 40 seconds so SA had a chance to timeout DNS lookups before
>> it could be totally killed by MailScanner.
>>
>> Thanks for your opinion, as ever :-)
>
> You're welcome Julian.

New config file says this:

# If SpamAssassin takes longer than this (in seconds), the check is
# abandoned and the timeout noted.
SpamAssassin Timeout = 75

# If SpamAssassin times out more times in a row than this, then it will be
# marked as "unavailable" until MailScanner next re-starts itself.
# This means that remote network failures causing SpamAssassin trouble will
# not mean your mail stops flowing.
Max SpamAssassin Timeouts = 10

I reckoned 75 was a perfect compromise between your suggestions of 60 and 90
:-)

> And thanks for making a great product, even if it does require the
> occasional tweak as SA and virus scanners evolve :)

Evolution is one of the major competitive advantages of the Open Source
model. Lots of eyes are watching out for effects and can help propose
solutions.

Which reminds me, I must get the book up to date with the phishing net code.
I'll have to find some time to do that.

Also, while I'm here: any volunteers to help get the rest of the
documentation on the website up to date please? Some of it is decidedly out
of date now, and I'm sure plenty of you are in a good position to be able to
help update it.

> (Next product improvement request: psychic adaptive spamassassin interface
> module, to keep abreast of unexpected changes in future versions of SA.)

Fortunately they didn't change the API too much, and they changed it in such
a way that it was easy to support both the old and the new in the same code.
If only all the virus scanning engines were that easy...


--
Julian Field
www.MailScanner.info
Buy the MailScanner book at www.MailScanner.info/store

PGP footprint: EE81 D763 3DB0 0BFD E1DC 7222 11F6 5947 1415 B654

------------------------ MailScanner list ------------------------
To unsubscribe, email jiscmail at jiscmail.ac.uk with the words:
'leave mailscanner' in the body of the email.
Before posting, read the MAQ (http://www.mailscanner.biz/maq/) and
the archives (http://www.jiscmail.ac.uk/lists/mailscanner.html).