SA 3.1.7 returning no result to MS?
Glenn Steen
glenn.steen at gmail.com
Fri Nov 10 16:27:56 GMT 2006
On 10/11/06, David Lee <t.d.lee at durham.ac.uk> wrote:
> (Linux/FC5; sendmail 8.13.7; MS 4.56.8)
>
> Yesterday on our two main inbound mailrelays I upgraded SA from 3.1.4 to
> 3.1.7. The MS config has:
> Log Spam = yes
> Log Non Spam = yes
>
> In the daily logs we now seem to be getting several occurence of:
> Message XXX from YYY to ZZZ is not spam, SpamAssassin (not cached, score=0, required 6, autolearn=)
>
> and:
> Message PPP from QQQ to RRR is not spam, SpamAssassin (cached, score=0, required 6, autolearn=)
>
> scattered amongst the occurences of more real data. (Around 7% of entries
> on one machine, around 4% on the other, are in such truncated/empty forms).
>
> The daily logs prior to this show no occurences at all.
>
> Any thoughts?
>
> Further data:
>
> 1. At the same time, I also got Razor2 working (from within SA) on these
> two machines.
>
> 2. When I check on a third (higher MX, lower preference) machine on which
> I did a similar upgrade, but on which Razor had been working properly
> working both before and after the upgrade, this has such entries both
> before and after. Which sort of points the finger towards Razor, rather
> than the SA upgrade.
>
>
>
> Anyone seen anything like this before? Is the apparently empty result
> from SA something that MS might be able to detect? How to debug something
> like this?
>
> (My next step might be to disable Razor and see if that seemed to stop
> these occurences. But that would simply provide an extra data point, not
> really provide a useful route to debug, understand and fix this overall
> MS/SA/Razor issue.)
>
Could be razor, I suppose. But couldn't this be the same sa-update
"madness" some have seen, where MS/SA simply fail to load the
moved/merged/updated rules from /var/lib/spamassassin/<version>....?
--
-- Glenn
email: glenn < dot > steen < at > gmail < dot > com
work: glenn < dot > steen < at > ap1 < dot > se
More information about the MailScanner
mailing list