[Rspamd-Users] Email hits BAYES_* after a few times

Sophie Loewenthal sophie at klunky.co.uk
Mon Jun 3 06:49:25 UTC 2019

> On 3 Jun 2019, at 8:11 am, Sophie Loewenthal <sophie at klunky.co.uk> wrote:
>> On 3 Jun 2019, at 4:49 am, Tim Harman via Users <users at lists.rspamd.com> wrote:
>>> Big type above:
>>>> Bayes should *not* be autolearning and configdump didn’t shed any light.
>> The only thing I can suggest is that other learning that you've done has learnt symbols that are being triggered later.
>> If you're not doing that at all, then maybe you've encountered a bug.
>> Seeing as you're using sqlite as the backend, I'm not sure you'll get a lot of support either as I think the preferred method now is redis.
>> You could try and enable some debugging too, but I'm not sure which debug module might help you, statistics only seems to have debug for redis (you could try an enable it to see if it helps 
> Hi Tim,
> sqlite is the default for the bayes stats. Redis is used for everything else

Can I use something this to drop my learnt Bayes,

redis-cli KEYS *BAYES* | xargs redis-cli DEL redis-cli KEYS RS* | xargs redis-cli DEL

More information about the Users mailing list