[Rspamd-Users] several skips in the log

[G.W. Haywood]

Hi there,

On Sat, 8 Apr 2023, tom at myposts.ovh wrote:

> My mailserver is in new setup.
> I saw in the log there were several skips info:
>
> 2023-04-08 18:05:08 #244121(rspamd_proxy) <8455fb>; proxy; rspamd_redis_connected: skip obtaining bayes tokens for BAYES_SPAM of classifier bayes: not enough learns 0; 200 required
> 2023-04-08 18:05:08 #244121(rspamd_proxy) <8455fb>; proxy; rspamd_redis_connected: skip obtaining bayes tokens for BAYES_HAM of classifier bayes: not enough learns 0; 200 required
> 2023-04-08 18:05:08 #244121(rspamd_proxy) <8455fb>; proxy; rspamd_stat_classifiers_process: skip statistics as SPAM class is missing
> 2023-04-08 18:05:08 #244121(rspamd_proxy) <8455fb>; lua; greylist.lua:331: Score too low - skip greylisting
> 2023-04-08 18:05:08 #244121(rspamd_proxy) <8455fb>; lua; neural.lua:442: skip ham sample to keep spam/ham balance; probability 1; 0 spam and 1 ham vectors stored
> 
> Can you tell why they are skipped?

The messages themselves seem to tell you why obtaining Bayes tokens
was skipped: 200 learns are required, and 0 exist.  I don't know why
you have no learns.  That *might* be what you should investigate, but
it *might* just be that your system has not yet processed enough mail
to build a representative body of ham and spam samples.  If that isn't
the case, perhaps you need to change something in the configuration to
get the Bayesian ham/spam classification working.  There is a lot of
documentation about this sort of thing, both on the rspamd Website and
elsewhere on the Web.  See for example

https://rspamd.com/doc/configuration/statistic.html
https://pipo.blog/articles/20210311-rspamd-trainer-bayes

which were the first two hits found from a Google search today for

"rspamd bayes learn"

I cannot vouch for the accuracy of any information found in searches. :/

-- 

73,
Ged.