[Rspamd-Users] Running the Docker version with Podman on my workstation - how to test before installing on my mail server?

[Philip Rhoades]

Gerald,

Thanks for responding - see inline comments:


On 2025-05-23 14:59, Gerald Galster wrote:
>> I have been varieties of using Qmail for decades and I used Spamdyke 
>> for a long time and then just MailDrop for simple spam filtering - 
>> which works pretty well for just me.  However, I am interested in 
>> RSpamd mostly as a practical and useful exercise to learn about 
>> Bayesian statisitics and for servers with more users.
>> 
>> So, I have got the RSpamd docker image running with Podman and have 
>> the web console working on my workstation - is there some way of 
>> testing this out in isolation eg with std input and output before I 
>> try and install it on my server with the MTA?
> 
> Most MUAs support exporting mails as plain text (.eml file).
> You can scan those via rspamc or curl on the cmdline.
> Search for rspamc here: https://rspamd.com/doc/faq.html


Ah excellent! - just what I want . .


> Beware that Bayes needs a certain amount of samples (training)
> before it will add to the spam score. See min_learns here:
> https://docs.rspamd.com/configuration/statistic/#statistics-configuration


Sure - I have these spams to test:

2015    771
2016    2410
2017    7178
2018    3181
2019    2711
2020    6527
2021    9341
2022    9740
2023    5530
2024    16674
2025    10610 - so far . .

and of course lots of ordinary mail.


> When running with the MTA you can temporarily set the reject
> action score to a higher value to avoid blocking:
> https://github.com/rspamd/rspamd/blob/master/conf/actions.conf


OK - I will check all that out.

Thanks!

P.
-- 
Philip Rhoades

PO Box 896
Cowra  NSW  2794
Australia
E-mail:  phil at pricom.com.au