[Rspamd-Users] skipped for bayes_user classifier: already in class ham

[users-rspamd131020 at jworld.ch]

Hi everyone,

I have a strange problem (Rspamd 2.5; I know...update pending, but still):

Legit Mail comes in with BAYES_SPAM_USER(8.09)[99.98%], when I want to despamify, I get the following:

# rspamc learn_ham mygoodmail.eml
Results for file: /root/mygoodmail.eml (0.021 seconds)
error = "<id at fqdn> is skipped for bayes_user classifier: already in class ham; probability 98.48%";
filename = "/root/mygoodmail.eml";
scan_time = 0.021575;

I use redis as backend:

# vi override.d/statistic.conf

classifier "bayes" {
    name = "bayes_user"
    expire = 2144448000;
    min_tokens = 11; 
    backend = "redis";
    servers = "localhost:6378";
    database = "1";
    cache {
    }   
    tokenizer {
        name = "osb";
    }   
    learn_condition = "return require(\"lua_bayes_learn\").can_learn";
    users_enabled = true;
    per_user = true;
    prefix = "%s%l%r";
    store_tokens = false;
    min_learns = 200;
    statfile {
        symbol = "BAYES_HAM_USER";
        spam = false;
    }   
    statfile {
        symbol = "BAYES_SPAM_USER";
        spam = true;
    }
    new_schema = true;
    signatures = false;
}

I guess something in redis got mixed up. How can I search and find just the entry for this mail and remove or edit in redis? Or what else do you suggest?

Thanks a lot,
Juerg