Verify that Junk-Folder will be scanned

tcz

Hi,

how can i verify, that E-Mails who moved to the Junk-Folder will processed by rspamd?
I have searched a lot of logs and also the rspamd-weblog, but i cant find any entry for this.

If i see right, based in the dovecot.conf mails will only processed, if they will be “copy” to the Junk folder. This event will be logged to the mailcow.log but only, if i used my iPhone. With Outlook there are only a “delete” event in the Old-Folder, but no “copy” event to the Junk-Folder.

Is there a safe way to check/verify the process of Junk-Folder?

thanks in advance,
tcz

esackbauer

https://community.mailcow.email/d/3482-debug-rspamd-not-training-with-inbox-junk-transitions/6

tcz

thanks! now i have the debug logging and found out a strange problem …
If i move a mail from/to junk, e.g. with sogo, in the log i found this:

... copy from INBOX: box=Junk ...
... expunge: box=INBOX ...

... copy from Junk: box=INBOX ...
... expunge: box=Junk...

This looks correct and trigger the learn spam/ham event.
If i do the same in outlook, i see the following in the log:

... delete: box=INBOX ...
... expunge: box=INBOX ... 
... save: box=Junk ...

... save: box=INBOX ...
... delete: box=Junk ...
... expunge: box=Junk ...

In this case, there will be not spam/ham learning triggered!
I think this is, because only the “COPY”-Event will be tragged:

imapsieve_mailbox1_name = Junk
imapsieve_mailbox1_causes = COPY
imapsieve_mailbox1_before = file:/usr/lib/dovecot/sieve/report-spam.sieve

If this is correct, then the Junk-Folder will not work in Outlook - can this anyone confirm? :-(
Or better - is there any solution for this “problem”?

additional, if i comment out the “causes” lines, it works also with outlook:

imapsieve_mailbox1_name = Junk
#imapsieve_mailbox1_causes = COPY
imapsieve_mailbox1_before = file:/usr/lib/dovecot/sieve/report-spam.sieve

But i am not sure, if this have any sideeffects/other problems?

additional, if i comment out the “causes” lines, it works also with outlook:

imapsieve_mailbox1_name = Junk
#imapsieve_mailbox1_causes = COPY
imapsieve_mailbox1_before = file:/usr/lib/dovecot/sieve/report-spam.sieve

But i am not sure, if this have any sideeffects/other problems?

DocFraggle

How exactly did you connect your account with Outlook? I use ActiveSync, there it works as intended.
I just enabled debug logging in rspamd like described in https://community.mailcow.email/d/3482-debug-rspamd-not-training-with-inbox-junk-transitions/6 and moved a SPAM mail from Junk to my Inbox. rspamd logged it as

rspamd-mailcow-1  | 2024-09-27 17:29:38 #38(controller) <86aa9c>; bayes; bayes_learn_spam: token 9851794553962298747 <#h:C-T:multipart/alternative; boundary="cIPN7pci1u9SLM03WPoPD=_jAYrIggUswE":#m:#mpth>: window: 1, total_count: 1, spam_count: 0, ham_count: 1
rspamd-mailcow-1  | 2024-09-27 17:29:38 #38(controller) <86aa9c>; bayes; bayes_learn_spam: token 6463352648909517147 <#F:Versand:#h:C-T:multipart/alternative; boundary="cIPN7pci1u9SLM03WPoPD=_jAYrIggUswE">: window: 1, total_count: 1, spam_count: 0, ham_count: 1
rspamd-mailcow-1  | 2024-09-27 17:29:38 #38(controller) <86aa9c>; bayes; bayes_learn_spam: token 10668326775548025847 <#u:neuheiten.nl:#F:Versand>: window: 1, total_count: 1, spam_count: 0, ham_count: 1
rspamd-mailcow-1  | 2024-09-27 17:29:38 #38(controller) <86aa9c>; bayes; bayes_learn_spam: token 14426626996836589832 <#dt:4:07:#u:neuheiten.nl>: window: 1, total_count: 1, spam_count: 0, ham_count: 1
rspamd-mailcow-1  | 2024-09-27 17:29:38 #38(controller) <86aa9c>; bayes; bayes_learn_spam: token 13963418669235991226 <#aur:s=pass,d=pass:bionovanutriments.com:#dt:4:07>: window: 1, total_count: 1, spam_count: 0, ham_count: 1

spam_count: 0, ham_count: 1

then I moved it back to Junk:

rspamd-mailcow-1  | 2024-09-27 17:30:35 #38(controller) <56e171>; bayes; bayes_learn_spam: token 9851794553962298747 <#h:C-T:multipart/alternative; boundary="cIPN7pci1u9SLM03WPoPD=_jAYrIggUswE":#m:#mpth>: window: 1, total_count: 0, spam_count: 1, ham_count: -1
rspamd-mailcow-1  | 2024-09-27 17:30:35 #38(controller) <56e171>; bayes; bayes_learn_spam: token 6463352648909517147 <#F:Versand:#h:C-T:multipart/alternative; boundary="cIPN7pci1u9SLM03WPoPD=_jAYrIggUswE">: window: 1, total_count: 0, spam_count: 1, ham_count: -1
rspamd-mailcow-1  | 2024-09-27 17:30:35 #38(controller) <56e171>; bayes; bayes_learn_spam: token 10668326775548025847 <#u:neuheiten.nl:#F:Versand>: window: 1, total_count: 0, spam_count: 1, ham_count: -1
rspamd-mailcow-1  | 2024-09-27 17:30:35 #38(controller) <56e171>; bayes; bayes_learn_spam: token 14426626996836589832 <#dt:4:07:#u:neuheiten.nl>: window: 1, total_count: 0, spam_count: 1, ham_count: -1
rspamd-mailcow-1  | 2024-09-27 17:30:35 #38(controller) <56e171>; bayes; bayes_learn_spam: token 13963418669235991226 <#aur:s=pass,d=pass:bionovanutriments.com:#dt:4:07>: window: 1, total_count: 0, spam_count: 1, ham_count: -1

spam_count: 1, ham_count: -1

tcz

The Account is added as IMAP-Account.

DocFraggle

tcz Hmm, seems you are right, doesn’t work for me with an IMAP account as well…

tcz

i now find out, that in the log spam/ham was learned if i commend out “COPY”, but the counters at the rspamd-status-page not increased. If i move a mail via sogo, the counters are increased … :-(

any idea how i can fix that?