Since I cannot use my email client to "unspam" the mails that have been marked by rspamd as spam, I would like to create a whitelist for rspamd. When I search for it, I only get instructions on how to set up a whitelist filter. This looks for different criteria, which are not fulfilled by the senders (yet the mails are legitimate). I tried to "ham" the messages using the "Scan/Learn" section in the browser UI, but that doesn't seem to work either. The problem is anyway, that the message always has a large attachment, and the source code is huge because of that. I don't remember if I tried, but I guess I did, to copy all of it into the window, which caused problems. So I only tried the scanning with the header of the message, but there are already entries like "Spam" etc., so I don't know if that's the reason why rspamd just doesn't want to look at those messages as ham. I also noticed that the messages are marked as spam only a few months ago (beginning in Decemper 2023), and rspamd is active much longer. What could cause rspamd to mark those messages (I get daily one from that particular sender) as spam after a long time?
Yes, but that doesn't help me, because I cannot and may not configure the sender's mail server. I could try to tell the sender why his mails get recognized as spam, but it is unlikely that this will change anything soon. There is e.g. no DKIM, among other things which are nowadays almost considered as standard. I just need to tell rspamd that it should let these mails pass through as ham, that's the part where I struggle, because it seems that rspamd doesn't make that easy. Or I've missed something, which is possible...
A customer pasted me the entire message that was marked as spam, but I still fail to convince rspamd that the sender is NOT a spammer. I have followed many guides and set up local rules.
Create SSH tunnel from localhost port <SOMETHING> to port rspamd uses on the RSPAMD host . Something like Code: ssh -L 127.0.0.1:$LOCALPORT:192.168.24.42:80 [email protected] Read the ssh man page for option -L. There should be instructions available on the Internet.
Can anybody, in plain language, explain why messages form sendrer's addresses (or domains) which have been whitelisted on the ISPConfig panel stlill get marked or rejected by rspamd ? Wen you look at the rspamd reports, it looks like in most of the cases, the result of the BAYES score is too high. But why in hell should rspamd apply a BAYES filter on mails from users that have been whitelisted in the first place ? Is there a way to disable the BAYES score (or force it's value to zero) when the sender's is whitelisted ????
Maybe there is syntax error in the whitelisting setting? That looks like the whitelisting is not applied. You have not provided enough information to know why that would be the case. Examine rspamd logs and postscript logs to see what happens to those e-mails.
I have the same problem. After investigation, it seems that the whitelist entries created using ISPConfig find their way to /etc/rspamd/local.d/users in a file named spamfilter_wblist_#.conf In there, the (anonymized) code looks like: Code: spamfilter_wblist-8 { priority = 50; from = "@allowed.com"; rcpt = "@mydomain.com"; want_spam = yes; apply { actions { reject = null; "add header" = null; greylist = null; "rewrite subject" = null; } } } There is no code in there that disables the Bayes filter, which therefore remains active and causes spam flagging of mails from whitelisted senders. One woud expect to see something like: Code: symbols_disabled = ["BAYES_HAM", "BAYES_SPAM"]; but it is not there. This seems like a bug. As a workaround, it would be possible to add the line above in the "apply" section of those files but, of course, that will be overwritten by ISPConfig at the next modification of the whitelist. Is this diagnosis correct, and is there a clean solution ?
