I had this problem at one point. SA has a size limit to emails it will check. It is set kind of low. 512k? I’m sure this is to save resources. I know I had to bump up the size to 1M at one point because spammers have learned to pad their spam to get it past the 512k limit to bypass filtering.
did you check the mail log at the point that the email was delivered ? This will give you more insight into what’s going on rather than looking at email headers
i tried, no changes
i continue to watch the problem, and i determinate when i use spamassasin this problematic virtual server wont check spam, but in the rest of the virtual server work perfect. if i swap to spamc, it works!, does anybody know if the size of the mailbox has relation with the process of spamassasin? i ask because the mailbox is huge!
That is HUGE! @ID10T has already pointed out above the inadequacies of SA.
Perhaps that is also why spamc appears to work.
I still think it (or part of it - 512k?) is being checked and not triggering as spam. → the score says so. It is not in learning mode so repeated hits are not increasing the score.
If the configuration is identical for all servers then I cannot see how one server shows as an exception to the others unless the spam being received is different.
yes , all of them use default config
this is the reason for suspect about mailbox size, because i dont find any diference. more i create other email adress, ( new and empty, but , part of the same virtual server) and it has the same behavior. any idea? i tried to “sell” a process of backup and clean process of the mailbox to my customer, but i afraid that this solution dont solve the problem and finally i became in a donkey ( for my customer)
dificult to know. really my customer is a Travel Agency, and this is the reason for a high volumen of received mails, they receive a lot of comercial campaign of mail, spam, and more. but my customer says that before to move to my server ( before in ionos) he receive more spam , and actually there are only a couple per week
Sometimes spam can be difficult to define. One person’s idea of spam can be totally different from another’s. I get 100’s per week. mostly trying to sell things I don’t want or videos, podcasts and the like I have no interest in. Surely your client can cope with “a couple” per week while they can be analysed to identify a trigger word/phrase/identity to train SA.
You said it works with spamc, so why not keep it as that?
