Lacre fails to send email #174
Labels
No labels
ANSIBLE
BUG
CODE
DEVELOPMENT
DOCUMENTATION
FEEDBACK
FIX
HOWTOs
IDEA
INFRA
ISSUE
MAILSERVER
RESEARCH
TESTS
To-Be-Reviewed
WEB
WEBSITE
No milestone
No project
No assignees
2 participants
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference: Disroot/gpg-lacre#174
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
I noticed an issue a while back with Lacre. I have not report it before as it was affecting emails that would be anyway rejected, however there might be bigger issue at hand so probably good to check what exactly causes the issue.
Currently on our server, before email is handed to Lacre, it's checked by Amavis (spam/virus check). I have few spamassasin filters looking for some obvious spam phrases and giving it high score (eg
5000
). When checking headers, postfix is ordered to reject any email with spam score higher then499.*
. Seems like this triggers Lacre to fail:I've tested this against other header filters and seems to be working ok. Just the X-SPAM seems to have issue.
I think what I could do to help narrowing down the issue or at least make it more possible to test is setting up similar scenario on lacre.io. This means spamassasin rule giving 5000 triggered on some keywords (for reproducibility), and add header check to postfix. Unless of course @pfm you see the issue right away based on the traceback posted above.
Quick update. I suspected high score number to trigger issue. But when removing the header check on postfix side, emails arrive to the other side marked as spam, so the issue seems to be with header_checks here which causes postfix to trip out. I will re-create scenario on lacre.io
I'll investigate it tomorrow morning, thanks for including the traceback.
I'll be at the coffee corner until noon tomorrow :)
I suspect it's related to overall message flow configuration, but I'll investigate further on
lacre.io
later.