Mail Server: Dovecot & Anti-Spam

In OpenDKIM, what is the division of labor between signing.table and key.table?

Postfix is running and can already send and receive mail on the server itself, so why do you still need Dovecot for external clients like Thunderbird or a phone?

A client can reach IMAP on port 143 with no encryption at all, yet port 993 refuses any unencrypted connection — what accounts for that difference?

Dovecot's config files run to 1,500-1,800 lines yet almost none of them do anything — how do you see just the settings that actually deviate from the defaults?

Dovecot auto-detects mailboxes and it worked in testing, so why is it still recommended to set mail_location explicitly in 10-mail.conf?

A macOS Mail user keeps hitting 'Maximum number of connections from user and IP exceeded' against your Dovecot server — what limit is being tripped and how do you raise it?

When you set up a mail account against a plain Dovecot server, where do the username and password actually come from?

Postfix supports SASL yet you still have to bring Dovecot into the picture to let remote users send mail — what does Dovecot actually contribute?

In the Postfix main.cf SASL block, smtpd_sasl_path is set to 'private/auth' rather than the full /var/spool/postfix/private/auth — why the relative path?

A mail client's setup screen warns it will send a 'Normal Password' / plain-text password — with STARTTLS enabled, is that actually a security problem?

How does SpamAssassin actually reach a spam verdict, and what does it stamp on the message so clients can filter it?

Legitimate mail is landing in the spam folder because SpamAssassin scores it too aggressively — which local.cf setting do you adjust, and in which direction?

How is SpamAssassin wired into Postfix's delivery path so that incoming mail actually gets scanned?

You want to prove your spam filter is live without waiting for real spam to arrive — what standard test message forces a spam verdict?

SpamAssassin already tags spam with headers, so why does moving that spam into a server-side Junk folder need two extra pieces of software?

To send incoming mail through Dovecot's LMTP so Sieve can file spam, which Postfix main.cf setting hands delivery over?

What role does ClamAV play on a mail server, and what does freshclam do for it?

You deploy ClamAV on a small 1 GiB cloud VM and the daemon keeps dying on its own — what is happening and what is the practical floor for RAM?

How do you confirm ClamAV is scanning mail without exposing your server to a real virus?

SPF, DKIM, and DMARC all live in DNS TXT records — how do their jobs actually differ?

An SPF record ends with a qualifier for 'everything else' — why is the softer ~all usually recommended over the strict -all?

What exactly does DKIM let a receiving mail server prove that SPF cannot?

OpenDKIM runs as a milter in 'Mode sv' — what do those two letters switch on?

What does opendkim-genkey produce, and which options control the key size and its name?

opendkim-testkey reports 'key not secure' on your new DKIM key — is that a problem you need to fix?

Wiring OpenDKIM into Postfix, why is milter_default_action set to accept, and where do you point the milter parameters?

Postfix has both smtpd_milters and non_smtpd_milters — which mail does each one filter?

You're enabling DMARC for the first time — which policy value should you start with, and why not jump straight to reject?

Beyond setting a policy, what practical intelligence do you get by adding a rua address to your DMARC record?

Why bother implementing SPF and DKIM at all if spammers can adopt them too?