I've been testing PMG for several weeks across multiple domains, but just discovered a deal-breaker problem if it can't be solved.
I set up a domain with a valid MX record pointing to the PMG, but with a transport that points to a server that has port 25 shut off -- to simulate a temporary outage at the destination server. I have "Verify Receivers" set to Yes.
Here's what I get when I send it mail (from outlook.com):
It neither buffers the email for later checking, nor does it send a bounce to the sender; it just fails.
This is in contrast to our existing Barracuda server (which I would dearly love to replace) that buffers everything for up to 48hrs, rechecking every 15m or so. Once it can contact the server and verify the address, THEN it sends a bounce or whatever as necessary.
Is this behavior possible with PMG? If not, I guess I'm done experimenting, as we have several customers with mail servers that often have poor connectivity.
I set up a domain with a valid MX record pointing to the PMG, but with a transport that points to a server that has port 25 shut off -- to simulate a temporary outage at the destination server. I have "Verify Receivers" set to Yes.
Here's what I get when I send it mail (from outlook.com):
Code:
2023-10-19T15:47:32.559490-07:00 pmg postfix/smtpd[642633]: connect from mail-co1nam11olkn2011.outbound.protection.outlook.com[40.92.18.11]
2023-10-19T15:47:32.998972-07:00 pmg postfix/cleanup[642635]: F3BBAC00B1: message-id=<20231019224732.F3BBAC00B1@pmg.real.dns>
2023-10-19T15:47:32.999899-07:00 pmg postfix/qmgr[642458]: F3BBAC00B1: from=<double-bounce@pmg.real.dns>, size=219, nrcpt=1 (queue active)
2023-10-19T15:47:39.008330-07:00 pmg postfix/smtpd[642633]: NOQUEUE: reject: RCPT from mail-co1nam11olkn2011.outbound.protection.outlook.com[40.92.18.11]: 450 4.1.1 <dave@domain-with-down-server.com>: Recipient address rejected: unverified address: Address verification in progress; from=<not-real-name@outlook.com> to=<dave@domain-with-down-server.com> proto=ESMTP helo=<NAM11-CO1-obe.outbound.protection.outlook.com>
2023-10-19T15:47:39.097904-07:00 pmg postfix/smtpd[642633]: disconnect from mail-co1nam11olkn2011.outbound.protection.outlook.com[40.92.18.11] ehlo=2 starttls=1 mail=1 rcpt=0/1 quit=1 commands=5/6
2023-10-19T15:47:46.665273-07:00 pmg postfix/postscreen[642584]: CONNECT from [147.78.103.182]:56887 to [66.42.80.17]:25
2023-10-19T15:47:46.774093-07:00 pmg postfix/postscreen[642584]: PREGREET 11 after 0.11 from [147.78.103.182]:56887: EHLO User\r\n
2023-10-19T15:47:46.990836-07:00 pmg postfix/postscreen[642584]: DISCONNECT [147.78.103.182]:56887
2023-10-19T15:48:03.056681-07:00 pmg postfix/smtp[642648]: connect to 216.32.74.103[216.32.74.103]:25: Connection timed out
2023-10-19T15:48:03.057448-07:00 pmg postfix/smtp[642648]: F3BBAC00B1: to=<dave@domain-with-down-server.com>, relay=none, delay=30, delays=0/0.03/30/0, dsn=4.4.1, status=undeliverable (connect to 216.32.74.103[216.32.74.103]:25: Connection timed out)
2023-10-19T15:48:03.057820-07:00 pmg postfix/qmgr[642458]: F3BBAC00B1: removed
It neither buffers the email for later checking, nor does it send a bounce to the sender; it just fails.
This is in contrast to our existing Barracuda server (which I would dearly love to replace) that buffers everything for up to 48hrs, rechecking every 15m or so. Once it can contact the server and verify the address, THEN it sends a bounce or whatever as necessary.
Is this behavior possible with PMG? If not, I guess I'm done experimenting, as we have several customers with mail servers that often have poor connectivity.