One bounce, 5 bounce counts
I recently upgraded Mailman to GNU Mailman 3.3.4 (Tom Sawyer) that once again handles bounces. Thank you.
In my setup, Mailman is with Maxking's Docker image maxking/mailman-core:0.3, and a Postfix outside Docker handling incoming and outgoing mails.
Since the upgrade some list administrators have expectedly complained about every time they reenable subscribers, they are disabled again. I have explained to them why they are disabled and why they may be disabled again when re-disabled.
My problem, though, is that some subscribers may be disabled even they are able to receive mails (confirmed with sendmail -bv), and I have been looking at one subscriber in particular.
- Can there be (i'm not saying there are, maybe it's my lack of knowledge) somthing regarding Mailman?
- And what would the appropeiate per list bounce settings, if not default?
Actual example:
- Mailman sent seemly 6 mails to a member on January 6th.
- I see one bouce message D1426CE22A in Postfix log that was delivered back to Mailman
- I see that Mailman is counting the bounce score from 1 to 5, one at the time, but that was day after receiving the bounce, and no other mails have been sent to the user on the 7th nor received as bounce on the 7th. I suspect that it must have been in the Postfix queue until then, but I wonder why that could be - but could be.
- The bounce would be in the Postfix queue until sent to Mailman and thus removed from the Postfix queue.
== /var/log/maillog == Jan 3 09:05:42 lstapp01pl postfix/smtp[10293]: EA1CC73105: to=recipient@example.com<mailto:heidi.ryssel@regionh.dk>, relay=rghhehspamfw001.regionh.dk[185.211.204.236]:25, delay=2.2, delays=0.14/0.03/0.1/1.9, dsn=2.0.0, status=sent (250 Ok: queued as 16A9084006E) Jan 3 11:27:35 lstapp01pl postfix/smtp[27118]: 0AD2F73117: to=recipient@example.com<mailto:heidi.ryssel@regionh.dk>, relay=rghhehspamfw002.regionh.dk[185.211.204.237]:25, delay=0.79, delays=0.14/0.03/0.15/0.47, dsn=2.0.0, status=sent (250 Ok: queued as C2CCDF20052) Jan 6 09:09:45 lstapp01pl postfix/smtp[2132]: EFAA673126: to=recipient@example.com<mailto:heidi.ryssel@regionh.dk>, relay=rghhvhspamfw002.regionh.dk[185.211.204.239]:25, delay=0.31, delays=0.09/0.01/0.07/0.14, dsn=2.0.0, status=sent (250 Ok: queued as 38094120005E) Jan 6 09:10:41 lstapp01pl postfix/smtp[2132]: 33F6373126: to=recipient@example.com<mailto:heidi.ryssel@regionh.dk>, relay=rghhehspamfw001.regionh.dk[185.211.204.236]:25, delay=0.35, delays=0.08/0/0.06/0.21, dsn=2.0.0, status=sent (250 Ok: queued as 86A0B840075) Jan 6 13:37:31 lstapp01pl postfix/qmgr[28946]: D1426CE22A: from=recipient@example.com<mailto:heidi.ryssel@regionh.dk>, size=11551, nrcpt=1 (queue active) Jan 6 14:06:43 lstapp01pl postfix/smtp[16749]: 89A68D5699: to=recipient@example.com<mailto:heidi.ryssel@regionh.dk>, relay=rghhvhspamfw001.regionh.dk[185.211.204.238]:25, delay=0.35, delays=0.09/0/0.07/0.2, dsn=2.0.0, status=sent (250 Ok: queued as D97EC37C0072) Jan 6 15:04:23 lstapp01pl postfix/smtp[6124]: A46527312C: to=recipient@example.com<mailto:heidi.ryssel@regionh.dk>, relay=rghhehspamfw002.regionh.dk[185.211.204.237]:25, delay=0.65, delays=0.12/0.03/0.03/0.48, dsn=2.0.0, status=sent (250 Ok: queued as 44ABBF2008C)
== /opt/mailman/var/logs/bounce.log == Jan 07 03:36:38 2022 (25) Member recipient@example.com<mailto:heidi.ryssel@regionh.dk> on list listname.lists.example.com, bounce score = 1. Jan 07 03:36:56 2022 (25) Member recipient@example.com<mailto:heidi.ryssel@regionh.dk> already scored a bounce on list listname.lists.example.com today. Jan 07 03:37:27 2022 (25) Member recipient@example.com<mailto:heidi.ryssel@regionh.dk> already scored a bounce on list listname.lists.example.com today. Jan 07 03:54:04 2022 (25) Member recipient@example.com<mailto:heidi.ryssel@regionh.dk> already scored a bounce on list listname.lists.example.com today. Jan 07 03:58:10 2022 (25) Member recipient@example.com<mailto:heidi.ryssel@regionh.dk> on list listname.lists.example.com, bounce score = 2. Jan 07 04:05:04 2022 (25) Member recipient@example.com<mailto:heidi.ryssel@regionh.dk> already scored a bounce on list listname.lists.example.com today. Jan 07 04:25:08 2022 (25) Member recipient@example.com<mailto:heidi.ryssel@regionh.dk> already scored a bounce on list listname.lists.example.com today. Jan 07 04:51:22 2022 (25) Member recipient@example.com<mailto:heidi.ryssel@regionh.dk> on list listname.lists.example.com, bounce score = 3. Jan 07 05:04:20 2022 (25) Member recipient@example.com<mailto:heidi.ryssel@regionh.dk> on list listname.lists.example.com, bounce score = 4. Jan 07 05:05:22 2022 (25) Member recipient@example.com<mailto:heidi.ryssel@regionh.dk> already scored a bounce on list listname.lists.example.com today. Jan 07 22:55:06 2022 (25) Member recipient@example.com<mailto:heidi.ryssel@regionh.dk> on list listname.lists.example.com, bounce score = 5. Jan 07 22:55:06 2022 (25) Disabling delivery for recipient@example.com<mailto:heidi.ryssel@regionh.dk> on list listname.lists.example.com by bounce Jan 07 22:55:06 2022 (25) Member recipient@example.com<mailto:heidi.ryssel@regionh.dk> on list listname.lists.example.com, bounce score 5 >= threshold 5, disabling delivery.
Regards, Henrik Rasmussen
On 1/13/22 6:17 AM, Henrik Rasmussen wrote:
My problem, though, is that some subscribers may be disabled even they are able to receive mails (confirmed with sendmail -bv), and I have been looking at one subscriber in particular.
- Can there be (i'm not saying there are, maybe it's my lack of knowledge) somthing regarding Mailman?
- And what would the appropeiate per list bounce settings, if not default?
Referring to the example below, it seems that your Postfix delivers successfully to the first hop MX of the recipient, but the mail bounces further downstream resulting in an emailed bounce.
One thing you can do to help mitigate scoring of bounces of emails due to spam filtering is to set
[mta]
verp_probes: yes
This will cause Mailman to not disable delivery when bounce score reaches threshold, but rather send a probe which hopefully won't be detected as spam and only disable delivery if the probe bounces.
Actual example:
- Mailman sent seemly 6 mails to a member on January 6th.
- I see one bouce message D1426CE22A in Postfix log that was delivered back to Mailman
- I see that Mailman is counting the bounce score from 1 to 5, one at the time, but that was day after receiving the bounce, and no other mails have been sent to the user on the 7th nor received as bounce on the 7th. I suspect that it must have been in the Postfix queue until then, but I wonder why that could be - but could be.
- The bounce would be in the Postfix queue until sent to Mailman and thus removed from the Postfix queue.
I suspect the issue is something like a connection problem from the recipient MX to a relay MX that is not being reported as a bounce until the retry interval has expired.
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
participants (2)
-
Henrik Rasmussen -
Mark Sapiro