PMG 8.2.5 - Emails in the queue that bounce back fill the log until the disk is full.

[JoelTimm]

Dear community,

I have tried to solve this problem, but I can't find a real solution. The emails in the queue that are sent back to our internal mail server fill up the log until the hard drive is full.
Then PMG stops responding. Of course, this always happens when I'm traveling or on vacation. This happens very often with old email addresses that our internal mail server no longer accepts. I tried to create a rule to block (invalid email list with ndr) these email addresses in PMG as well, but PMG itself tries to send SPAM reports to these non-existent email addresses. The logs are each 1.5 GB in size (syslog.1 and mail.log.1). We don't have a large hard drive for PMG. This becomes critical here because it happens too often.

I have set up daily log rotation, but that is not enough.

Is there anything that can be done to solve this problem? Maybe I'm overlooking something.

regards,

Joel.

 

[Stoiko Ivanov]

. This happens very often with old email addresses that our internal mail server no longer accepts.

This is usually handled by enabling recipient verification:
https://pmg.proxmox.com/pmg-docs/pmg-admin-guide.html#pmgconfig_mailproxy_options

The logs are each 1.5 GB in size (syslog.1 and mail.log.1). We don't have a large hard drive for PMG. This becomes critical here because it happens too often.

you can disable the logging to /var/log/mail.log in rsyslog.conf (/var/log/syslog is used for the tracking center)

but PMG itself tries to send SPAM reports to these non-existent email addresses.

also fixed by recipient verification

I hope this helps!

 

[JoelTimm]

Thank you for your reply.

The problem is that recipient verification is already enabled (with 550) and the issue still occurs.

Is there any further advice for our case?

Thank you in advance.

 

[Stoiko Ivanov]

Thank you for your reply.

The problem is that recipient verification is already enabled (with 550) and the issue still occurs.

This is odd - two things come to mind:
* make sure that mails from the internet are received on the _external port_ of PMG (this is where recipient verification is used)
* check if your downstream/internal mailserver responds with a permanent error-code 5xx for non-exisiting recipients after the RCPT TO command (some do so only after DATA - IIRC some versions of Exchange needed a configuration change for this to be the case).

If this does not help - discard the verification database (GUI->Administration->Queues) - send an email to a non-existing address in your domain and share the logs

I hope this helps!