Issue Recurring Amavis timeout issue — "conversation with 127.0.0.1 timed out" recurring every N months

[mcbus94]

Server operating system version

AlmaLinux 9.8

Plesk version and microupdate number

Plesk Obsidian v18.0.79_build1800260720.04 os_RedHat el9

• Plesk Obsidian on AlmaLinux 9.8
• Amavisd 2.13.1-1.el9 (from EPEL)
• AMD EPYC 7313 (32 cores), 66GB RAM, nginx + PHP 8.4 FPM, MariaDB

Emails fail to send with the following error recurring every N months (pattern not yet identified):

4.4.2 deferred (delivery temporarily suspended: conversation with 127.0.0.1[127.0.0.1] timed out while receiving the initial server greeting)

Last occurrence (today, 24 Jul 2026):

• 130+ messages queued and blocked, spanning ~12 hours (from ~12:36 until now)
• All postfix SMTP processes show status "delivery temporarily suspended"
• Amavisd logs show: (!!)TROUBLE in child_init_hook: BDB can't connect db env. for multiple child processes

Temporary fix (applied successfully):
Following https://support.plesk.com/hc/en-us/articles/15657655076375, restarting services resolves it:

systemctl restart amavisd
systemctl restart postfix
postsuper -r ALL

Messages then process normally.

Problem recurs periodically (frequency TBD, likely monthly or quarterly). Is this a known issue with Amavisd 2.13.1 on AlmaLinux 9.8?

Thank you in advance

 

[Sebahat.hadzhi]

Hi, @mcbus94 . Does the issue appear in relatively the same time - is there any pattern you can catch, or does it appear randomly? Please check if OOM has occurred:

journalctl -k | grep -i -e memory -e oom

Also, you may use the sar utility to check CPU and system load:

sar -q 2 5

If unavailable, you can install it as follow:

yum install sysstat
service sysstat start

 

[mcbus94]

I can't define a specific time because I only notice the problem when I log into Plesk or in the evening I realize I haven't received a single email.
No memory and no cpu issue, we are using Plesk monitoring.

 

[Sebahat.hadzhi]

Thank you for the confirmation. Please navigate to Tools & Settings > Plesk Email Security > Server Settings > Advanced, and confirm if "Concurrent Amavis Processes" is set to the default value of 10. If yes, please try increasing it to 20, for example.

 

[mcbus94]

Hi. The issue recurred today. Email are not sent/delivered.
The first email "blocked" has been sent at 13:40:11

<.......@.........>: delivery temporarily suspended: conversation with 127.0.0.1[127.0.0.1] timed out while receiving the initial server greeting
recipient=.........@.............
offset=747
dsn_orig_rcpt=rfc822;..............@..............
status=4.4.2
action=delayed
reason=delivery temporarily suspended: conversation with 127.0.0.1[127.0.0.1] timed out while receiving the initial server greeting

journalctl -k | grep -i -e memory -e oom
Aug 04 13:15:43 ............. kernel: kcare: allocating memory in module space...

Here below more details that could be useful to find the issue:

$ sudo journalctl -k --since "2026-08-04 13:00:00" | grep -E "kcare|kpatch"
Aug 04 13:15:43 ............. kernel: kcare: allocated 15188191 bytes for patch at ffffac492930d000
Aug 04 13:15:43 ............. kernel: kcare: allocating memory in module space...
Aug 04 13:15:43 ............. kernel: kcare: allocated 15187680 bytes at ffffffffc1f10000
Aug 04 13:15:43 ............. kernel: kpatch: 13 entries unpatched
Aug 04 13:15:43 ............. kernel: kpatch: patching vmlinux, nf_conntrack, xfs, kvm, tls, drm_kms_helper, bnxt_en, dm_mod...

$ sudo journalctl -u amavisd --since "2026-08-04 13:15:00" --until "2026-08-04 13:40:00"
Aug 04 13:22:29 ............. amavis[87163]: NOTICE: reconnecting in response to: err=2006, HY000, DBD::MariaDB::st execute failed: Server has gone away at /usr/share/perl5/vendor_perl/Amavis/Out/SQL/Connection.pm line 175.
Aug 04 13:35:50 ............. amavis[87164]: NOTICE: reconnecting in response to: err=2006, HY000, DBD::MariaDB::st execute failed: Server has gone away
Aug 04 13:35:50 ............. amavis[87165]: NOTICE: reconnecting in response to: err=2006, HY000, DBD::MariaDB::st execute failed: Server has gone away
Aug 04 13:35:50 ............. amavis[87166]: NOTICE: reconnecting in response to: err=2006, HY000, DBD::MariaDB::st execute failed: Server has gone away
Aug 04 13:35:50 ............. amavis[87167]: NOTICE: reconnecting in response to: err=2006, HY000, DBD::MariaDB::st execute failed: Server has gone away
Aug 04 13:35:50 ............. amavis[87168]: NOTICE: reconnecting in response to: err=2006, HY000, DBD::MariaDB::st execute failed: Server has gone away

$ sudo postqueue -p | tail -1
-- 4188 Kbytes in 130 Requests.

I have just set the "Concurrent Amavis Processes" from 10 to 20. Amavis and postfix services have been restarted and email delivery is now OK. But I woudl like to find a solution without reboot amavis and postfix.

Thank you in advance