Postfix Problems

Quote from seaxels

found this thread after realizing I haven't received healthchecks messages for over two weeks. Similar issue - same errors from postfix (proxymap and cleanup connection refused). Reboot doesn't help. Uninstalling flashmemory plugins restores the service, but would be great if anyone had a proper solution.

The correct solution at the moment is described in the post above. Install the test Flash plugin, after removing the plugin that was installed from the repository. As for me, the latest test version works as it should...

Hi, after upgradeing my 2 OMV Server from 6 to 7. I have the maybe same problem as you all.

The postfix service did not run well after the upgrades.

I figured out, that disable and enable notfication makes the services runing till a reboot.

And I figured out, that if you delete the file "/var/lib/postfix/master.lock" it is posible to start the postfix service again. But disable and enable the service helps only for some houres.

Maybe there is a problem with the permisons of the master.lock file.

As a partial solution, I am now using discord notifications instead of email notifications:

Discord notification for OpenMediaVault

Discord notification for OpenMediaVault. GitHub Gist: instantly share code, notes, and snippets.

gist.github.com

I've been hit by this also after several years using OMV (jumping from Pi4 to CM4 to amd64) without any issue regarding postfix@-.service and always using flashmemory plugin.

That flashmemory plugin was updated to the zram one and ran also without issues (except the one's from zram itself)

A few days (weeks) ago, it was reverted to the old folder2ram and that's when I saw the problem:

postfix would fail on a reboot.

What I've been able to gather is that the failed postfix only happens if I use reboot (on the GUI).

If I use a shutdown command (did that several times due to house reno) , after I boot back the system, there's no failed units and postfix works normal.

Can any of you guys try this:

Having the flashmemory plugin in use, do a shutdown.

After a few moments, boot up the system and once the OMV GUI becomes available (means that all services should be UP) go on CLI and run:

sudo systemctl --failed

~ $ sudo systemctl --failed

  UNIT LOAD ACTIVE SUB DESCRIPTION
0 loaded units listed.

ryecoaaron Does this makes any sense or help in any way to triage what happens?

If you need me to reproduce this, it's easy.

Just tested. flashmemory plugin on amd64 system. Get after both (reboot and shutdown):

sudo systemctl --failed
UNIT LOAD ACTIVE SUB DESCRIPTION
0 loaded units listed.

Usually I do a shutdown over night with the autoshutdown plugin. The next morning postfix won't start. (That's why i run a cronjob to restart porstfix).

Quote from moflix

That's why i run a cronjob to restart porstfix

If you can, disable the cronjob and test again after a reboot and after a shutdown.

The cronjob is just a workaround. Not a fix.

The ronjob runs at a fixed time and didn't interfere with above tests.

Soma

I´ve made a clean reinstall of my system around two weeks ago... Using folder2ram... 6.1.32 kernel image.

I guess i never had postfix service errors after shutdown and reboot...

But, other services like nginx, php, etc. sometimes not coming up correct and

i´m gettting email notifications about... Looks like timing, but i´m not an expert!

May be it´s a basic problem anywhere...?!

Best regards

prtigger

Seems like i'm also affected by this phenomenon.

This was the situation so far:

On all my 4 NAS I had problems with e-mail notifications.

On all my 4 NAS I see log messages like this one:

postfix/pickup[2419527]: warning: connect #1 to subsystem public/cleanup: Connection refused

On all my 4 NAS the plugin Flashmemory was installed.

(I'm not convinced, that there is relation to the plugin, though)

What I did so far:

-uninstall the plugin on all NAS

-either reboot the NAS or restart postfix

-installed again the plugin Flashmemory on one of my NAS

Judging from the mails, that came in, sending mails was blocked for at least 3 weeks, maybe longer.

I did this yesterday evening. So far mail notification is still working for all NAS (including the one with Flashmemory installed). But I fear it will eventually stop working on one or more of my NAS.

It seems that yesterday mail were stuck again on one of my NAS.

It is that NAS where I activated Flashmemory.

Is there anything I could provide here as information, before I restart postfix again?

Hi, I just upgraded (reinstalled) to OMV7 and I'm seeing this issue for the first time ever.

Using flashmemory plugin.

I am getting in the logs.

May 16 12:45:15 nas postfix/smtpd[4601]: warning: connect #3 to subsystem private/proxymap: Connection refused
May 16 12:45:25 nas postfix/smtpd[4601]: warning: connect #4 to subsystem private/proxymap: Connection refused
May 16 12:45:35 nas postfix/smtpd[4601]: warning: connect #5 to subsystem private/proxymap: Connection refused
May 16 12:45:45 nas postfix/smtpd[4601]: warning: connect #6 to subsystem private/proxymap: Connection refused
May 16 12:45:55 nas postfix/smtpd[4601]: warning: connect #7 to subsystem private/proxymap: Connection refused

Emails are not being delivered, but sometimes they do.

What's the status of this issue?

Thank you for this update ryecoaaron we hope in OMV 8 we can take advantage of your great work on the plugin.

So for OMV7 on my 3 servers thus far, the workaround whilst keeping flashmemory plugin installed and running is per my previous post.

The solution and workaround discussed was to set a delay of 30 seconds to the service, before launching it, likely fixing the issue. Run systemctl edit postfix.service at the prompt, and enter the data below (those with brackets, obviously )

% systemctl edit postfix.service

[Unit]
StartLimitIntervalSec=300

[Service]
ExecStartPre=/bin/sleep 30

After adding these parameters, rebooting the system should now get a working postfix service after a few seconds; test with sendmail -q command. I have had no issues, personally.

Quote from Konsti

Thank you for this update ryecoaaron we hope in OMV 8 we can take advantage of your great work on the plugin.

So for OMV7 on my 3 servers thus far, the workaround whilst keeping flashmemory plugin installed and running is per my previous post.

The solution and workaround discussed was to set a delay of 30 seconds to the service, before launching it, likely fixing the issue. Run systemctl edit postfix.service at the prompt, and enter the data below (those with brackets, obviously )

Code

% systemctl edit postfix.service

[Unit]
StartLimitIntervalSec=300

[Service]
ExecStartPre=/bin/sleep 30

After adding these parameters, rebooting the system should now get a working postfix service after a few seconds; test with sendmail -q command. I have had no issues, personally.

Display More

Hi, I made the changes you recommended.

I rebooted and doublechecked they have been apllied, but my Postfix is still a mess:

EDIT: I've run systemd-analyze blame and these tewo services take more than 30 secs to start up:

33.489s openmediavault-issue.service

33.217s folder2ram_startup.service

I have no idea of what's going on should I open a separate discussion?
I'll try to increase your reccomendation to 40 seconds.

EDIT2: 40 seconds is enough for my system!

Sorry for the spam, but the solution unfortunately is not reliable, I can't get postfix to work reliably.
I've done another reboot and it's stuck:

root@nas:~# journalctl -f
May 29 18:41:02 nas postfix/smtpd[13978]: warning: connect #1 to subsystem private/proxymap: Connection refused
May 29 18:41:12 nas postfix/smtpd[13978]: warning: connect #2 to subsystem private/proxymap: Connection refused
May 29 18:41:22 nas postfix/smtpd[13978]: warning: connect #3 to subsystem private/proxymap: Connection refused
May 29 18:41:32 nas postfix/smtpd[13978]: warning: connect #4 to subsystem private/proxymap: Connection refused
May 29 18:41:42 nas postfix/smtpd[13978]: warning: connect #5 to subsystem private/proxymap: Connection refused
May 29 18:41:52 nas monit[795]: Mail: Error receiving data from the mailserver -- Resource temporarily unavailable
May 29 18:41:52 nas monit[795]: Alert handler failed, retry scheduled for next cycle
May 29 18:41:52 nas postfix/smtpd[13978]: warning: connect #6 to subsystem private/proxymap: Connection refused
May 29 18:42:02 nas postfix/smtpd[13978]: warning: connect #7 to subsystem private/proxymap: Connection refused
May 29 18:42:12 nas postfix/smtpd[13978]: warning: connect #8 to subsystem private/proxymap: Connection refused
May 29 18:42:22 nas postfix/smtpd[13978]: warning: connect #9 to subsystem private/proxymap: Connection refused
May 29 18:42:32 nas postfix/smtpd[13978]: warning: connect #10 to subsystem private/proxymap: Connection refused

Hi, I am hijacking this thread just to make ryecoaaron aware that I've found another issue related to the flashmemoryplugin:

Post

RE: "doest not exist nginx" alerts on a new OMV7 install

I can confirm that by uninstalling the flashmemory plugin the email alert goes away. I've rebooted the system twice.

I already have the latest stable kernel: Linux 6.12.22+bpo-amd64

auanasgheps

Jun 6th 2025

It's just an alert, the service doesn't get stuck like postfix, but I wanted to make you aware.