+++ This bug was initially created as a clone of Bug #1954802 +++

Description of problem:

On very busy systems all of the 'systemd --user' processes will all echo messages that the primary systemd process does.  This creates a massive cascade of logs which is completely unnecessary.

In RHEL 8.3 we added a mount succeeded message and on systems with tons of autofs and users logged in we get every single 'systemd --user' process all echoing the mount succeeded message.

From unit.c in RHEL 8.3.0: 

    void unit_log_success(Unit *u) {
	    assert(u);

	    log_struct(LOG_INFO,
		       "MESSAGE_ID=" SD_MESSAGE_UNIT_SUCCESS_STR,
		       LOG_UNIT_ID(u),
		       LOG_UNIT_INVOCATION_ID(u),
		       LOG_UNIT_MESSAGE(u, "Succeeded."));
    }

Version-Release number of selected component (if applicable):

RHEL 8.3 systemd-239-41.el8_3.1.x86_64

How reproducible:

Always

Steps to Reproduce:
1. Have a user logged in and mount a filesystem

Actual results:

systemd overwhelms logs on very busy systems.

Expected results:

Only the primary systemd process should indicate that these mounts succeeded or this should be hidden behind a debug or other logging.

Additional info:

--- Additional comment from Jan Macku on 2021-05-24 13:43:37 UTC ---

Hi,
Anthony I was able to reproduce the issue with the following reproducer.

Set persistent journal, then:
> # useradd testuser
Open session for user testuser e.g. via ssh. As root mount tmpfs and unmount it.
> # mount -t tmpfs none /mnt
> # umount /mnt
> # journalctl -u user
> ...
> ... systemd[5322]: mnt.mount: Succeeded.

Once I have the fix I'll give you an update.

Jan