Control: fixed -1 254.1-2

On Fri, 2 Jun 2023 13:22:02 -0400 Jason Franklin <ja...@oneway.dev> wrote:
Package: systemd
Version: 252.6-1
Severity: normal

Greetings:

It appears that user@.service is left in a failed state when a user logs
in over ssh and is forced to reset their password due to expiration.

I was able to regularly reproduce this behavior with the process
described below.

# Create a test user.
$ sudo adduser fish
...

# Ensure no failed units.
$ systemctl list-units --failed
...

# Expire the user's password.
$ sudo passwd -e fish

# Log in via ssh. Properly change the user's password when prompted.
$ ssh fish@localhost
...
# Here, you will be kicked back out to your prompt after the new
# password is set.

# Now, note that a failed unit for the user is left behind.
$ systemctl list-units --failed
  UNIT               LOAD   ACTIVE SUB    DESCRIPTION
* user@1001.service  loaded failed failed User Manager for UID 1001
...

I think the above accurately describes the behavior I'm seeing.

It seems odd to me that the failed service lingers even though the
user's password has been changed correctly, without any real failures in
the process. The user is now able to log in and what not, but systemd
indicates a failure.

This is an issue for me because these failures can stack up on systems
at my work, and monitoring of failed units then indicates a problem
where there is not one.

Please let me know any thoughts on this. It could be another piece of
software that's causing the error. Or, it could be that I have something
improperly configured or that I am misinterpreting things.

Seems to work in trixie/sid, so marking as fixed for that version.


Attachment: OpenPGP_signature.asc
Description: OpenPGP digital signature

Reply via email to