daemon-reload runs alsa-restore.service and others

[Unit 193]

When systemd is reloaded (manually, upgrades), the volume is reset, likely using alsa-restore.

[Pupykin-s (pupykin-s)]

Steps to reproduce:
- became root
- run alsamixer, change master volume
- in another terminal run "systemctl disable qwe.socket"
master volume changes back.

It looks like systemctl disable qwe.socket cause systemctl start alsa-restore.service

Changing sound volume is not what I expected when disabling .socket)

ArchLinux x86_64, kernel 3.18.2-2, systemd 218-1.

[Pupykin-s (pupykin-s)]

qwe.socket does not exist, but it does not matter. Probably systemctl disable reloads target including alsa-* services.

[Pupykin-s (pupykin-s)]

May be the prolem is related to alsa:

/usr/lib/systemd/system/basic.target.wants contains

alsa-restore.service -> ../alsa-restore.service
alsa-state.service -> ../alsa-state.service

How to fix this issue properly?

[Lennart-poettering (lennart-poettering)]

Could you please turn on debug mode in PID with "systemd-analyze set-log-level debug". Then please reproduce the issue and attach the logs generated from PID 1 about this. Thanks!

[zbyszek (zbyszek-in)]

Should be fixed with http://cgit.freedesktop.org/systemd/systemd/commit/?id=6e392c9c45.

[Intelfx100 (intelfx100)]

(In reply to Zbigniew Jedrzejewski-Szmek from comment #4)
> Should be fixed with
> http://cgit.freedesktop.org/systemd/systemd/commit/?id=6e392c9c45.

Thanks for finally reviewing it :)

Is it going to be backported?

[zbyszek (zbyszek-in)]

That's really up to distributions. For Fedora, most likely.

[Intelfx100 (intelfx100)]

(In reply to Zbigniew Jedrzejewski-Szmek from comment #6)
> That's really up to distributions. For Fedora, most likely.

I've meant the stable branches (systemd-stable).

[zbyszek (zbyszek-in)]

(In reply to Ivan Shapovalov from comment #7)
> (In reply to Zbigniew Jedrzejewski-Szmek from comment #6)
> > That's really up to distributions. For Fedora, most likely.
>
> I've meant the stable branches (systemd-stable).

Yes.

[Martin Pitt (pitti)]

This can be seen in journalctl -f when doing daemon-reload:

Mär 12 08:33:48 donald systemd[1]: Reloading.
Mär 12 08:33:48 donald systemd[1]: Unit type .busname is not supported on this system.
Mär 12 08:33:48 donald systemd[1]: Started Set Up Additional Binary Formats.
Mär 12 08:33:48 donald systemd[1]: Mounted Configuration File System.
Mär 12 08:33:48 donald systemd[1]: Started Commit a transient machine-id on disk.
Mär 12 08:33:48 donald systemd[1]: Started Various fixups to make systemd work better on Debian.
Mär 12 08:33:48 donald systemd[1]: Starting Tell Plymouth To Write Out Runtime Data...
Mär 12 08:33:48 donald systemd[1]: Started Dispatch Password Requests to Console Directory Watch.
Mär 12 08:33:48 donald systemd[1]: Started Manage Sound Card State (restore and store).
Mär 12 08:33:48 donald systemd[1]: Starting Restore Sound Card State...
Mär 12 08:33:48 donald systemd[1]: Started Recovery mode menu.

[Martin Pitt (pitti)]

After that happens, can you please give me the output of "systemctl status alsa-restore.service" and "ls -la /var/lib/alsa"?

[Unit 193 (unit193)]

● alsa-restore.service - Restore Sound Card State
   Loaded: loaded (/lib/systemd/system/alsa-restore.service; static; vendor preset: enabled)
   Active: inactive (dead) since Fri 2015-03-13 05:18:58 EDT; 11s ago
  Process: 20877 ExecStart=/usr/sbin/alsactl -E HOME=/var/run/alsa restore (code=exited, status=0/SUCCESS)
 Main PID: 20877 (code=exited, status=0/SUCCESS)

Mar 13 05:18:58 Alpha systemd[1]: Starting Restore Sound Card State...
Mar 13 05:18:58 Alpha systemd[1]: Started Restore Sound Card State.

total 20
drwxr-xr-x 2 root root 4096 Mar 12 18:27 .
drwxr-xr-x 76 root root 4096 Mar 7 21:27 ..
-rw-r--r-- 1 root root 6849 Mar 12 18:27 asound.state
-rw-r--r-- 1 root root 11 Aug 20 2014 asound.state.lock

Last bit is interesting.

[Martin Pitt (pitti)]

For the record, I don't see the behaviour from comment 1 with current systemd any more.

[Martin Pitt (pitti)]

I tried with apt-get install --reinstall systemd udev, several "systemctl daemon-reexec" and "daemon-reload", but I just don't get this bug any more. Unit193, do you?

[Joe Barnett (thejoe)]

I still see the green screen problem (marked as dup of this bug). although it happens on non-systemd upgrades...

[Martin Pitt (pitti)]

@Joe: unduplicating then, following up on your original bug

[Unit 193 (unit193)]

I see this on neither of the two systems I had seen it on. Looks like a recent upload fixed it.

[Martin Pitt (pitti)]

Thanks for confirming. Tentatively closing for now. If anyone still sees this, please yell.