systemd: Yet Another Init Replacement

unikum · 2011-03-30 17:45:51

Somebody hear about systemadm (Graphical frontend for the systemd system and session manager)? I not found package provides this program from any distributive.

litemotiv · 2011-03-30 17:58:24

Ok using systemctl's shutdown command is paramount, since the chance of non-contiguous files with a regular shutdown is quite high. Also nice is that halt for systemd is actually halt instead of poweroff, while gnu halt just shuts everything down on my system.

litemotiv · 2011-03-31 08:47:20

Upgrading seems to error out:

( 6/14) upgrading initscripts-systemd
Failed to get D-Bus connection: Connection terminated during authentication.
error: command failed to execute correctly

falconindy · 2011-03-31 10:10:23

Yep. In this one case, you're going to need a reboot (changes with dbus that I can't quite pinpoint in the commit log). It'd be really great if upstream doesn't make this a habit.

lymphatik · 2011-03-31 14:59:26

Hi,

I have two questions regarding systemd

1- I get a small trouble with ntfs partition they do not get mounted. As suggested I tried a fuser to see if something was accessing it but nothing is stated.

Mar 31 16:22:21 myhost kernel: mount[894]: Mount is denied because the NTFS volume is already exclusively opened.
Mar 31 16:22:21 myhost kernel: mount[894]: The volume may be already mounted, or another software may use it which
Mar 31 16:22:21 myhost kernel: mount[894]: could be identified for example by the help of the 'fuser' command.
Mar 31 16:22:21 myhost kernel: systemd[1]: storage.mount mount process exited, code=exited status=16
Mar 31 16:22:21 myhost kernel: systemd[1]: Unit storage.mount entered failed state.

2- I wanted to boot directly into with the old startx trick in xinit. So I have done this. And everything is quite nice except that I get a sh as shell and I start at / instead of my home directory.

[Unit]
Description=xinit for %i
Requires=dev-tty3.device
After=dev-tty3.device systemd-user-sessions.service

[Service]
User=%i
ExecStart=/bin/zsh -c startx

[Install]
WantedBy=graphical.target

Otherwise systemd is quite amazing thanks for the great work

Last edited by lymphatik (2011-03-31 15:01:45)

Zom · 2011-03-31 15:04:41

Huh. Since the update to systemd 21-1, wicd.service is no longer started during boot. I have to start it manually by doing 'systemctl start wicd.service'. I tried disabling and re-enabling it, but no go. Anybody else have the same issue?

litemotiv · 2011-03-31 15:04:50

lymphatik wrote:

2- I wanted to boot directly into with the old startx trick in xinit.

I created a /etc/systemd/system/xinit.service file like this:

[Unit]
Description=Direct login to X
Requires=dev-tty7.device
After=dev-tty7.device systemd-user-sessions.service

[Service]
ExecStart=/bin/su litemotiv -l -c "/bin/bash --login -c xinit >/dev/null 2>&1"
StandardOutput=syslog

[Install]
WantedBy=graphical.target

benneque · 2011-03-31 16:30:27

Hi. What's the current state about /run directory?
I now have /run AND /var/run, both with the same files and same timestamps. I hope systemd makes a symlink or sth like that
When will /var/run go away?

Zom · 2011-03-31 19:33:28

benneque wrote:

Hi. What's the current state about /run directory?
I now have /run AND /var/run, both with the same files and same timestamps. I hope systemd makes a symlink or sth like that
When will /var/run go away?

Straight from the mail with the announcement:

- There's now /run, which will act as new place for /var/run, but is
available during early boot. There's going to be a longer announcement
about this on fedora-devel soon.

falconindy · 2011-03-31 19:44:48

benneque wrote:

Hi. What's the current state about /run directory?
I now have /run AND /var/run, both with the same files and same timestamps. I hope systemd makes a symlink or sth like that
When will /var/run go away?

You're looking at a bind mount of tmpfs.

It's going to be years before /var/run goes away, if it ever does.

benneque · 2011-03-31 20:45:17

Yeah maybe you're right. Let's try to guess when Debian STABLE will have systemd

falconindy · 2011-04-01 02:12:38

Just pushed systemd-22. It's built with optional support for systemadm, which requires gtk2 and libnotify. Astute observers will also notice systemd-analyze, which is supposed to be some silly boot time analyzer (similar to bootchartd). You'll need dbus-python for that.

Zom · 2011-04-01 08:28:11

Alright, I think I solved my problem with wicd.service. Apparently, systemd doesn't use network.target any more on either of my machines. I don't know if this is by design or if I'm just doing something wrong. Anyway, I simply changed to multi-user.target under the [Install] section and put the new wicd.service in /etc/systemd/system.

[Unit]
Description=Wicd Network Manager
After=syslog.target

[Service]
Type=dbus
BusName=org.wicd.daemon
ExecStart=/usr/sbin/wicd -f
Restart=always

[Install]
#WantedBy=network.target
WantedBy=multi-user.target

After a 'systemctl disable wicd.service' and 'systemctl enable wicd.service' it works again.

benneque · 2011-04-01 10:26:32

Just did a "dmesg | grep systemd" and there was a small warning:
Where's the depmod.service gone?
Is it no more needed? Which service does now manage modules?

EDIT: software:
community/initscripts-systemd v21-1 (systemd) [installed]
community/systemd 22-2 (systemd) [installed]
community/systemd-arch-units 20110320-2 (systemd) [installed]

Last edited by benneque (2011-04-01 10:34:41)

falconindy · 2011-04-01 12:13:13

depmod.service did nothing but call depmod -A (which refreshes module dependency files only if required). This was ripped straight from arch initscripts, but upon closer inspection, it's really not something that makes sense. It's called late in the boot process -- after your root is mounted rw, for obvious reasons. Because it's called so late, it's not fixing anything crucial that's going to prevent you from booting. Packages which install kernel modules should be calling 'depmod $(uname -r)' or 'depmod -a' on install, so there's really not much point in keeping this silly depmod -A check around as well.

This coincides with its imminent destruction in /etc/rc.sysinit as well: http://mailman.archlinux.org/pipermail/ … 00508.html

benneque · 2011-04-01 15:21:28

Thanks!

But still nobody answered my question: (It's again plymouth)
Where can I see services that were started as dependencies? plymouth-quit.service gets called at every boot (I think). But "systemctl | grep -i ply" doesn't give any hint. "systemctl --all | grep -i ply" gives me:

plymouth...t-wait.service loaded inactive dead Wait for Plymouth Boot Screen to Quit
plymouth-start.service loaded inactive dead Show Plymouth Boot Screen
systemd-...ymouth.service error inactive dead systemd-ask-password-plymouth.service

so no plymouth-quit (just plymouth-quit-wait).

And I have still NO plymouth at shutdown/reboot...

EDIT: FOUND MISSING FILES (or wrong units):
plymouth-start.service:
Wants=systemd-ask-password-plymouth.path
Before=systemd-ask-password-plymouth.service

Those 2 unit files don't exist (at least for me)

Last edited by benneque (2011-04-01 20:18:42)

ianux · 2011-04-02 01:18:44

Hi,

Is there a service doing the job of the "Removing Leftover Files" part of /etc/rc.sysinit ? Because /tmp (amongst other files/dirs) is not flushed at boot time.

falconindy · 2011-04-02 01:26:58

/etc/tmpfiles.d controls your temporary files. If you want /tmp regularly flushed on boot, mount it on tmpfs.

fflarex · 2011-04-02 01:59:11

systemd is ignoring /etc/crypttab. Is there any way I can avoid having to type 3 separate passwords each boot? I searched around a bit but I didn't find anything useful.

falconindy · 2011-04-02 02:09:14

fflarex wrote:

systemd is ignoring /etc/crypttab. Is there any way I can avoid having to type 3 separate passwords each boot? I searched around a bit but I didn't find anything useful.

I'm seeing /etc/crypttab being parsed in src/cryptsetup-generator.c. The fopen call is line 232 in current git HEAD. Check your logs to make sure it isn't barfing on a malformed file (might need to boot with 'verbose'). I'm not sure how closely Arch's cryptsetup adheres to the standard.

Last edited by falconindy (2011-04-02 02:09:52)

fflarex · 2011-04-02 02:24:40

Fixed it. My passwords in /etc/crypttab were interpreted as paths to key files. I'm not sure if this is a bug or not, but it certainly doesn't support all the features of crypttab that Arch does. I also had to remove Arch-specific handling of encrypted swap.

lymphatik · 2011-04-02 09:05:33

I do not know if it's a bug or not but every time I want to launch lighttpd service I have to create a directory "/var/run/lighttpd"

So I created a conf file in /etc/tmpfile.d but I am not sure it should be done that way

d /var/run/lighttpd 755 http http

Last edited by lymphatik (2011-04-02 09:06:01)

benneque · 2011-04-02 11:04:44

@lymphatik: That's exactly the way it should be done! (The other "right" way would be, if lighthttpd creates the folder itself)

Cosmin · 2011-04-03 12:10:09

I get this error when I install systemd

Failed to get D-Bus connection: Failed to connect to socket /org/freedesktop/systemd1/private: Connection refused

any ideas?

Last edited by Cosmin (2011-04-03 12:10:30)

falconindy · 2011-04-03 13:41:22

Cosmin wrote:

I get this error when I install systemd
Failed to get D-Bus connection: Failed to connect to socket /org/freedesktop/systemd1/private: Connection refused
any ideas?

If you installed (not upgraded) just systemd, something in awry. Nothing in the install scriptlet could generate this error. Did you mean that this came from installing the initscripts package? In the end, you can ignore this error and reboot. It's harmless if systemd isn't even running (which is generally what causes the error).

Arch Linux

#326 2011-03-30 17:45:51

Re: systemd: Yet Another Init Replacement

#327 2011-03-30 17:58:24

Re: systemd: Yet Another Init Replacement

#328 2011-03-31 08:47:20

Re: systemd: Yet Another Init Replacement

#329 2011-03-31 10:10:23

Re: systemd: Yet Another Init Replacement

#330 2011-03-31 14:59:26

Re: systemd: Yet Another Init Replacement

#331 2011-03-31 15:04:41

Re: systemd: Yet Another Init Replacement

#332 2011-03-31 15:04:50

Re: systemd: Yet Another Init Replacement

#333 2011-03-31 16:30:27

Re: systemd: Yet Another Init Replacement

#334 2011-03-31 19:33:28

Re: systemd: Yet Another Init Replacement

#335 2011-03-31 19:44:48

Re: systemd: Yet Another Init Replacement

#336 2011-03-31 20:45:17

Re: systemd: Yet Another Init Replacement

#337 2011-04-01 02:12:38

Re: systemd: Yet Another Init Replacement

#338 2011-04-01 08:28:11

Re: systemd: Yet Another Init Replacement

#339 2011-04-01 10:26:32

Re: systemd: Yet Another Init Replacement

#340 2011-04-01 12:13:13

Re: systemd: Yet Another Init Replacement

#341 2011-04-01 15:21:28

Re: systemd: Yet Another Init Replacement

#342 2011-04-02 01:18:44

Re: systemd: Yet Another Init Replacement

#343 2011-04-02 01:26:58

Re: systemd: Yet Another Init Replacement

#344 2011-04-02 01:59:11

Re: systemd: Yet Another Init Replacement

#345 2011-04-02 02:09:14

Re: systemd: Yet Another Init Replacement

#346 2011-04-02 02:24:40

Re: systemd: Yet Another Init Replacement

#347 2011-04-02 09:05:33

Re: systemd: Yet Another Init Replacement

#348 2011-04-02 11:04:44

Re: systemd: Yet Another Init Replacement

#349 2011-04-03 12:10:09

Re: systemd: Yet Another Init Replacement

#350 2011-04-03 13:41:22

Re: systemd: Yet Another Init Replacement

Board footer