systemd: Yet Another Init Replacement

WorMzy · 2011-10-23 17:37:51

falconindy wrote:

No, they can be removed. There's a large table of guaranteed mounts in mount-setup.c:

static const MountPoint mount_table[] = {
        { "proc",     "/proc",                  "proc",     NULL,                MS_NOSUID|MS_NOEXEC|MS_NODEV, true },
        { "sysfs",    "/sys",                   "sysfs",    NULL,                MS_NOSUID|MS_NOEXEC|MS_NODEV, true },
        { "devtmpfs", "/dev",                   "devtmpfs", "mode=755",          MS_NOSUID,                    true },
        { "tmpfs",    "/dev/shm",               "tmpfs",    "mode=1777",         MS_NOSUID|MS_NODEV,           true },
        { "devpts",   "/dev/pts",               "devpts",   "mode=620,gid=" STRINGIFY(TTY_GID), MS_NOSUID|MS_NOEXEC, false },
        { "tmpfs",    "/run",                   "tmpfs",    "mode=755",          MS_NOSUID|MS_NODEV, true },
        { "tmpfs",    "/sys/fs/cgroup",         "tmpfs",    "mode=755",          MS_NOSUID|MS_NOEXEC|MS_NODEV, false },
        { "cgroup",   "/sys/fs/cgroup/systemd", "cgroup",   "none,name=systemd", MS_NOSUID|MS_NOEXEC|MS_NODEV, false },
};

Aw. And I thought I was being clever looking for mount scripts. :C

I guess they're essential mounts for systemd to run properly/efficiently? Or is there a different reason for why they're hard-coded into the array?

Last edited by WorMzy (2011-10-23 17:38:15)

falconindy · 2011-10-23 17:53:39

Right. Because they're essential for proper operation.

ron9 · 2011-10-24 07:17:58

nemesis2all wrote:

Need help setting up systemd. I have sshd and dhcdbd fail. I assume it is because it is trying to start those services before network has finished. Can someone help me with proper setup for this?

Have you tried with systemd.automount?
With comment=systemd.automount in fstab it will be mounted when it is needed.
http://0pointer.de/public/systemd-man/s … mount.html
http://0pointer.de/public/systemd-man/s … mount.html

.

nemesis2all · 2011-10-25 13:48:44

ron9 wrote:

nemesis2all wrote:
Need help setting up systemd. I have sshd and dhcdbd fail. I assume it is because it is trying to start those services before network has finished. Can someone help me with proper setup for this?
Have you tried with systemd.automount?
With comment=systemd.automount in fstab it will be mounted when it is needed.
http://0pointer.de/public/systemd-man/s … mount.html
http://0pointer.de/public/systemd-man/s … mount.html
.

I have not tried that yet. Would it have an adverse effect on booting without using systemd?

ron9 · 2011-10-25 18:13:54

nemesis2all wrote:

I have not tried that yet. Would it have an adverse effect on booting without using systemd?

Nope. I dont think so. But I use only native systemd.

misc · 2011-10-28 00:27:31

Looks like the default getty@.service isn't suited for ngetty, at least if that is used as intended (as a single daemon providing login sessions for all given tty).

crabman · 2011-10-31 09:54:07

I noticed an error in the ntpd.service file. The environment path should be

EnvironmentFile=/etc/conf.d/ntp-client.conf

instead of /etc/conf.d/ntp.conf

crabman

Last edited by crabman (2011-10-31 10:13:08)

t-8ch · 2011-10-31 11:45:49

booting [community] systemd-37 and [testing] cryptsetup-1.4.0 leads to following error:

Oct 31 11:15:22 laptop [   29.318759] systemd-cryptsetup[677]: /lib/systemd/systemd-cryptsetup: error while loading shared libraries: libcryptsetup.so.1: cannot open shared object file: No such file or directory
Oct 31 11:15:22 laptop [   29.319085] systemd[1]: cryptsetup@home.service: main process exited, code=exited, status=127
Oct 31 11:15:22 laptop [   29.336899] systemd[1]: Job cryptsetup.target/start failed with result 'dependency'.
Oct 31 11:15:22 laptop [   29.336909] systemd[1]: Job home.mount/start failed with result 'dependency'.
Oct 31 11:15:22 laptop [   29.336915] systemd[1]: Job dev-mapper-home.device/start failed with result 'dependency'.
Oct 31 11:15:22 laptop [   29.336920] systemd[1]: Unit cryptsetup@home.service entered failed state.

rebuilding systemd from abs fixes this.
So care if you have an ecrypted (non-root) device and are using [testing].

falconindy: Should I also file a flyspray bug against systemd?

t-8ch · 2011-10-31 11:50:09

and cronie.service should list:

EnvironmentFile=/etc/conf.d/crond

falconindy · 2011-10-31 12:54:04

t-8ch wrote:

booting [community] systemd-37 and [testing] cryptsetup-1.4.0 leads to following error:

Oct 31 11:15:22 laptop [   29.318759] systemd-cryptsetup[677]: /lib/systemd/systemd-cryptsetup: error while loading shared libraries: libcryptsetup.so.1: cannot open shared object file: No such file or directory
Oct 31 11:15:22 laptop [   29.319085] systemd[1]: cryptsetup@home.service: main process exited, code=exited, status=127
Oct 31 11:15:22 laptop [   29.336899] systemd[1]: Job cryptsetup.target/start failed with result 'dependency'.
Oct 31 11:15:22 laptop [   29.336909] systemd[1]: Job home.mount/start failed with result 'dependency'.
Oct 31 11:15:22 laptop [   29.336915] systemd[1]: Job dev-mapper-home.device/start failed with result 'dependency'.
Oct 31 11:15:22 laptop [   29.336920] systemd[1]: Unit cryptsetup@home.service entered failed state.

rebuilding systemd from abs fixes this.
So care if you have an ecrypted (non-root) device and are using [testing].

falconindy: Should I also file a flyspray bug against systemd?

Nope. I'll have a rebuild in community-testing shortly.

Glaucous · 2011-11-05 16:56:04

crabman wrote:

I noticed an error in the ntpd.service file. The environment path should be
EnvironmentFile=/etc/conf.d/ntp-client.conf
instead of /etc/conf.d/ntp.conf
crabman

https://github.com/falconindy/systemd-a … ts/pull/91

Was probably added to extra a bit too early.
testing/ntp works properly.

crabman · 2011-11-07 12:15:55

Glaucous wrote:

crabman wrote:
I noticed an error in the ntpd.service file. The environment path should be
EnvironmentFile=/etc/conf.d/ntp-client.conf
instead of /etc/conf.d/ntp.conf
crabman
https://github.com/falconindy/systemd-a … ts/pull/91
Was probably added to extra a bit too early.
testing/ntp works properly.

True, since ntp-4.2.6.p4-2 is in extra, the path was changed to /etc/conf.d/ntpd.conf and the unit file works again.

vostok4 · 2011-11-11 02:21:08

Is there any way to persist NETWORK_PERSIST="yes" in systemd? I have my rootfs on NFS, system boots with no problem, but I cannot shut down or reboot (I just get a black screen and the system is crashed in a weird state). I think that something is preventing the halt or reset, but I have no idea how to diagnose what that is.

tomegun · 2011-11-11 04:33:58

vostok4 wrote:

Is there any way to persist NETWORK_PERSIST="yes" in systemd? I have my rootfs on NFS, system boots with no problem, but I cannot shut down or reboot (I just get a black screen and the system is crashed in a weird state). I think that something is preventing the halt or reset, but I have no idea how to diagnose what that is.

My guess is that we should look at how dracut deals with this and see if there is anything we could do to make NETWORK_PERSIST obsolete. As far as I know there is no support for anything like NETWORK_PERSIS in fedora, but I might be wrong...

regitator · 2011-11-13 09:56:22

Quick question: how do you get into X if you use a lightweight window manager like awesome? Running a full-blown DM seems like overkill. In the Gentoo Wiki i found a "startx.service" but that won't work.

tomegun · 2011-11-13 10:02:21

regitator wrote:

Quick question: how do you get into X if you use a lightweight window manager like awesome? Running a full-blown DM seems like overkill. In the Gentoo Wiki i found a "startx.service" but that won't work.

How do you do it now? I assume a line in inittab? Could you paste the line you have and I'll make an example out of it.

lymphatik · 2011-11-13 14:00:43

regitator wrote:

Quick question: how do you get into X if you use a lightweight window manager like awesome? Running a full-blown DM seems like overkill. In the Gentoo Wiki i found a "startx.service" but that won't work.

Name the following xinit@.service then enable this service.

[Unit]
Description=Direct login to X for user %I
Requires=dev-tty2.device
After=dev-tty2.device systemd-user-sessions.service

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

[Install]
WantedBy=graphical.target

you might want to change zsh to bash if required

Last edited by lymphatik (2011-11-13 14:01:15)

Awebb · 2011-11-13 14:06:38

Interesting… It's one and a half year after I first saw this thread. Systemd seems to be a major player in the linux init league now. This Lennart Poettering guy is amazing. Pulseaudio, avahi, systemd - everybody hates it, everybody uses it, but nobody would admit that. So, as a so called professional, I'll have to learn how to use it, better now than next year.

Zom · 2011-11-13 14:53:22

Awebb wrote:

Interesting… It's one and a half year after I first saw this thread. Systemd seems to be a major player in the linux init league now. This Lennart Poettering guy is amazing. Pulseaudio, avahi, systemd - everybody hates it, everybody uses it, but nobody would admit that. So, as a so called professional, I'll have to learn how to use it, better now than next year.

Far from everybody hate them, but it's mostly the complaints that are heard.

archreg · 2011-11-13 16:02:12

lymphatik wrote:

Name the following xinit@.service then enable this service.

[Unit]
Description=Direct login to X for user %I
Requires=dev-tty2.device
After=dev-tty2.device systemd-user-sessions.service

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

[Install]
WantedBy=graphical.target

you might want to change zsh to bash if required

What is your output of ck-list-sessions when logging in like that?

Awebb · 2011-11-13 17:09:48

Zom wrote:

Far from everybody hate them, but it's mostly the complaints that are heard.

Yeah, the usual thing. The haters raise their voice, but the moment somebody says he likes something, he's either a hippie, a fanboy or somebody with unexpected sexual preferences. I have no particular opinion: I have yet to try systemd more than half an hour, avahi is somewhat cool and pulseaudio, well, is standard on so many distros, it can't be THAT broken anymore.

Zom · 2011-11-13 17:59:52

Awebb wrote:

Yeah, the usual thing. The haters raise their voice, but the moment somebody says he likes something, he's either a hippie, a fanboy or somebody with unexpected sexual preferences. I have no particular opinion: I have yet to try systemd more than half an hour, avahi is somewhat cool and pulseaudio, well, is standard on so many distros, it can't be THAT broken anymore.

Personally, I think pulseaudio was implemented a bit too soon, and that might've given people somewhat of a sour taste, but yeah.

Anyhow, I've used systemd now for more than 8 months. It's been working quite nice all this time as far as I can tell. So yeah, give it another go and see what you think. Make sure it lasts more than half an hour though.

lymphatik · 2011-11-14 16:58:45

archreg wrote:

lymphatik wrote:
Name the following xinit@.service then enable this service.
[Unit]
Description=Direct login to X for user %I
Requires=dev-tty2.device
After=dev-tty2.device systemd-user-sessions.service

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

[Install]
WantedBy=graphical.target
you might want to change zsh to bash if required
What is your output of ck-list-sessions when logging in like that?

I do not have console kit, so sorry about that

Vamp898 · 2011-11-15 18:32:31

I didnt understand the Vala Problem.

Glib is in every base system, if you like it or not.

And you can use Vala to output plain C and re-package this plain C. So you dont need any Vala on any system.

So you have Plain C Code which most dependency is glib (which is on nearly every base system anyway) but its developed much faster in Vala than than in plain C (for this developer)

So we have the advantage that we get Plain C Code and the developer have the advantage that he does not have to write C Code and so develops faster.

I think its a good idea at all.

Last edited by Vamp898 (2011-11-15 18:33:58)

PReP · 2011-11-16 22:48:39

No burning hate on systemd, but i hope the usual archlinux initscripts keep being the default.

I lack nothing as it is and has been for a great while, and as long as that is the case
- i personally see no need for systemd, or pulseaudio, or avahi.
neither do i udiskie or ck-launch-session, or mysql-like databases.

Things boot fast, audio works with alsa and my applications,
and my network is configured and working as it is.
Openbox works fine, Kde does to, though it is a bit more stuff to scale down there.

Perhaps in the future, i will be forced to look into systemd, just because,
but for now i keep reading this thread now and then.

Arch Linux

#1001 2011-10-23 17:37:51

Re: systemd: Yet Another Init Replacement

#1002 2011-10-23 17:53:39

Re: systemd: Yet Another Init Replacement

#1003 2011-10-24 07:17:58

Re: systemd: Yet Another Init Replacement

#1004 2011-10-25 13:48:44

Re: systemd: Yet Another Init Replacement

#1005 2011-10-25 18:13:54

Re: systemd: Yet Another Init Replacement

#1006 2011-10-28 00:27:31

Re: systemd: Yet Another Init Replacement

#1007 2011-10-31 09:54:07

Re: systemd: Yet Another Init Replacement

#1008 2011-10-31 11:45:49

Re: systemd: Yet Another Init Replacement

#1009 2011-10-31 11:50:09

Re: systemd: Yet Another Init Replacement

#1010 2011-10-31 12:54:04

Re: systemd: Yet Another Init Replacement

#1011 2011-11-05 16:56:04

Re: systemd: Yet Another Init Replacement

#1012 2011-11-07 12:15:55

Re: systemd: Yet Another Init Replacement

#1013 2011-11-11 02:21:08

Re: systemd: Yet Another Init Replacement

#1014 2011-11-11 04:33:58

Re: systemd: Yet Another Init Replacement

#1015 2011-11-13 09:56:22

Re: systemd: Yet Another Init Replacement

#1016 2011-11-13 10:02:21

Re: systemd: Yet Another Init Replacement

#1017 2011-11-13 14:00:43

Re: systemd: Yet Another Init Replacement

#1018 2011-11-13 14:06:38

Re: systemd: Yet Another Init Replacement

#1019 2011-11-13 14:53:22

Re: systemd: Yet Another Init Replacement

#1020 2011-11-13 16:02:12

Re: systemd: Yet Another Init Replacement

#1021 2011-11-13 17:09:48

Re: systemd: Yet Another Init Replacement

#1022 2011-11-13 17:59:52

Re: systemd: Yet Another Init Replacement

#1023 2011-11-14 16:58:45

Re: systemd: Yet Another Init Replacement

#1024 2011-11-15 18:32:31

Re: systemd: Yet Another Init Replacement

#1025 2011-11-16 22:48:39

Re: systemd: Yet Another Init Replacement

Board footer