systemd: Yet Another Init Replacement

tomegun · 2012-01-13 21:37:10

mwknowles92 wrote:

I've been getting some slightly annoying messages while booting. It doesn't cause any problems, but I still wonder why these errors are showing up.
[ 1.771982] EXT4-fs (sda5): couldn't mount as ext2 due to feature incompatibilities
[ 1.727911] EXT4-fs (sda5): couldn't mount as ext3 due to feature incompatibilities
I haven't been able to find anything on google yet, any ideas what's going on?

Are you using mkinitcpio? I have seen this kind of message when not using an initramfs. AFAIU the kernel (unlike 'mount') will try ext2, ext3 and then ext4 until it finds one that works. This is not a bug I think and nothing to worry about.

This is based on very vague memories, so I might be entirely wrong...

mwknowles92 · 2012-01-14 03:18:20

No I'm not using mkinitcpio and I'm not using initramfs. What's interesting is that I didn't even compile ext2 or ext3 into the kernel or make them as modules. Weird.

falconindy · 2012-01-14 03:34:47

The ext4 module has backwards support for mounting ext2 and ext3 FSs for some time now.

swiftgeek · 2012-01-14 07:35:35

Somebody should add to ArchWiki info about systemd failing to set proper locale when "keymap" hook is enabled in initramfs
I wasted half of the day for that

Also it would be good to provide some universal template of unit file for etc/rc.d/* daemons
BTW. I created something for pacman-cage http://pastebin.com/gYGNsrb9

tomegun · 2012-01-14 09:44:12

swiftgeek wrote:

Somebody should add to ArchWiki info about systemd failing to set proper locale when "keymap" hook is enabled in initramfs
I wasted half of the day for that

This sounds like a bug. We should try to fix it rather than document it...

swiftgeek wrote:

Also it would be good to provide some universal template of unit file for etc/rc.d/* daemons
BTW. I created something for pacman-cage http://pastebin.com/gYGNsrb9

systemd-initscripts will put some in /run/systemd/generator/. This needs an update though...

BasT · 2012-01-15 16:43:40

On arch-dev-public it was suggested to put

-session        optional        pam_systemd.so

in /etc/pam.d/kde*
(arch wiki suggests the same except without the "-")

I checked my own configuration which apparently I set to

session		required	pam_systemd.so

although I cannot remember my reasons for doing so. Maybe that's just what I read elsewhere?

What exactly is the difference between these settings?

Edit: meant to paste the systemd line, not the pam_limits one.

Last edited by BasT (2012-01-15 17:22:06)

tomegun · 2012-01-15 16:57:37

BasT wrote:

-session        optional        pam_systemd.so
session		required	pam_limits.so
What exactly is the difference between these settings?

They are very different, have a look at the man pages:

http://linux.die.net/man/8/pam_limits

http://0pointer.de/public/systemd-man/pam_systemd.html

The '-' in front of 'session' means that in case systemd is not installed the error will just be ignored (so it is safe to set it to this by default).

BasT · 2012-01-15 17:19:15

Oops i guess i copied the wrong line.

What I have is:

-session	required	pam_systemd.so

Maki · 2012-01-18 08:20:01

Sorry for being a noob, but how do i run a nfs server via systemd?, there are no native files and it only works when i enable arch-daemons. Nfsd.service starts but gives access denied on the client.

falconindy · 2012-01-18 16:51:06

There are native files for nfs in the units package.

rpc-idmapd.service
rpc-mountd.service
rpcbind.service

And add nfsd to a conf file in /etc/modules-load.d

Maki · 2012-01-18 21:42:12

Ha, thanks a lot, but now all i get is mount.nfs4: access denied by server , works just fine via rc.conf

tomegun · 2012-01-18 22:07:22

I believe fedora has some unit files that we might want to steal for nfs servers. I don't have a server to test on though so have not been pushing it....

fsckd · 2012-01-18 22:09:28

Hi Maki, your post was accidentally sent to the moderators via report. Don't worry, happens to the best of us, all the time.

Maki wrote:

I noticed that there was a rpcbind service but it didn't cross my mind that the other daemons will have a different name , thanks a lot, will add them to the wiki asap

Time, about one hour before this post.

Maki · 2012-01-18 22:09:54

 var-lib-nfs-rpc_pipefs.mount   loaded failed failed        RPC pipe filesystem

this is the only failed unit in systemctl, google shows nothing, btw how can i search this topic only?

@fsckd no problem, i we all click Report by mistake sometimes , tnx for letting me know, i got confused when i saw a post missing

Last edited by Maki (2012-01-18 22:32:01)

Agamemnon · 2012-01-20 15:54:28

Hi all!

I have a problem with the last update of gnome-shell. The color manager is completly broken and xsession-errors show this:

(gnome-control-center:865): color-cc-panel-WARNING **: failed to connect to colord: Erreur lors de l'appel de StartServiceByName pour org.freedesktop.ColorManager : GDBus.Error:org.freedesktop.systemd1.LoadFailed: Unit colord.service failed to load: No such file or directory. See system logs and 'systemctl status colord.service' for details.

Apparently, it seems to be systemd related. Before, I had no problem with it... Any idea?

tomegun · 2012-01-20 16:01:46

Agamemnon wrote:

Hi all!
I have a problem with the last update of gnome-shell. The color manager is completly broken and xsession-errors show this:
(gnome-control-center:865): color-cc-panel-WARNING **: failed to connect to colord: Erreur lors de l'appel de StartServiceByName pour org.freedesktop.ColorManager : GDBus.Error:org.freedesktop.systemd1.LoadFailed: Unit colord.service failed to load: No such file or directory. See system logs and 'systemctl status colord.service' for details.
Apparently, it seems to be systemd related. Before, I had no problem with it... Any idea?

This is a colord packaging bug. The dbus service file points to the systemd service file, but the systemd service file is not installed. I'll fix our packager.

christopher72 · 2012-01-24 22:28:36

I'm having this curious shutdown problem which I can't figure out, because by trying to find out what causes it, I basically solve the problem. :-)
I've got a cifs share in fstab:

//192.168.178.25/Backups /mnt/storage   cifs    _netdev,noauto,suid,users,comment=systemd.automount 0   0

Mounts just fine when I access it.
However, shutdown/reboot fails. I just get a blinking cursor at the top left of the screen. If I comment the line in fstab, shutdown/reboot works fine, so I guess it has something to do with umounting the cifs mount. However, if I try to make systemd more verbose by removing the "quiet" option from the kernel command line, shutdown/reboot is working again, even with the fstab entry enabled.

Btw, shutdown/reboot seems to be a very delicate thing in systemd... it should be more robust. But that's just an opinion.

Any ideas or workarounds as to how to solve this?

Last edited by christopher72 (2012-01-24 22:31:38)

christopher72 · 2012-01-25 12:50:50

Well, I "solved" my shutdown problem by using nfs instead of cifs... which is more of a linux way of doing things anyway.

TheRealSoup · 2012-01-27 13:23:07

Sorry if anyone has covered this already, but I'm struggling to boot my laptop using mdraid with systemd.

I have 5 mdraid arrays, but systemd only brings up /root. Once I get the emergency message after failing dependencies of /home and others, /proc/mdstat reports only the one array.

Options I use in fstab for all of these arrays is simply noatime and the partition type is fd as one would expect.

I'm not sure what I'm missing. This is the first time I'm trying systemd. Has anyone encountered something like this before?

ImNtReal · 2012-01-27 14:45:08

TheRealSoup wrote:

I have 5 mdraid arrays, but systemd only brings up /root. Once I get the emergency message after failing dependencies of /home and others, /proc/mdstat reports only the one array.

Is /usr one of them? If so, have you read this: http://www.freedesktop.org/wiki/Softwar … -is-broken

=-Jameson

TheRealSoup · 2012-01-27 15:03:57

Nope, /usr lives on / and not a separate partition.

Just seems as if how systemd handles &/ discovers rootfs if different to everything else if finds.

tomegun · 2012-01-27 15:06:12

TheRealSoup wrote:

Nope, /usr lives on / and not a separate partition.
Just seems as if how systemd handles &/ discovers rootfs if different to everything else if finds.

The rootfs is not mounted by systemd (how could it? systemd lives there). It is done from the initrd, so that explains why the rootfs works.

TheRealSoup · 2012-01-28 01:04:37

Sorry, don't know how, but I seem to have been missing a few systemd unit files.

Maki · 2012-01-28 11:34:13

Fiddling again with systemd and my nfs problem, i started modprobed nfsd, started rpcbind idmapd and nfsd with systemd but still couldn't mount it getting permission denied, i reloaded the config with exportfs -r, and now it works, this seems wierd since i haven't touched the config these days, could be that the units don't set everything properly ?

Update: i found an exportfs service, will try that after a reboot

Last edited by Maki (2012-01-28 11:42:51)

ego.abyssi · 2012-01-29 19:10:23

I've been using systemd for a while now, through a few different installations of Arch on different laptops. Here are some current issues that I've been experiencing, and some (perhaps incomplete) solutions:

* * *

Problem: /etc/vconsole is properly configured but systemd-vconsole-setup.service fails to set the console font

Solution: systemd restart systemd-vconsole-setup.service

* * *

Problem: /lib/systemd/system/wpa_supplicant.service doesn't associate with my wireless card, wlan0

Solution: create /lib/systemd/system/wpa_supplicant\@.service:

[Unit]
Description=wpa_supplicant on %I
Wants=network.target
Before=dhcpcd@.service

[Service]
Type=dbus
BusName=fi.epitest.hostap.WPASupplicant
ExecStart=/usr/sbin/wpa_supplicant -i %I -c /etc/wpa_supplicant/wpa_supplicant.conf -u

Then,

ln -s /lib/systemd/system/wpa_supplicant\@.service /etc/systemd/system/multi-user.target.wants/wpa_supplicant\@wlan0.service

* * *

Problem: /etc/systemd/system/multi-user.target.wants/wpa_supplicant\@wlan0.service and /etc/systemd/system/multi-user.target.wants/dhcpcd\@wlan0.service fail on boot

Solution: systemd restart wpa_supplicant\@wlan0.service && systemd restart dhcpcd\@wlan0.service

* * *

Other than the networking specifics, the problems I am experiencing above seem to stem from devices not yet existing when these services run. I'm assuming there is a better solution than restarting the services, perhaps along the lines of adding appropriate Wants, Requires, Before, or After clauses to the service specifications. I'll be damned if I quite know how to arrange which clauses into which service specifications to alleviate the problems.

Arch Linux

#1101 2012-01-13 21:37:10

Re: systemd: Yet Another Init Replacement

#1102 2012-01-14 03:18:20

Re: systemd: Yet Another Init Replacement

#1103 2012-01-14 03:34:47

Re: systemd: Yet Another Init Replacement

#1104 2012-01-14 07:35:35

Re: systemd: Yet Another Init Replacement

#1105 2012-01-14 09:44:12

Re: systemd: Yet Another Init Replacement

#1106 2012-01-15 16:43:40

Re: systemd: Yet Another Init Replacement

#1107 2012-01-15 16:57:37

Re: systemd: Yet Another Init Replacement

#1108 2012-01-15 17:19:15

Re: systemd: Yet Another Init Replacement

#1109 2012-01-18 08:20:01

Re: systemd: Yet Another Init Replacement

#1110 2012-01-18 16:51:06

Re: systemd: Yet Another Init Replacement

#1111 2012-01-18 21:42:12

Re: systemd: Yet Another Init Replacement

#1112 2012-01-18 22:07:22

Re: systemd: Yet Another Init Replacement

#1113 2012-01-18 22:09:28

Re: systemd: Yet Another Init Replacement

#1114 2012-01-18 22:09:54

Re: systemd: Yet Another Init Replacement

#1115 2012-01-20 15:54:28

Re: systemd: Yet Another Init Replacement

#1116 2012-01-20 16:01:46

Re: systemd: Yet Another Init Replacement

#1117 2012-01-24 22:28:36

Re: systemd: Yet Another Init Replacement

#1118 2012-01-25 12:50:50

Re: systemd: Yet Another Init Replacement

#1119 2012-01-27 13:23:07

Re: systemd: Yet Another Init Replacement

#1120 2012-01-27 14:45:08

Re: systemd: Yet Another Init Replacement

#1121 2012-01-27 15:03:57

Re: systemd: Yet Another Init Replacement

#1122 2012-01-27 15:06:12

Re: systemd: Yet Another Init Replacement

#1123 2012-01-28 01:04:37

Re: systemd: Yet Another Init Replacement

#1124 2012-01-28 11:34:13

Re: systemd: Yet Another Init Replacement

#1125 2012-01-29 19:10:23

Re: systemd: Yet Another Init Replacement

Board footer