You are not logged in.

#301 2011-03-27 08:58:20

litemotiv
Forum Fellow
Registered: 2008-08-01
Posts: 5,026

Re: systemd: Yet Another Init Replacement

falconindy wrote:

Nothing should be printed to stdout unless the unit isn't installed, as you discovered.

$ systemctl is-enabled lvm-activate.service && echo "lvm is enabled" || echo "lvm is not enabled"

So if i understand correctly is-enabled actually means is-installed? Ok that makes things more clear, maybe the parameter would be better renamed then?


ᶘ ᵒᴥᵒᶅ

Offline

#302 2011-03-27 13:23:42

falconindy
Developer
From: New York, USA
Registered: 2009-10-22
Posts: 4,111
Website

Re: systemd: Yet Another Init Replacement

Right. It only applies to units with an install section and it returns true when the install section's directives are already fulfilled.

Naming conventions are a scary topic on the systemd-devel list...

Offline

#303 2011-03-28 20:14:16

frankieboy
Member
Registered: 2009-04-12
Posts: 65

Re: systemd: Yet Another Init Replacement

tomegun wrote:
frankieboy wrote:

I installed systemd-git, initscripts-systemd-git and systemd-arch-units-git, disabled all md*, dmraid services, no luck. Enabled mdadm.service, still no luck. Commented out all partitions but the root, I get a login prompt, but unable to login (don't know why), enabled all partitions on the ssd (/dev/mapper/sil_) in fstab, no luck.

Hmmm... we are basically shooting in the dark here, with no form of error message. Are you using the "verbose" kernel parameter? This should give you some more output, so maybe it will tell you where the problem is (when you get stuck at the login prompt, I'm assuming you are trying to log in as root, right?).

-t

I boot it into single mode using systemd from community repo, gave the root password, and typed /sbin/mdadm --assemble --scan, which resulted the following:
mdadm: failed to set writable for /dev/md0: device or resource busy (and the same for md1, md2, md3)
/dev/md0 has 640 premissions and root.disk group.
Hope that helps.

Thx

Frankieboy

Offline

#304 2011-03-28 21:31:29

tomegun
Developer
From: France
Registered: 2010-05-28
Posts: 661

Re: systemd: Yet Another Init Replacement

@Frankieboy: could you be experiencing this <http://www.righteoushack.net/?p=197>? Maybe you could compare lsmod between systemd and sysvinit to see what might be using your md devices.

Offline

#305 2011-03-29 10:36:50

benneque
Member
Registered: 2011-03-19
Posts: 53

Re: systemd: Yet Another Init Replacement

If you look at cgit of systemd and plymouth , you can see that there's something going on!
Let's hope that with the next versions everything gets better & better smile
http://cgit.freedesktop.org/plymouth
http://cgit.freedesktop.org/systemd

Offline

#306 2011-03-29 14:39:46

eworm
Package Maintainer (PM)
From: Oberhausen, Germany
Registered: 2010-01-30
Posts: 107
Website

Re: systemd: Yet Another Init Replacement

I've installed systemd on a number of arch systems. It works fine on most of them (old Acer Notebook, recent Samsung Netbook), but has some problems on Thinkpad Notebooks (T400s and T410s). It basically works, but I have a boot delay of several minutes. I do see the messages about fsck (beside some intel gpu warnings at least on one of the systems), then the system is sitting there and doing nothing until the login manager manically appears. Any ideas?
The systems boots as expected (and far faster) with sysvinit.


ArchLinux - make it simple & lightweight

Offline

#307 2011-03-29 17:18:32

tomegun
Developer
From: France
Registered: 2010-05-28
Posts: 661

Re: systemd: Yet Another Init Replacement

@eworm: Probably a service is failing, so it waits the standard 3 minutes timeout before ignoring the service and continuing.
You can figure out which one by doing "systemctl" or "systemctl --all" after you have successfully booted up. If that does not give a clear answer, then try booting with the "verbose" kernel parameter.

Offline

#308 2011-03-30 00:34:52

ngoonee
Forum Fellow
From: Between Thailand and Singapore
Registered: 2009-03-17
Posts: 7,358

Re: systemd: Yet Another Init Replacement

I've been thinking to try out autofs, a quick google of the phrase - "systemd" autofs - reveals that at least in suse autofs seems to be supported by systemd. However I don't find any autofs-related .service files or similar. Should I try to write one (/etc/rc.d/autofs doesn't seem all that complicated) or would it just be duplication?


Allan-Volunteer on the (topic being discussed) mailn lists. You never get the people who matters attention on the forums.
jasonwryan-Installing Arch is a measure of your literacy. Maintaining Arch is a measure of your diligence. Contributing to Arch is a measure of your competence.
Griemak-Bleeding edge, not bleeding flat. Edge denotes falls will occur from time to time. Bring your own parachute.

Offline

#309 2011-03-30 09:57:38

tomegun
Developer
From: France
Registered: 2010-05-28
Posts: 661

Re: systemd: Yet Another Init Replacement

ngoonee wrote:

I've been thinking to try out autofs, a quick google of the phrase - "systemd" autofs - reveals that at least in suse autofs seems to be supported by systemd. However I don't find any autofs-related .service files or similar. Should I try to write one (/etc/rc.d/autofs doesn't seem all that complicated) or would it just be duplication?

Depends on how you want to use autofs. Autofs is used by systemd internally so some feautres are alrady implemented. Just add "comment=systemd.automount" to your fs in fstab and it will be mounted on first access. I don't know much about the other feauters of autofes, but I think there are things that are not implemented by systemd, so probably a service file would be needed.

-t

Offline

#310 2011-03-30 10:13:08

Dieter@be
Forum Fellow
From: Belgium
Registered: 2006-11-05
Posts: 2,001
Website

Re: systemd: Yet Another Init Replacement

tomegun wrote:

Depends on how you want to use autofs. Autofs is used by systemd internally so some feautres are alrady implemented. Just add "comment=systemd.automount" to your fs in fstab and it will be mounted on first access. I don't know much about the other feauters of autofes, but I think there are things that are not implemented by systemd, so probably a service file would be needed.

one of the other nice things about autoFS is that it can autodiscover hosts and their shares, so:

$ # assuming /var/autofs is set to the "basedir" in the autofs master config
$ ls /var/autofs # this discovers all hosts automatically
hostname1
host2
$ ls /var/autofs/host2/ # discovers shares on host2
myshare
share2
$ cd /var/autofs/host2/share2 # mounted transparantly for you, you can start working here

< Daenyth> and he works prolifically
4 8 15 16 23 42

Offline

#311 2011-03-30 10:17:28

tomegun
Developer
From: France
Registered: 2010-05-28
Posts: 661

Re: systemd: Yet Another Init Replacement

Dieter@be wrote:

one of the other nice things about autoFS is that it can autodiscover hosts and their shares, so:

Nice! I think we would need a service file to get this.

Offline

#312 2011-03-30 10:38:36

ngoonee
Forum Fellow
From: Between Thailand and Singapore
Registered: 2009-03-17
Posts: 7,358

Re: systemd: Yet Another Init Replacement

tomegun wrote:
ngoonee wrote:

I've been thinking to try out autofs, a quick google of the phrase - "systemd" autofs - reveals that at least in suse autofs seems to be supported by systemd. However I don't find any autofs-related .service files or similar. Should I try to write one (/etc/rc.d/autofs doesn't seem all that complicated) or would it just be duplication?

Depends on how you want to use autofs. Autofs is used by systemd internally so some feautres are alrady implemented. Just add "comment=systemd.automount" to your fs in fstab and it will be mounted on first access. I don't know much about the other feauters of autofes, but I think there are things that are not implemented by systemd, so probably a service file would be needed.

-t

I'm mainly after auto-mounting of USB devices. Seems like that would need a service file?


Allan-Volunteer on the (topic being discussed) mailn lists. You never get the people who matters attention on the forums.
jasonwryan-Installing Arch is a measure of your literacy. Maintaining Arch is a measure of your diligence. Contributing to Arch is a measure of your competence.
Griemak-Bleeding edge, not bleeding flat. Edge denotes falls will occur from time to time. Bring your own parachute.

Offline

#313 2011-03-30 10:42:57

tomegun
Developer
From: France
Registered: 2010-05-28
Posts: 661

Re: systemd: Yet Another Init Replacement

ngoonee wrote:

I'm mainly after auto-mounting of USB devices. Seems like that would need a service file?

Yes. The inbuilt functionality is that it will mount a device when you first access the mountpoint (useful for seldomly accessed mounts like /boot, or for allowing boot to proceed even if a required mount (like /home) is not mounted yet).

Offline

#314 2011-03-30 10:44:23

ngoonee
Forum Fellow
From: Between Thailand and Singapore
Registered: 2009-03-17
Posts: 7,358

Re: systemd: Yet Another Init Replacement

tomegun wrote:
ngoonee wrote:

I'm mainly after auto-mounting of USB devices. Seems like that would need a service file?

Yes. The inbuilt functionality is that it will mount a device when you first access the mountpoint (useful for seldomly accessed mounts like /boot, or for allowing boot to proceed even if a required mount (like /home) is not mounted yet).

And for devices which aren't in fstab?

autofs' automounter functions as a replacement for nautilus' auto-mount, mounting plugged-in devices as soon as they're detected to /media or wherever you want.


Allan-Volunteer on the (topic being discussed) mailn lists. You never get the people who matters attention on the forums.
jasonwryan-Installing Arch is a measure of your literacy. Maintaining Arch is a measure of your diligence. Contributing to Arch is a measure of your competence.
Griemak-Bleeding edge, not bleeding flat. Edge denotes falls will occur from time to time. Bring your own parachute.

Offline

#315 2011-03-30 10:46:34

tomegun
Developer
From: France
Registered: 2010-05-28
Posts: 661

Re: systemd: Yet Another Init Replacement

ngoonee wrote:

And for devices which aren't in fstab?

autofs' automounter functions as a replacement for nautilus' auto-mount, mounting plugged-in devices as soon as they're detected to /media or wherever you want.

Then you'd need a service file.

Offline

#316 2011-03-30 11:51:59

antantros
Member
Registered: 2011-03-30
Posts: 1

Re: systemd: Yet Another Init Replacement

Interesting project. Just`ve tryed it but have not so exciting results (faster booting).  Best results are 18 seconds from grub to kdm. With standart init - 20 seconds.  May be this is normal for notebook with usual sata drive? I did "systemctl enable" for all daemons listed in rc.conf (dcron.service, syslog-ng.service), enabled console-kit-daemon.service, kdm.service, systemd-readahead-collect.service and systemd-readahead-replay.service. In DAEMONS section I leave @net-profiles and @bluetooth because of I can not find *.service for them.  Did i do everything right? I use systemd, initscripts-systemd and systemd-arch-units  from community.

Last edited by antantros (2011-03-30 12:18:33)

Offline

#317 2011-03-30 12:02:04

Dieter@be
Forum Fellow
From: Belgium
Registered: 2006-11-05
Posts: 2,001
Website

Re: systemd: Yet Another Init Replacement

tomegun wrote:
Dieter@be wrote:

one of the other nice things about autoFS is that it can autodiscover hosts and their shares, so:

Nice! I think we would need a service file to get this.

beware though. autofs has been broken for a looooong time.
See https://bbs.archlinux.org/viewtopic.php?id=79513
https://bugs.archlinux.org/task/14474

ultimately they just moved into AUR because noone could fix the brokenness.

edit: hmm, i see it's in community now, I guess I could try it out again :roll:

Last edited by Dieter@be (2011-03-30 12:02:35)


< Daenyth> and he works prolifically
4 8 15 16 23 42

Offline

#318 2011-03-30 12:31:35

eworm
Package Maintainer (PM)
From: Oberhausen, Germany
Registered: 2010-01-30
Posts: 107
Website

Re: systemd: Yet Another Init Replacement

tomegun wrote:

@eworm: Probably a service is failing, so it waits the standard 3 minutes timeout before ignoring the service and continuing.
You can figure out which one by doing "systemctl" or "systemctl --all" after you have successfully booted up. If that does not give a clear answer, then try booting with the "verbose" kernel parameter.

Here is everything that failed or had errors:

# systemctl --all | egrep '(failed|error)'
systemd-...-plymouth.path error  inactive dead          systemd-ask-password-ply
apache.service            loaded failed   failed        Legacy unit for apache.s
auditd.service            error  inactive dead          auditd.service
mysql.service             loaded failed   failed        Legacy unit for mysql.se
systemd-...ymouth.service error  inactive dead          systemd-ask-password-ply
udev-settle.service       loaded failed   failed        udev Wait for Complete D

Where do apache.service and mysql.service come from? They fail because /etc/rc.d/{apache,mysql} do not exist. Everything is fine with httpd.service and mysqld.service though.

udev-settle.service is fine after restarting it manually. Before that status gives:

udev-settle.service - udev Wait for Complete Device Initialization
          Loaded: loaded (/lib/systemd/system/udev-settle.service)
          Active: failed since Wed, 30 Mar 2011 13:42:35 +0200; 38min ago
         Process: 992 ExecStart=/sbin/udevadm settle (code=exited, status=1/FAILURE)
          CGroup: name=systemd:/system/udev-settle.service

What are the other units for?

Booting with verbose gives a lot more output, but the relevant error message (I can see a red "failed" flash by) is shown for a very short time before agetty and slim come up.


ArchLinux - make it simple & lightweight

Offline

#319 2011-03-30 12:53:52

Zom
Member
From: Sweden
Registered: 2007-10-27
Posts: 430

Re: systemd: Yet Another Init Replacement

antantros wrote:

Interesting project. Just`ve tryed it but have not so exciting results (faster booting).  Best results are 18 seconds from grub to kdm. With standart init - 20 seconds.  May be this is normal for notebook with usual sata drive? I did "systemctl enable" for all daemons listed in rc.conf (dcron.service, syslog-ng.service), enabled console-kit-daemon.service, kdm.service, systemd-readahead-collect.service and systemd-readahead-replay.service. In DAEMONS section I leave @net-profiles and @bluetooth because of I can not find *.service for them.  Did i do everything right? I use systemd, initscripts-systemd and systemd-arch-units  from community.

I think readahead is enabled by default to be honest.

However, I don't think systemd runs the DAEMONS from rc.conf by default. Bluetooth also doesn't need to be started as a service, it'll start when you connect a bluetooth device to the computer (Like, say, a USB bluetooth dongle) IIRC.

Offline

#320 2011-03-30 13:08:36

eworm
Package Maintainer (PM)
From: Oberhausen, Germany
Registered: 2010-01-30
Posts: 107
Website

Re: systemd: Yet Another Init Replacement

eworm wrote:

Where do apache.service and mysql.service come from? They fail because /etc/rc.d/{apache,mysql} do not exist. Everything is fine with httpd.service and mysqld.service though.

Oops, I had apache and mysql in my DAEMONS array in /etc/rc.conf... This problem is solved.
Still... What about udev-settle.service and the others?

Last edited by eworm (2011-03-30 13:08:56)


ArchLinux - make it simple & lightweight

Offline

#321 2011-03-30 14:29:36

falconindy
Developer
From: New York, USA
Registered: 2009-10-22
Posts: 4,111
Website

Re: systemd: Yet Another Init Replacement

udev-settle is going away (or at least won't be enabled by default) with the next udev. If your hardware is working, I think you can just ignore that.

plymouth can be ignored, since its not installed. same with audit.

I really should just delete the audit service file since I'm not even compiling in support for it.

Offline

#322 2011-03-30 14:40:55

eworm
Package Maintainer (PM)
From: Oberhausen, Germany
Registered: 2010-01-30
Posts: 107
Website

Re: systemd: Yet Another Init Replacement

falconindy wrote:

udev-settle is going away (or at least won't be enabled by default) with the next udev. If your hardware is working, I think you can just ignore that.

Ignoring it is a stupid idea... It brings three minutes of boot delay. :-p
I added "--timeout=10" to /lib/systemd/system/udev-settle.service, resulting in a 10 seconds boot delay.

2011-03-30T16:19:06.193870+02:00 io kernel: udevadm[1025]: timeout waiting for udev queue
2011-03-30T16:19:06.193872+02:00 io kernel: systemd[1]: udev-settle.service: main process exited, code=exited, status=1
2011-03-30T16:19:06.193873+02:00 io kernel: systemd[1]: Unit udev-settle.service entered failed state.

I think I have to figure out what make my udev go wild...


ArchLinux - make it simple & lightweight

Offline

#323 2011-03-30 15:38:21

falconindy
Developer
From: New York, USA
Registered: 2009-10-22
Posts: 4,111
Website

Re: systemd: Yet Another Init Replacement

I suggest turning up udev logging in /etc/udev/udev.conf. Something's not right there.

Offline

#324 2011-03-30 16:23:33

eworm
Package Maintainer (PM)
From: Oberhausen, Germany
Registered: 2010-01-30
Posts: 107
Website

Re: systemd: Yet Another Init Replacement

I already did. That floods the logs with a lot of udev messages, but I could not find anything usable till now.

This is what udev prints when it exits:

udevadm settle - timeout of 10 seconds reached, the event queue contains:
  /sys/devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.4/1-1.4:1.0/bluetooth/hci0 (1601)
  /sys/devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.4/1-1.4:1.0/bluetooth/hci0/rfkill2 (1602)
  /sys/devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.4/1-1.4:1.0/bluetooth/hci0/rfkill2 (1603)
  /sys/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.8/2-1.8:1.0/0003:1B96:0001.0003 (1633)
  /sys/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.8/2-1.8:1.1/0003:1B96:0001.0004 (1634)
  /sys/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.8/2-1.8:1.0/0003:1B96:0001.0003/hidraw/hidraw2 (1666)
  /sys/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.8/2-1.8:1.1/0003:1B96:0001.0004/hidraw/hidraw3 (1727)

ArchLinux - make it simple & lightweight

Offline

#325 2011-03-30 17:40:45

vostok4
Member
Registered: 2010-12-16
Posts: 70

Re: systemd: Yet Another Init Replacement

Question about mounts. I have the following in my fstab:

/dev/sdb1 /mnt/hd ext3 noatime 0 0
/dev/sdc1 /mnt/sd reiserfs noatime 0 0

But these directories are not getting mounted at boot by systemd (or by... what? what should mount it?) What's the proper course of action here?

Also, should I put all my new .service files in /lib/systemd/system or /etc/systemd/system?

Offline

Board footer

Powered by FluxBB