Opensuse 12.1 hangs at bootup for 5 minutes after systemd-fsck

English Install/Boot/Login
#1

Hi,

I recently upgraded from 11.4 to 12.1 and things were working fine,
but now after a job lot of updates (I think there were something like 168 of them,
and I’m afraid I don’t know what they all were), the system hangs for about 5 mins
at boot time after running systemd-fsck.

It does eventually boot into KDE and everything else is fine, it’s just annoying that
I have to wait for this process.

It briefly shows a runlevel 3 login screen before the KDE login,
which I hadn’t noticed before.

here’s a shot of my boot screen where it hangs:

http://i298.photobucket.com/albums/mm264/moondustjunky/systemd-fsck.jpg

the root partition is on /dev/sdb1.
/dev/sda1, /dev/sdb3 and /dev/sdb4 are all data partitions.

I had to reinstall the proprietary nvidia driver for my GeForce FX5700,
but as I say I do eventually get into KDE alright.

Thanks

#2

5 minutes is the default timeout if you’re booting in systemd. You have to find out which service fails and disable it (or fix it).

#3

On 02/09/2012 06:16 PM, Wild Goose wrote:
> here’s a shot of my boot screen where it hangs:
>
> [image:
> http://i298.photobucket.com/albums/mm264/moondustjunky/systemd-fsck.jpg]

that is not a ‘hang’…that is a process in progress…looks like an
fsck of sdb3, which guess may be the /home of your installed openSUSE on
your second disk…

the question is: why is it running fsck! how did you shut down? was it a
normal shutdown by clicking “Leave” and then “Turn off computer” or
“Restart Computer” (i do not know which desktop environment you are
using [since you didn’t say] so those words may not be correct…)

or, are you shutting down from a command line? if so, what
command/switches are you using?

or, are you hibernating/sleeping…

and, you might as well tell us how you “upgraded from 11.4 to 12.1”
which of these supported guides did you follow step by step:
http://tinyurl.com/35p966c
http://tinyurl.com/6kvoflv

or, did you do it some other way…

–
DD http://tinyurl.com/DD-Caveat
Read what Distro Watch writes: http://tinyurl.com/SUSEonDW

#4

I am shutting down and restarting KDE from the graphical menu,
“leave” “turn off” etc.

I’m not using hibernate or sleep

I upgraded by downloading the 12.1 iso image and booting from the dvd,
choosing the upgrade option.

I have enabled verbose booting, and the end of dmesg gives:

   10.405222] 8139too 0000:00:13.0: eth0: link up, 100Mbps, full-duplex, lpa 0x45E1
  313.220854] agpgart-amd64 0000:00:00.0: AGP 3.0 bridge
  313.220870] agpgart-amd64 0000:00:00.0: putting AGP V3 device into 8x mode
  313.220918] nvidia 0000:01:00.0: putting AGP V3 device into 8x mode
  340.428608] EXT4-fs (sdb1): re-mounted. Opts: acl,user_xattr,commit=0
  340.450125] EXT4-fs (sdb3): re-mounted. Opts: commit=0
  340.464639] EXT4-fs (sdb4): re-mounted. Opts: commit=0
  340.480133] EXT4-fs (sda1): re-mounted. Opts: acl,user_xattr,commit=0
  392.165621] fuse init (API version 7.17)
  • so now it seems to pause after eth0 comes up for some reason?
#5

On 02/09/2012 08:06 PM, Wild Goose wrote:
> - so now it seems to pause after eth0 comes up for some reason?

pay attention to what please-try-again posted…he know far more about
12.1 and the new systemd…

you might try booting and at the first green screen press F5 and select
“system V” like here:
http://dl.dropbox.com/u/10573557/12.1_Misc/12.1_F5_sysV.jpg

then run a little while and reboot…if the ‘hang’ is not there then you
are pretty sure it has something to do with systemd, an if you follow
the advice in #2 above you will probably smile again…

–
DD http://tinyurl.com/DD-Caveat
Read what Distro Watch writes: http://tinyurl.com/SUSEonDW

#6

Not a solution, but a workaround: Yast - Software - Softwaremanagement. Search for sysvinit-init and install it. It will tell you it has to uninstall systemd-sysvinit. Your booting will be done the old way using sysvinit, which will remove the 5 minute time out during boot.

#7

In principle, the time out is not a bad thing. systemd will hang for 5 minutes where sysinit might hang forever.

#8

On 2012-02-09 21:36, Knurpht wrote:
>
> Not a solution, but a workaround: Yast - Software - Softwaremanagement.
> Search for sysvinit-init and install it. It will tell you it has to
> uninstall systemd-sysvinit. Your booting will be done the old way using
> sysvinit, which will remove the 5 minute time out during boot.

That should not be done before testing it by pressing F5 at boot and
choosing systemv.

–
Cheers / Saludos,

Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)

#9

Is there a second nic on this system? And are you using samba, especially the nmb service?

#10

On 2012-02-09 18:16, Wild Goose wrote:

> about 5 mins
> at boot time after running systemd-fsck.

In a terminal, as root, run:


systemd-analyze blame

and find out where the time is spent. Alternatively, you can also do


systemd-analyze plot > p.svg

Then you can use “display p.svg” to see the graphics, or use the tool you
prefer. The graphics shows where time is used.

–
Cheers / Saludos,

Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)

#11

Waow !!!

#12

I only have one nic, but yes samba (I’m using smb and nmb) does seem to be the problem.

when I disabled it at boot time it booted up fine.

also that p.svg graph mentioned before gave a huge spike for “network.service”

any ideas why samba would be doing this?

#13

On 2012-02-09 22:46, please try again wrote:
>> prefer. The graphics shows where time is used.
>> >
>> >
> Waow !!!

Yes, it is a cute thing, isin’t? :slight_smile:

–
Cheers / Saludos,

Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)

#14

It’s a bug. You can try my patch here: http://forums.opensuse.org/english/get-technical-help-here/install-boot-login/469461-bug-nmb-systemd-more-than-1-ethernet-card.html#post2436984. But I have seen this behaviour only on machines with more than one nic - actually there isn’t any machine with a single nic here.

#15

Yes, it’s a nice one!
I added a bootchart option (that creates a bootchart entry) in updategrub: http://forums.opensuse.org/english/other-forums/development/programming-scripting/458234-updategrub-opensuse-legacy-grub-not-update-grub-2.html#post2421328. But this is lot faster.

#16

The bug (see post #14 above) is reportedly fixed, but does not seem to have made it to an update yet.

Note that, in my experience, the delay is 5 min if it is a static IP address, but only 15 secs(or so) with DHCP, if that will work for you.

#17

I know that it is reported as fixed, but last time I checked (can not find the page in bugzilla anymore), the fix didn’t work. My patch works (at least for me). You have to use “restart” instead of “start” to start the nmb service.