Problem with clients booting from Ubuntu 10.0
Brought to you by:
steven_shiau
Menu
▾
▴
Description of problem:
Upgraded Ubuntu 8.04 to Ubuntu 10.04 (not a fresh install)
The clients fail to correctly start up leaving them at the shell login prompt.
There are a number of issues/errors in the start-up information:
init.orig: procps main process (process #) terminated with status 255
initct: event failed
init: drbl_client_boot main process (process #) terminated with status 1
Then there are the following messages:
"Rather than invoking init scripts through /etc/init.d, use the service (8) utility, e.g. service S18portmap start
start: unknown job: S18portmap"
The are a number of services which get reported like this:
s10acpid
s12dbus
s30gdm
s43portmap
The last error before it exits to the shell login prompt is:
init.orig: plymouth-stop pre-start process (process #) terminated with status 1
If you login via the login prompt and use su to start the gdm service it allows the user to login and use the client. All the services like portmap, dbus, acpid etc… all seem to be running.
How reproducible:
I'm hoping to do a fresh install of Ubuntu 10.04 and DRBL on our test Server to see if the issue persists.
Steps to Reproduce:
Actual results:
Expected results:
Some info about the DRBL environment (PLEASE DO NOT EDIT THEM!):
OS version: Ubuntu 10.04
Server arch: i686
Server CPU: Intel(R) Xeon(TM) CPU 3.00GHz
Server memory size: 2061148 kB
Server Kernel version: 2.6.32-22-generic
Installed DRBL-related packages: drbl-1.9.6-17 clonezilla-2.3.6-24 mkswap-uuid-0.1.1-1 drbl-partimage-0.6.8-1drbl drbl-ntfsprogs-2.0.0-4 partclone-0.2.9-1drbl drbl-chntpw-0.0.20040818-7 drbl-lzop-1.02-0.8drbl pigz-2.1.5-1drbl pbzip2-1.0.5-1ubuntu1 mkpxeinitrd-net-1.5-3 udpcast-20100130-1drbl drbl-etherboot-5.4.3-2 gpxe-1.0.0-1drbl freedos-1.0-14drbl
Client kernel version: 2.6.32-22-generic
Client kernel arch: i586
NICs with private IP address in server: eth0 eth1
Private IP address in server: 192.168.1.73 192.168.2.1
Total client no: 20
Client IP address: 192.168.2.20 192.168.2.21 192.168.2.22 192.168.2.23 192.168.2.24 192.168.2.25 192.168.2.26 192.168.2.27 192.168.2.28 192.168.2.29 192.168.2.30 192.168.2.31 192.168.2.32 192.168.2.33 192.168.2.34 192.168.2.35 192.168.2.36 192.168.2.37 192.168.2.38 192.168.2.39
===
"init.orig: procps main process (process #) terminated with status 255 initct: event failed init: drbl_client_boot main process (process #) terminated with status 1 Then there are the following messages: "Rather than invoking init scripts through /etc/init.d, use the service (8) utility, e.g. service S18portmap start start: unknown job: S18portmap" The are a number of services which get reported like this: s10acpid s12dbus s30gdm s43portmap " -> Basically you can ignore those warning messages. It's nothing to do with your gdm service.
How about if you run:
ls -alFh /etc/init/gdm.conf
on your DRBL server?
Does it exist?
Steven.
Hi Steven,
Yes gdm.conf exists
I've installed a fresh Ubuntu 10.04 and DRBL on our test Server and it all works. So the issue must be related to the upgrade from 8.04 to 10.04
Yes, I think so. Somehow the service of gdm does not work…
If you still want to test the upgraded system, follow
http://drbl.sourceforge.net/one4all/#upgrade
to reinstall drbl.
Steven.
The client is left at the command line login prompt. I can log in and start gdm and that works. So the drbl client start up is not calling the gdm for some reason. I've got our clients working here with a work-around - I placed a shell script in rcS.d to call 'service gdm start'. Now when the client boots it launches the gdm. We've been using this workaround for a day and bit now without any issues so far.
Hopefully I'll soon get a chance reinstall DRBL on our upgraded system and I'll post back any results. In the meantime I have spotted something on both the upgraded and fresh install. On a client machine when you open a Terminal the prompt no longer contains the machine name, it just displays:
bob@(none):~$
Thanks.
As for "it just displays: bob@(none):~$ " -> Is this on the upgraded system? Or the fresh install one?
Normally you can modify the prompt in the /etc/bash.bashrc.
Steven.
It occurs on both, and the login screen does not the display the machines name
The (none) seems to indicate the machines does not have a name
So what's the result if you run this on DRBL server:
cat /tftpboot/nodes/$IP/etc/hostname
(replace the $IP with that of your client's IP address)?
Steven.
I ran cat /tftpboot/nodes/$IP/etc/hostname and it outputs the correct hostname fo the client
I solved the issue. It seems in Ubuntu 10.04 it no longer likes underscores in the hostname. The hostname service fails, but it does not give you any error messages. Once I had removed the undescores the service was able to start.
OK, thanks for sharing that.
Steven.