ImportError: No module named cobbler.services
by Pacey, Mike
Hi folks,
I've been using the EPEL yum package of cobbler on Scientific Linux 6.4 for a few weeks now without a problem. This week, a system build failed with the error "Unable to download the kickstart file". Looking through various logs, I think I can trace the problem to /var/log/httpd/error_logs:
[Tue Aug 13 15:27:42 2013] [error] [client 10.41.6.6] mod_wsgi (pid=2544): Target WSGI script '/var/www/cobbler/svc/services.py' cannot be loaded as Python module.
[Tue Aug 13 15:27:42 2013] [error] [client 10.41.6.6] mod_wsgi (pid=2544): Exception occurred processing WSGI script '/var/www/cobbler/svc/services.py'.
[Tue Aug 13 15:27:42 2013] [error] [client 10.41.6.6] Traceback (most recent call last):
[Tue Aug 13 15:27:42 2013] [error] [client 10.41.6.6] File "/var/www/cobbler/svc/services.py", line 27, in <module>
[Tue Aug 13 15:27:42 2013] [error] [client 10.41.6.6] from cobbler.services import CobblerSvc
[Tue Aug 13 15:27:42 2013] [error] [client 10.41.6.6] ImportError: No module named cobbler.services
If I understand that correctly, python isn't finding the cobbler/services* files located in /usr/lib/python2.6/site-packages. I can run "python /var/www/cobbler/svc/services.py" happily from the command line, so I'm puzzled as to why httpd/wsgi can't find it. There's a few ways I could hack the setup to ensure it finds it, but I'd prefer to work out what exactly has broken and why.
I installed a fresh version of cobbler on a test machine, but can't see any difference. Specifically, the cobbler.conf and wsgi.conf in /etc/httpd/conf.d and the main httpd.conf file are identical.
Can anyone suggest any further diagnostics I can run?
Regards,
Mike.
-----
Dr Mike Pacey, Email: M.Pacey(a)lancaster.ac.uk<mailto:M.Pacey@lancaster.ac.uk>
HPC Manager, Phone: 01524 510659
Information Systems Services, Fax: 01524 594459
ISS Building, Lancaster University,
Lancaster LA1 4WA
10 years, 8 months
Blank interface name?
by Bret Wortman
I have a system which is reporting a blank interface name. In part:
# cobbler system dumpvars --name asipa
:
interface_master_ :
interface_type_ :
interfaces : {'': {'ipv6_address': '', 'interface_type': '', 'static':
True, ...}
:
#
The actual interface is eth0, not something weird. Any idea where I've gone
off the rails here? I set up the system using:
# cobbler system add --name=asipa --profile=F18-x86_64
--mac=00:50:56:12:12:12 --ip-address=1.2.3.4 --gateway=1.2.3.1 --hostname=
asipa.foo.com --static=1
It's a VMware VM, but I've done this with VMware VMs before without any
issues.
*
*
*Bret Wortman*
http://damascusgrp.com/
http://about.me/wortmanbret
10 years, 8 months
Re: [cobbler] Cobbler sync going to /var/lib/tftpboot instead of /tftpboot
by Michael
A general remark about tftpboot location on CentOS 6.4:
The package *syslinux-tftpboot* places its files in.... /tftpboot
Don't be confused while using cobbler.
Regards,
Michael.
2013/8/12 Bryan Hughes <bhughes(a)gisfederal.com>
> awesome. Thanks for the quick response and help.
>
> Appreciate it
>
> On Aug 12, 2013, at 1:27 PM, James Cammarata <jimi(a)sngx.net> wrote:
>
> No there is not, /var/lib/tftptboot is the correct location for in.tftpd
> on centos 6.4. If you really want it in /tftpboot, create a symlink or bind
> mount.
>
> On Mon, Aug 12, 2013 at 12:02 PM, Bryan Hughes <bhughes(a)gisfederal.com>wrote:
>
>> CentOS 6.4.
>>
>> Is there a way to force it to go to /tftpboot or no?
>> On Aug 12, 2013, at 1:01 PM, James Cammarata <jimi(a)sngx.net> wrote:
>>
>> It is going there because that is what is detected for your OS. What
>> version of Linux are you running?
>>
>> On Mon, Aug 12, 2013 at 11:50 AM, Bryan Hughes <bhughes(a)gisfederal.com>wrote:
>>
>>> How do I change the tree cobbler uses for tftp? I can't seem to find
>>> the setting anywhere. the /etc/xinetd.d/tftp is pointing to /tftpboot but
>>> when I run cobbler sync it puts everything to /var/lib/tftpboot. My
>>> question is, how do I change it to point to /tftpboot like it should be?
>>> _______________________________________________
>>> cobbler mailing list
>>> cobbler(a)lists.fedorahosted.org
>>> https://lists.fedorahosted.org/mailman/listinfo/cobbler
>>>
>>
>> _______________________________________________
>> cobbler mailing list
>> cobbler(a)lists.fedorahosted.org
>> https://lists.fedorahosted.org/mailman/listinfo/cobbler
>>
>>
>>
>> _______________________________________________
>> cobbler mailing list
>> cobbler(a)lists.fedorahosted.org
>> https://lists.fedorahosted.org/mailman/listinfo/cobbler
>>
>>
> _______________________________________________
> cobbler mailing list
> cobbler(a)lists.fedorahosted.org
> https://lists.fedorahosted.org/mailman/listinfo/cobbler
>
>
>
> _______________________________________________
> cobbler mailing list
> cobbler(a)lists.fedorahosted.org
> https://lists.fedorahosted.org/mailman/listinfo/cobbler
>
>
10 years, 8 months
Cobbler sync going to /var/lib/tftpboot instead of /tftpboot
by Bryan Hughes
How do I change the tree cobbler uses for tftp? I can't seem to find the setting anywhere. the /etc/xinetd.d/tftp is pointing to /tftpboot but when I run cobbler sync it puts everything to /var/lib/tftpboot. My question is, how do I change it to point to /tftpboot like it should be?
10 years, 8 months
koan install ubuntu server 12.04
by Ronald J. Yacketta
Never used ubunut before but looking to use it as our cloud base server OS. Imported 12.04 into cobbler and attempting to koan install an instance to play with and ran into the following
koan --virt --server=cobbler.srvr --profile=ubuntu-12.04-x86_64 --virt-type=kvm --virt-name=nuvola --virt-path=/mnt/sas3_1/ -w &
[1] 19989
[root@bs3-1 ~]# - looking for Cobbler at http://cobbler.srvr:80/cobbler_api
- reading URL: http://cobbler.srvr/cblr/svc/op/ks/profile/ubuntu-12.04-x86_64
install_tree: http://cobbler.srvr/cblr/ks_mirror/ubuntu-12.04-x86_64/
libvirtd (pid 2021) is running...
- adding disk: /mnt/sas3_1//nuvola-disk0 of size 35 (driver type=raw)
- ['virt-install', '--connect', 'qemu:///system', '--name', 'nuvola', '--ram', '1024', '--vcpus', '1', '--autostart', '--vnc', '--virt-type', 'kvm', '--extra-args=ks=http://cobbler.srvr/cblr/svc/op/ks/profile/ubuntu-12.04-x86_64 ksdevice=link lang= text ', '--location', 'http://cobbler.srvr/cblr/ks_mirror/ubuntu-12.04-x86_64/', '--arch', 'x86_64', '--os-variant', 'generic26', '--disk', 'path=/mnt/sas3_1//nuvola-disk0,size=35,driver_type=raw', '--network', 'bridge=br9', '--wait', '0', '--noautoconsole']
Starting install...
Retrieving file version.info... | 116 B 00:00 ...
ERROR Couldn't find hvm kernel for Ubuntu tree.
command failed (1)
can someone shed some light here?
Thanks
-Ron
10 years, 8 months
Making a mgmtclass definition causes 500 error
by Bryan Hughes
When I have a management class and set is definition? to true. I get a 500 error when doing /usr/bin/cobbler-ext-nodes <node>. If I change the if definition? to false and then run the same ext-nodes command it will work?
Do I do something different with a definition instead of a class?
10 years, 9 months
Integrate existing puppet baseline into cobbler
by Bryan Hughes
Hello,
I have an existing puppet baseline using node classifiers but want to move this all into cobbler's ENC. Is there a good way to move all the puppet classes (/etc/puppet/modules) into cobbler management classes?
Also, can you do paramertized classes and overwrite the variable on a per server basis or do you need a separate management class for each value of the variable. For instance if there is a class named foo and a parameter of bar. bar can be set to true or false. Do I need a management class for bar set to true and bar set to false or can I set that variable somewhere besides the management class?
Thanks,
10 years, 9 months
Problem Importing A New Distribution -- Ubuntu 13.04
by Bob Cochran
I have not been using cobbler in a while but now I need to. I basically
need to install virtual guests (running under the Mac OS X version of
Virtual Box...and I have a Mac OS X host). My goal is to import the
Ubuntu 13.04 amd64 desktop distribution into cobbler and then have my
VirtualBox guest pxe boot to the installer for itand do the entire
install over the network.
Cobbler version:
[root@cobbler1 ~]# cobbler version
Cobbler 2.4.0
source: 923574f, Thu Jun 20 00:42:48 2013 -0500
build time: Sun Aug 4 11:58:30 2013
On reading the documentation on cobblerd.org for Distros:
"In general, it’s really a lot easier to follow the import workflow --
it only requires waiting for the mirror content to be copied and/or
scanned. Imported mirrors also save time during install since they don’t
have to hit external installation sources. Please read the Cobbler
Import <http://www.cobblerd.org/manuals/2.4.0/4/2/4_-_Import.html>
documentation for more details."
...I decided to try this without first creating "distro" or "system"
objects:
cobbler import --path=/media/stuff --name=ubuntu-13-x86_64 --arch=x86_64
I have this result:
received on stderr:
Found a matching signature: breed=ubuntu, version=raring
Adding distros from path /var/www/cobbler/ks_mirror/ubuntu-13-x86_64-x86_64:
No distros imported, bailing out
!!! TASK FAILED !!!
Do I need to manually create a "distro" and "system" first?
Thanks
Bob Cochran
10 years, 9 months
need help with a import problem: rhel6.4
by Tan, Jeff
Hi,
I am having problems importing the latest rhel 6.4 iso. It keeps coming back with : No signature matched in /var/www/cobbler/ks_mirror/rhel6.4local.3
The command I am running is : cobbler import --name=rhel6.4local.3 --path=/media
/media is a DVD mounted on a vmware CentOS 6.2 host. I tried to import also from a local repo but am getting the same error message.
I tried to do : cobbler signature update then restarted cobblerd but it's no go. What am I missing?
Regards,
Jeff Tan
10 years, 9 months
using parameterized mgmt classes
by Greg Chavez
I suspect I am being obtuse here. I'm trying to use the arusso/vnc module
from puppet forge which is declared like this:
class { 'vnc':
refresh => true,
servers => [ 'user' => 'pete',
'args' => '-geometry 1920x1200 -depth 16' ],
service_ensure => true,
xstartup_template => 'myclass/xstartup.erb',
}
Via the Web console, I did the following:
--------------------
(1) Created management class "vnc":
**** name: vnc
**** class name: vnc
**** parameters/variables: service_ensure xstartup_template refresh servers
(2) In system object:
**** selected class "vnc"
**** Added yaml to management parameters which was converted to:
{ service_ensure: True, xstartup_template: vnc/xstartup.erb, refresh: True,
servers: [{'user': 'pete'}, {'args': '-geometry 1920x1200 -depth 16'}] }
--------------------
However, this is what cobbler-ext-node returns for the system:
~> cobbler-ext-nodes pete.local
classes:
site: {}
vnc:
refresh: '~'
servers: '~'
service_ensure: '~'
xstartup_template: '~'
environment: production
parameters:
from_cobbler: 1
refresh: true
servers:
- user: pete
- args: -geometry 1920x1200 -depth 16
service_ensure: true
tree: http://@@http_server@@/cblr/links/centos-6.4-x86_64
xstartup_template: vnc/xstartup.erb
I didn't think that this would work the way I did it, but I was hoping the
resultant breakage would give me a clue as to how to fix. No such luck. I
am stumped.
Can anyone tell me what I'm doing wrong? Mahalo in advance.
--
\*..+.-
--Greg Chavez
+//..;};
10 years, 9 months