June 2010 - cobbler - Fedora Mailing-Lists

cobbler is not creating the core-0 repo on new installs

by Jeremy Hansen

It seems that cobbler is not creating the core-0 repo on new installs. My distro tree actually lives on another server and I manually added the distro definition to the cobbler server that doing the install. Here's the distro config: Name : CentOS-5.2-x86_64 Architecture : x86_64 Breed : redhat Comment : rhel5.2 Initrd : /var/www/cobbler/ks_mirror/CentOS-5.2-x86_64/images/pxeboot/initrd.img Kernel : /var/www/cobbler/ks_mirror/CentOS-5.2-x86_64/images/pxeboot/vmlinuz Kernel Options : {} Kernel Options (Post Install) : {} Kickstart Metadata : {'tree': 'http://fopp-cbl0000.ash3.blah.com/cblr/links/CentOS-5.2-x86_64/'} Management Classes : [] OS Version : rhel5 Owners : ['admin'] Red Hat Management Key : <<inherit>> Red Hat Management Server : <<inherit>> Template Files : {} How can I trace down the reason for it not adding the core-0 definition to the cobbler-config.repo in /etc/yum.repos.d? Thanks -jeremy

13 years, 10 months

1
0
0 / 0

reposync: error: no such option: -m

by Jeremy Hansen

I recently upgraded a cobbler server from 1.6.8 to 2.0.3 and now I'm seeing this issue with reposync: creating: /var/www/cobbler/repo_mirror/cloudera-testing/.origin/cloudera-testing.repo running: /usr/bin/reposync -l -m -d --config=/var/www/cobbler/repo_mirror/cloudera-testing/.origin/cloudera-testing.repo --repoid=cloudera-testing --download_path=/var/www/cobbler/repo_mirror usage: Reposync is used to synchronize a remote yum repository to a local directory using yum to retrieve the packages. /usr/bin/reposync [options] reposync: error: no such option: -m returned: 2 Exception occured: cobbler.cexceptions.CX Exception value: 'cobbler reposync failed' Exception Info: File "/usr/lib/python2.4/site-packages/cobbler/utils.py", line 122, in die raise CX(msg) Any ideas? Thanks -jeremy

13 years, 10 months

1
1
0 / 0

cobbler-ext-nodes ignoring no_proxy

by Nicolás Pace

Hi people, I found out that cobbler-ext-nodes is ignoring the no_proxy environment variable. I need cobbler-ext-nodes to ignore http_proxy, or to pay attention to no_proxy, in order to fetch data from the local network and not from the server. I've seen that cobbler-ext-nodes is using urlgrabber to fetch this info, and that's why it is not using no_proxy (it is an old bug in urlgrabber that didn't seem to be patched yet :S). Regards, -- Ing. Nicolás Pace http://www.linkedin.com/in/nickar/ http://www.jornadasdelsur.org.ar/

13 years, 10 months

2
2
0 / 0

specifying another next_server dynamically??

by Jeremy Hansen

Is it possible to specify an alternate next_server for a given system profile? We have some solaris boxes which we'd like to exist in the cobbler db and we'd like cobbler to manage its dns and dhcp but we want to specify another next_server for this profile so it will boot from a jumpstart server at another address. I was hoping for a simple "--next-server=" flag to system add, but I don't see such an option. Thanks! -jeremy

13 years, 10 months

2
2
0 / 0

Missing xxxx-primary.sqlite.bz2

by Baptiste AGASSE

Hi all, I use cobbler to mirror EPEL repo for some packages I have created the repo in cobbler with rpm-list option : --rpm-list="cobbler chkrootkit func python-simplejson certmaster nagios* nrpe* php* perl-* mod_* pam_* debmirror*" This sync package as i want but not download/generate the xxxx-primary.sqlite.bz2 for the mirrored repo, and when i want to query this repo, I get an error from yum : # yum install debmirror Loaded plugins: allowdowngrade, fastestmirror, priorities, security Loading mirror speeds from cached hostfile http://cobbler.exemple.com/cobbler/repo_mirror/CentOS-5-epel-x86_64/repod...: [Errno 14] HTTP Error 404: Not Found Trying other mirror. Error: failure: repodata/40512a231767c5c4efe65d9dd267793d58a8c4ab-primary.sqlite.bz2 from epel: [Errno 256] No more mirrors to try. cobbler version (from EPEL repo for EL5): Cobbler 2.0.3.1 source: ?, ? build time: Mon Mar 1 20:08:11 2010 Anyone have the same problem ? Regards.

13 years, 10 months

2
2
0 / 0

Confirming when a VM has been provisioned?

by lists＠truthisfreedom.org.uk

Hi all, Is there anyway of querying the cobbler XML-RPC API to find out if a node has finished building? At the moment, the best way I can find is to check the last two lines of the logs, but I can't help thinking that if the server knows when the build has finished so it can put it in the log, the API must also know when the build is finished so i can "tell" the server to mark the build as completed in the log. I'm running some automated testing at the moment and I need to know that my VMs are up before I run the config-mgmt system against it! :) Thanks in advance, Matt

13 years, 10 months

8
11
0 / 0

Cobbler Export ?

by Tom Brown

Hi I know 'cobbler export' does not exist as a way for dumping the systems and profiles from within a cobbler install but whats the best way to migrate from one install server to another? Could such a feature as 'cobbler system export' possibly exist in the future? I'd rather not just copy /var/lib/cobbler/config thanks

13 years, 10 months

1
0
0 / 0

Mac netboot support?

by Jennings, Jared L CTR USAF AFMC 46 SK/CCI

Does Cobbler support anything besides BIOS, PXE and PXELINUX? I've configured my dhcpd like it says in http://www.afp548.com/article.php?story=20061220102102611 to provide Apple Boot Service Discovery Protocol (BSDP) extensions to its DHCP replies, and successfully gotten a Mac to fetch EFI executables over the network and run them. It appears that elilo (http://elilo.sourceforge.net/) can run as an EFI executable, and fetch a kernel and initrd via TFTP, as PXELINUX does, and start everything up. (Appears: I haven't gotten there yet.) But elilo's means of choosing a configuration is different from that of PXELINUX, so I think to support it Cobbler will need to make another different tree of config files under the tftp root. Has anyone done this yet?

13 years, 10 months

2
1
0 / 0

Setup help -- will pay (Cobbler 1.2.9)

by Frank

Dear All: So far I was able to install Cobbler 1.2.9 on Fedora 8. This is a VPS (Virtuozzo) server ... couldn't get SWAP set up for CentOS, so I am using this older Fedora version. All worked fine -- I was able to import "CentOS 4.8" and "CentOS 5.4" (32 and 64-bit versions) into the Cobbler server, using the command .... cobbler import --path=rsync: (...) But now I am lost. I see these distributions fine in the Cobbler web interface listed. I did set the Cobbler server IP in /etc/cobbler/settings But then I am not sure what to do from there. I have several servers in the same rack, same network, than this VPS Cobbler server. What I want to do is to (a) be able to *remotely* reinstall CentOS (above versions) to any of these other servers if needed, (b) to also remotely install CentOS on new servers that have been set up in the rack (if possible). (I can reboot all servers remotely using a device to interrupt power.) I would be happy for all concrete advice, or for someone to pay for configuring this from here. Thanks! Frank EMAIL: support AT server-1.com

13 years, 11 months

2
1
0 / 0

Re: Config help? .. Cobbler 1.2.9

by Dylan Swift

On 12 June 2010 23:12, Frank <support(a)server-1.com> wrote: > Thank you Dylan. I continue to be confused. Aren't the values (e.g. > the IP or the domain name, etc.) in /etc/dhcpd.conf handed on to the > server where an OS should be installed? I'm not sure what is confusing you - is it the address you specified for the system is not in the dhcpd.conf file, or the address in the kickstart file does not match that in the dhcpd.conf file. I'll try to clarify (apologies if this is stuff you already understand)... The DHCP server will make an offer of IP address, subnet, gateway, dns, etc to a dhcp request. The client (ie where the os is to be installed, the system that made the dhcp request) will then accept this offer and the address will be allocated, by the dhcp server, to the MAC address of the client, for the duration of the lease, or until the client relinquishes the address. This is what DHCP does. In addition, a 'next-server' parameter is also passed in the dhcp offer. This tells the client where to go next to obtain further instruction, only if this is a PXE boot. The further instruction is a program (eg memtest) or a kernel (for installation / live distro running) If you use the templated version of dhcp (ie "manage_dhcp; 1" in /etc/cobbler/settings and you have a configured '/etc/cobbler/dhcp.template' file) then when you define a new system and 'sync' cobbler, a new entry will be added to the working dhcpd.conf file that defines specific instructions for the system you have just defined. Such as: host generic1 { hardware ethernet e4:1f:13:60:9a:64; fixed-address 192.168.42.1; option host-name "node1.xxx.yyy.localdomain"; filename "/pxelinux.0"; next-server 192.168.42.152; } where: * my cobbler server (cobbler + dhcp + tftpboot) is at 192.168.42.152 * the client will set up IP networking at 192.168.42.1 * my dhcp pool is 192.168.42.1-192.168.42.150 One thing to note - the IP address in the entry above is an address from the dhcp pool, NOT the address you defined for the server in cobbler system settings. That address will be in the kickstart file when it is served to the client at installation time. There should also be a generic installation entry - one that does not define an IP address, but does tell the client where to go next. In this case the client would boot to a menu that offers any of the profiles for installation. It should also default to boot from disk, meaning that a human operator would have to initiate a re-install. > I added a "System" in Cobble_web -- is the IP, subnet, hostname info > then taken from that profile? One add-on question to this: Is there > any way to tell Cobbler how large the SWAP partition should be, and > is there a way to tell it how to partition the HDDs? The system defines specific information, such as hostnames, IP Addresses etc that are applied to a profile. A profile is a combination of kickstart + distro. You can define the hard disk layout in the kickstart file (usually 'autopart'), but if you look in most regular kickstart files you will find a section (commented out) that defines how the hard disks have been partitioned. These lines usualy start with 'clearpart', 'volgroup' 'logvol' and/or 'part'. Each of these keywords is followed by various parameters. There is good documentation online in the RedHat manuals (google is your friend). Hope that helps Dylan

13 years, 11 months

2
1
0 / 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

cobbler June 2010