cobbler April 2009

cobbler@lists.fedorahosted.org

51 participants
71 discussions

repo issues after 1.6.1 to 1.6.2
by Tom Brown 09 Apr '09

09 Apr '09

I upgraded my test system to 1.6.2 and i am having similar issues with repo's not showing. This is what a cobbler check says # cobbler check The following potential problems were detected: #0: One or more repos referenced by profile objects is no longer defined in cobbler: uuCentOS-5.2-x86_64-base, uusoftware-5-x86_64 Now these were working in 1.6.1 but why are they prefixed with 'uu' i had noted the use of 'u' for unicode data i think but 'uu' ? # cobbler repo list CentOS-5.2-x86_64-base software-5-x86_64 # cobbler repo report repo : CentOS-5.2-x86_64-base arch : x86_64 breed : yum comment : created : Mon Mar 30 14:19:03 2009 createrepo_flags : -c cache environment : {} keep updated : True mirror : http://10.10.10.7/mirror/5.2/os/x86_64/ mirror locally : False modified : Mon Mar 30 14:19:03 2009 owners : [u'admin'] priority : 99 rpm list : yum options : {} repo : software-5-x86_64 arch : x86_64 breed : yum comment : created : Mon Mar 30 14:20:18 2009 createrepo_flags : -c cache environment : {} keep updated : True mirror : http://10.10.10.7/yum/bf-software/5/x86_64/ mirror locally : False modified : Mon Mar 30 14:20:18 2009 owners : [u'admin'] priority : 99 rpm list : yum options : {} # cobbler profile report CentOS-5.2-xen-x86_64 profile : CentOS-5.2-xen-x86_64 distro : CentOS-5.2-xen-x86_64 comment : created : Thu Mar 12 10:43:23 2009 dhcp tag : default enable menu : True kernel options : {u'dns': u'192.168.11.1'} kickstart : /var/lib/cobbler/kickstarts/CentOS-5.2-xen-x86_64.ks ks metadata : {} mgmt classes : [] modified : Mon Mar 30 15:04:56 2009 name servers : [u'10.5.216.215', u'10.5.216.216'] name servers search : [] owners : [u'admin'] post kernel options : {} redhat mgmt key : <<inherit>> redhat mgmt server : <<inherit>> repos : ['uuCentOS-5.2-x86_64-base', 'uusoftware-5-x86_64'] server : <<inherit>> template_files : {} virt bridge : xenbr0 virt cpus : 1 virt file size : 5 virt path : virt ram : 512 virt type : xenpv any thoughts ? This seems very similar to what happened during the upgrade last time.

3 36

Workaround for EL4 unicode problem with repos field
by Michael DeHaan 08 Apr '09

08 Apr '09

Regarding the unicode thread: https://fedorahosted.org/pipermail/cobbler/2009-April/003722.html Here is a fix: http://git.fedorahosted.org/git/?p=cobbler;a=commit;h=b4f0fe496d05e9acce310… Steps to test: CentOS 4.7 base install in KVM Configure for both EPEL 4 and EPEL 4 testing # git clone git://git.fedorahosted.org/cobbler # cd cobbler # make devinstall # touch /tmp/foo # cobbler distro add --name=d1 --initrd=/tmp/foo --kernel=/tmp/foo # cobbler repo add --name=r1 --mirror=http://x # cobbler repo add --name=r2 --mirror=http://y # cobbler profile add --name=p1 --distro=d1 --repos="r1 r2" # cobbler profile report You should see the unicode version: repos: [ u"r1", u"r2" ] or the ascii version: repos: [ "r1", "r2" ] Both of which are correct repos: [ "ur1", "ur2" ] Is not correct. 1.6.3 will be released approximately next Friday. Enterprising readers may note that the above patch also contains some streamlining of the repo assignment code that is also *not related* to the bugfix. No sense not taking the opportunity to do some minor cleanup at the same time. Please ignore the comments from me simplejson versioning, we will not be seeking an update for that. --Michael

2 1

Problems importing Scientific Linux distros
by Jens Ahrens 08 Apr '09

08 Apr '09

Hi everybody, I found a solution to a problem that has already been addressed several times: https://fedorahosted.org/cobbler/ticket/374 http://www.mail-archive.com/et-mgmt-tools@redhat.com/msg02552.html The import of a DVD-Image of a Scientific linux distro fails with: No distro signature for kernel at /var/www/cobbler/ks_mirror This problem occurred with cobbler 1.4.3-4 and also with the latest version 1.6.2-1. I just patched the file /usr/lib/python2.4/site-packages/cobbler/action_import.py adding the following lines(*) ------------------------------------ [ 'CentOS/rpms' , "redhat" ], [ 'CentOS' , "redhat" ], (*) [ 'SL' , "redhat" ], (*) [ 'SL/RPMS' , "redhat" ], (*) [ 'SL/rpms' , "redhat" ], [ 'Packages' , "redhat" ], [ 'Fedora' , "redhat" ], ------------------------------------- Now the import works fine. Maybe you want to integrate this in the next release. Thanks and cheers, Jens

2 1

[Fwd: A preview of the deployment functions for 1.8]
by Michael DeHaan 08 Apr '09

08 Apr '09

I originally sent this to the -devel list, though I meant to send this here. Currently works now for specific hosts, working on assigning things to a simple pool of hosts shortly. # cobbler deploy --system=NameOfSystemRecordOfGuest --virt-host=NameOfSystemRecordOfHost [--method=ssh|func] It may look a bit like a SSH wrapper in the first form, because it is... though it will pull hostnames from cobbler and use those and make sure the right parameters are sent to koan. The Func method uses func for authentication, so you don't need to run ssh-agent. Rather simple ATM, but growing ... comments welcome. --Michael

2 2

cobbler 1.6.2 sync error
by Marcos A. Mondragon 06 Apr '09

06 Apr '09

Hello, I was upgrading a test server and when I run cobbler sync this pops out: # *** ERROR *** # # There is a templating error preventing this file from rendering correctly. # # This is most likely not due to a bug in Cobbler and is something you can fix. # # Look at the message below to see what things are causing problems. # # (1) Does the template file reference a $variable that is not defined? # (2) is there a formatting error in a Cheetah directive? # (3) Should dollar signs ($) be escaped that are not being escaped? # # Try fixing the problem and then investigate to see if this message goes # away or changes. # # # cannot find 'system' while searching for 'iface.system.netboot_enabled' # File "/usr/lib/python2.4/site-packages/cobbler/templar.py", line 115, in render # data_out = str(t) # # File "/usr/lib/python2.4/site-packages/Cheetah/Template.py", line 982, in __str__ # def __str__(self): return getattr(self, mainMethName)() # # File "cheetah_DynamicallyCompiledCheetahTemplate_1239035073_59_21648.py", line 394, in respond # Error templating file: /etc/dhcpd.conf I'm going to guess from the message that I did something wrong but the question is where to start looking... Thanks, Marc Mondragon Fox River Financial Resources 2100 Enterprise Avenue Geneva, IL 60134

3 3

[TESTING] 1.60 preview rpms now available
by Michael DeHaan 05 Apr '09

05 Apr '09

Hi folks, Source rpms for 1.6 are now available for those that want to help test 1.6 and see what's new. http://mdehaan.fedorapeople.org/files/cobbler/testing/ This is a testing release, so you don't want to install these in production. We'll test on this for 1.5-2 weeks and then see about a release. You can see the list of things that have changed here: https://fedorahosted.org/pipermail/cobbler/2009-March/003205.html Thanks! --Michael

5 18

post network snippet w/ dhcp
by matt whiteley 04 Apr '09

04 Apr '09

With the new version of the post_install_network_snippet (1.6.2) and a dhcp host, I am getting some odd behavior. The snippet writes over the current /etc/resolv.conf (that was generated by dhclient-script) but doesn't rerun dhclient to regenerate it. This causes anything following it in the post script to fail name resolution. Micahel suggested on irc the possibility of just running this snippet last, I am fine with this if that is the general consensus but I thought maybe this behavior was unintended. /etc/resolv.conf before snippet: ; generated by /sbin/dhclient-script search cic.pdx.edu nameserver 131.252.120.128 nameserver 131.252.120.129 /etc/resolv.conf after snippet: ; generated by /sbin/dhclient-script search This host is in a profile that doesn't have 'name servers' or 'name servers search' defined since all hosts using it use dhcp. I will add some settings and reload to test if that fixes things. -- matt whiteley <whiteley(a)pdx.edu>

2 2

dhcp - bonded interfaces - pxe
by Tom Brown 03 Apr '09

03 Apr '09

This is a useage question i think. Historically bonds have been applied by snippets in %post but to me it seems to make most sense to have cobbler handle it natively now that it seems to be able to. Lets say we have a box with 4 nics and those nics make 2 bonds - is it possible that this box can still do a DHCP and PXE off just 1 interface at install time as i presume you need to PXE/DHCP to a real interface as opposed to a bond? How is this handled within cobbler as i see the bonds seem to be sorted out in the post_network_config snippet so i am just wondering how this all works in practice? thanks

3 2

Is this a problem related to Cobbler?
by Christopher Dale 03 Apr '09

03 Apr '09

[root@sg07 ~]# !koan koan --virt --server=192.168.56.102 --system=command01 --virt-name=command01 - reading URL: http://192.168.56.102/cblr/svc/op/ks/system/command01 install_tree: http://192.168.56.102:80/cblr/links/rhel53-x86_64 libvirtd (pid 6327) is running... downloading initrd initrd.img to /var/lib/xen/initrd.img url=http://192.168.56.102/cobbler/images/rhel53-x86_64/initrd.img - reading URL: http://192.168.56.102/cobbler/images/rhel53-x86_64/initrd.img downloading kernel vmlinuz to /var/lib/xen/vmlinuz url=http://192.168.56.102/cobbler/images/rhel53-x86_64/vmlinuz - reading URL: http://192.168.56.102/cobbler/images/rhel53-x86_64/vmlinuz libvir: Xen error : Domain not found: xenUnifiedDomainLookupByUUID libvir: Xen error : Domain not found: xenUnifiedDomainLookupByName libvir: Xen Daemon error : POST operation failed: (xend.err "Error creating domain: (2, 'Invalid kernel', 'xc_dom_parse_elf_kernel: ELF image has no shstrtab\\n')") libvirt.libvirtError virDomainCreateLinux() failed POST operation failed: (xend.err "Error creating domain: (2, 'Invalid kernel', 'xc_dom_parse_elf_kernel: ELF image has no shstrtab\\n')") File "/usr/lib/python2.4/site-packages/koan/app.py", line 191, in main k.run() File "/usr/lib/python2.4/site-packages/koan/app.py", line 408, in run self.virt() File "/usr/lib/python2.4/site-packages/koan/app.py", line 683, in virt return self.net_install(after_download) File "/usr/lib/python2.4/site-packages/koan/app.py", line 602, in net_install after_download(self, profile_data) File "/usr/lib/python2.4/site-packages/koan/app.py", line 681, in after_download self.virt_net_install(profile_data) File "/usr/lib/python2.4/site-packages/koan/app.py", line 1105, in virt_net_install virt_type = self.virt_type File "/usr/lib/python2.4/site-packages/koan/xencreate.py", line 174, in start_install guest.start_install() File "/usr/lib/python2.4/site-packages/virtinst/Guest.py", line 822, in start_install return self._do_install(consolecb, meter, wait) File "/usr/lib/python2.4/site-packages/virtinst/Guest.py", line 843, in _do_install self.domain = self.conn.createLinux(install_xml, 0) File "/usr/lib64/python2.4/site-packages/libvirt.py", line 573, in createLinux if ret is None:raise libvirtError('virDomainCreateLinux() failed', conn=self)

4 6

Re: [PATCH] keep_ssh_host_keys and logging-snippets
by Jasper Capel 03 Apr '09

03 Apr '09

On Thu, 2009-03-12 at 17:03 +0000, Tom Brown wrote: > >> dont suppose you could forward me the ssh snippet could you as i could > >> do with testing it and i cant seem to get git going at the moment > >> > >> thanks > >> > > > > Sure, there you go! > > > > > > > sorry to follow this up directly with you but do you know why this might > occur > *snip* > it only happens when i try to use the ssh snippet and its when the pre > snippets get run - any clue how i can do about debugging this? > > thanks (And on-list again, please use the list for issues like this, others can learn from it, or help) Hmm. No, I'll run a few tests of my own sometime next week. Unfortunately, debugging %pre is a pain. You actually can't, in text-only mode, as you can't swap out to the console / logs using ALT +f2 / ALT+f3. You'll have to use virt-manager to get to those screens, I guess. Maybe someone has a better idea. If you add that log_ks_pre snippet I just posted to the top of the %pre section, you might have some useful information during that stage under /tmp. Did this happen for you using the old one as well? (I'd guess so, as the error doesn't indicate LVM). -Jasper

2 1

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

cobbler April 2009