VLANs and bonds problem.
by Jakov Sosic
Hi.
I've encountered very strange problem.
I want to set up the following network configuration:
eth0+eth1 = bond0
vlan10 => bond0.10
vlan11 => bond0.11
Also, for the purpose of PXE installation my network staff can set the
native vlan of eth0 to vlan 10.
Problem is that cobbler doesn't manage ISC DHCP very well in this case.
Address has to be set on bond0 also, and I didn't test how would that
affect final configuration of installed system...
This is the file I think that's culprit:
/usr/lib/python2.6/site-packages/cobbler/modules/manage_isc.py
And this is the code snippet:
if interface["interface_type"] in ("slave","bond_slave","bridge_slave"):
if interface["interface_master"] not in system.interfaces:
# Can't write DHCP entry; master interface does not
# exist
continue
ip = system.interfaces[interface["interface_master"]]["ip_address"]
interface["ip_address"] = ip
host = system.interfaces[interface["interface_master"]]["dns_name"]
else:
ip = interface["ip_address"]
host = interface["dns_name"]
After cobbler sync, i get this in /etc/dhcp/dhcpd.conf:
host generic8 {
hardware ethernet 00:11:5c:34:29:ff;
option host-name "myhost.mydomain";
option routers 192.168.1.1;
filename "/pxelinux.0";
next-server 192.168.1.2;
}
If you can notice, it misses the "fixed-address <IP>;" line in there, so
PXE won't boot it because the machine doesn't get the IP assigned...
What happens is that the snippet of python code checks if the interface
is slave, and if it is, then checks for the IP of the master interface.
In my case, bond0 is master interafce, but it does not have IP address
assigned because I do VLAN tagging over the bond.
And if I manually set the IP address on eth0, code again doesn't set the
IP because eth0 is flagged as slave, and because of that it checks the
IP of the master interface :D
Only viable solution is to set same IP address on bond0 and bond0.10,
and after the initial boot remove the IPADDR/NETMASK/NETWORK settings
from ifcfg-bond0 file and reboot again...
Is there any better way to achieve this? Can manage_isc.py be hacked in
someway, or some new option added?
I've tried with --interface=eth0 --managed=yes, but it doesn't help.
Ty and sorry for long post
11 years, 8 months
How to PXE boot F17 now?
by Bret Wortman
Has anyone figured out how to PXE boot F17 from Cobbler now? Specifically,
I'm trying to figure out how to specify the "inst.repo=" line now that it's
required.
I tried adding it via --kopts, but that hasn't appeared to work. I'll keep
experimenting, but was wondering if anyone had solved this already.
--
Bret Wortman
The Damascus Group
Fairfax, VA
http://bretwortman.com/
http://twitter.com/BretWortman
11 years, 8 months
No such option --autostart???
by Scot Floess
All,
I've got a CentOS 5.8 xen based server (doesn't support hardware virtualization). I'm running KOAN 2.2.3 and I get this error when I attempt to install a VM:
usage: virt-install --name NAME --ram RAM STORAGE INSTALL [options]
virt-install: error: no such option: --autostart
I've downgraded to KOAN 1.6.6 and everything works fine... Is there a command line option or something else I can do to continue using 2.2.3? I don't mind 1.6.6 - just when I yum update, I get the latest KOAN :)
Thanks!
Flossy
--
Make It So Number One
11 years, 8 months
cobbler reposync on centos 6.3
by Shake Chen
Hi
I run cobbler 2.2.3 in centos 6.2 , the cobbler reposync seem have no
problem. but in centos 6.3,
cobbler reposync
[epel6-x86_64: 7684 of 7685 ] Skipping existing
zvbi-devel-0.2.33-6.el6.x86_64.rpm
[epel6-x86_64: 7685 of 7685 ] Skipping existing
zvbi-fonts-0.2.33-6.el6.noarch.rpm
received on stderr:
Exception occured: <class 'cobbler.cexceptions.CX'>
Exception value: 'cobbler reposync failed'
Exception Info:
File "/usr/lib/python2.6/site-packages/cobbler/utils.py", line 126, in die
raise CX(msg)
Exception occured: <class 'cobbler.cexceptions.CX'>
Exception value: 'cobbler reposync failed'
Exception Info:
File "/usr/lib/python2.6/site-packages/cobbler/action_reposync.py", line
125, in run
self.sync(repo)
File "/usr/lib/python2.6/site-packages/cobbler/action_reposync.py", line
169, in sync
return self.yum_sync(repo)
File "/usr/lib/python2.6/site-packages/cobbler/action_reposync.py", line
402, in yum_sync
utils.die(self.logger,"cobbler reposync failed")
File "/usr/lib/python2.6/site-packages/cobbler/utils.py", line 134, in
die
raise CX(msg)
reposync failed, tries left: 0
Exception occured: <class 'cobbler.cexceptions.CX'>
Exception value: 'reposync failed, retry limit reached, aborting'
Exception Info:
File "/usr/lib/python2.6/site-packages/cobbler/utils.py", line 126, in die
raise CX(msg)
Exception occured: <class 'cobbler.cexceptions.CX'>
Exception value: 'reposync failed, retry limit reached, aborting'
Exception Info:
File "/usr/lib/python2.6/site-packages/cobbler/remote.py", line 89, in run
rc = self._run(self)
File "/usr/lib/python2.6/site-packages/cobbler/remote.py", line 251, in
runner
name=None, nofail=nofail, logger=self.logger)
File "/usr/lib/python2.6/site-packages/cobbler/api.py", line 758, in
reposync
return reposync.run(name)
File "/usr/lib/python2.6/site-packages/cobbler/action_reposync.py", line
147, in run
utils.die(self.logger,"reposync failed, retry limit reached, aborting")
File "/usr/lib/python2.6/site-packages/cobbler/utils.py", line 134, in
die
raise CX(msg)
!!! TASK FAILED !!!
but when I run
/usr/bin/reposync -l -m -d
--config=/var/www/cobbler/repo_mirror/epel6-x86_64/.origin/epel6-x86_64.repo
--repoid=epel6-x86_64 --download_path=/var/www/cobbler/repo_mirror -a x86_64
seem no problem.
--
Shake Chen
11 years, 8 months
Systems don't inherit profile's kickstart metadata?
by Joshua J. Kugler
So, this problem:
Doing koan installs and allowing auto-register gives me a system with
"localhost" as the host name and a weird floating point number as the
system name in Cobbler, instead of what I expected, which was the value
I provided to koan as --virt-name.
lead to this attempted solution:
I defined a system using the profile I wanted, and then try to run koan
with --system=SYSTEM_NAME.
which lead to this problem:
System installs fine, but the kickstart meta data does not take affect;
in this case, the chef job that was run when I installed via --profile
was not run when I installed via a --system= that used the same
previously mentioned profile.
which lead to this attempt:
koan --virt --profile=builder-el6-x86_64 --system=builder-el6-64-01\
--server=cobbler.example.com --virt-path=vg_group
which lead to this interesting bug:
- looking for Cobbler at http://cobbler.example.com:80/cobbler_api
- reading URL: http://172.29.1.101/cblr/svc/op/ks/system/builder-el6-
x86_64
Traceback (most recent call last):
<snip>
raise HTTPError(req.get_full_url(), code, msg, hdrs, fp)
HTTPError: HTTP Error 404: NOT FOUND
Why was it trying to pull the profile from the system section of the
API?
Can't you "override" the profile when you specify a system?
This is Cobbler/Koan 2.2.2 by the way.
Thanks for any help you can render!
j
--
Joshua J. Kugler - Fairbanks, Alaska
Azariah Enterprises - Programming and Website Design
joshua(a)azariah.com - Jabber: pedahzur(a)gmail.com
PGP Key: http://pgp.mit.edu/ ID 0x73B13B6A
11 years, 8 months
koan would no woking because DHCP entries only get generated if netboot_enable = true
by Shake Chen
Hi
Now in cobbler 2.3.2, only netboot_enable = true, would add the dhcp
entries. when the client finish instanlled , the netboot_enabled would
change to 0, also would delete the dhcp entries.
but when the client plan to use koan
koan --server=10.1.199.12 --replace-self --system=node15
the install can not finish, just because can not get the IP address.
--
Shake Chen
11 years, 8 months
cobbler/koan network problem
by Tim Coote
Hullo
I'm trying to upgrade a physical box to fedora 17. My cobbler server's an f15 box, running cobbler-2.2.2-1.fc15.noarch.
I can create a vm and install f17 on it.
I can re-install f16 on my physical computer
But if I change the distro to use to f17, the install breaks with dracut complaining that it cannot open the file /tmp/pre_install_network (or possibly /tmp/pre_network_install - I cannot see the physical box from here). When I look in /tmp on the partially installed system, there is no file, but there are a bunch of files with the p1p1 style of interface names.
Have I run into a problem with i/f naming because my cobbler server's quite an old fedora build, or is there likely to be some other problem. Either way, any ideas on diagnosis/workaround.
regards
Tim
11 years, 8 months
how to add two repo to profile in command
by Shake Chen
Hi
Now I want to add two repo epel6-x86_64 epel6-x86_64-testing to profle, but
no working.
cobbler profile edit --name=CentOS6.3-x86_64
--repos="epel6-x86_64,epel6-x86_64-testing"
--
Shake Chen
11 years, 9 months
DHCP entries only get generated if netboot_enable = true
by Dan Rao
Hi all,
I upgraded my Cobbler install to 2.2.3-1.el6 (from epel-testing) last
night.
Interestingly, when I run cobbler sync, only machines that have netboot
enabled get entries in my dhcpd.conf.
The offending line(s) appear to be in
/usr/lib/python2.6/site-packages/cobbler/modules/manage_isc.py - lines
164-165
----
if not interface["netboot_enabled"]:
continue
----
(This is in the section that generates the ISC dhcpd config file).
I commented those lines out to get my server back, but I was curious as
to why that was there in the first place.
Any insight would be appreciated!
Best,
Dan
11 years, 9 months
Can't read out variable repos in snippet
by Manuel.Laesser@swisslife.ch
Hello togehter
I would like to read out variable repos (cobbler profile dumpvars --name=rhel5-intranet | grep -w repos) in my snippet create_repos_snippet and download a yum repo file if the string is not empty.
[root@cobbler sl]# cobbler profile dumpvars --name=rhel5-intranet | grep -w repos
repos : rhel58-x86_64-20120723
<snipp>
#if 'rhel58' in $repos
#echo $repos
wget http://cobbler/pub/files/5_rhel5-updates.repo -O /etc/yum.repos.d/rhel5-updates.repo
#end if
</snipp>
Unfortunately this doesn't work. I have to specifie the whole name of the repo like
<snipp>
#if 'rhel58-x86_64-20120723' in $repos
#echo $repos
wget http://cobbler/pub/files/5_rhel5-updates.repo -O /etc/yum.repos.d/rhel5-updates.repo
#end if
</snipp>
Has anybody any idea how to test $repos without using the whole name of the repo? I also tried #if $repos != ' ' without any success.
Thank you for your help.
Manuel
11 years, 9 months