How does one set up dnsmasq/DHCP to play nice with Windows DNS/DHCP
by Dan White
I am in a mixed environment where the main DNS/DHCP is run on Windows machines and controlled by other folks.
I want to explore the possibility of running my own DHCP with dnsmasq on my Cobbler Server so that I do not have to wait for the cooperation of the Windows Admin to be able to PXE-boot my Linux machines. The network environment is such that all the Cobbler/PXE action takes place on a single subnet defined by a single value of the third octet of the IP address ( xxx.xxx.(my sub-net).xxx)
What I am looking for are the necessary considerations to co-exist peacefully on the network with the other DHCP server. I do not need to run my own DNS but it would speed up my tasking if I could do the DHCP/PXE setup myself.
“Sometimes I think the surest sign that intelligent life exists elsewhere in the universe is that none of it has tried to contact us.”
Bill Waterson (Calvin & Hobbes)
12 years, 4 months
cobbler 2.2 and empty subnet
by Christian Horn
Hi,
using a RHEL6.2 and the current cobble in epel, 2.0.11-2,
a system object can be added, and is handled properly.
Using cobbler 2.2.1-1 from testing leads to this:
# cobbler system add --name=tester6 --profile=rhel6u2-x86_64
# cobbler system edit --name=tester6 --hostname=tester6 \
--interface=eth0 --ip-address=10.0.100.6 \
--subnet=255.255.255.0 --mac=00:ff:06:00:00:01 --static=1
# cobbler system dumpvars --name=tester6 |grep subnet_
subnet_eth0 :
# cobbler system report --name=tester6 |grep Subnet
Subnet Mask : 255.255.255.0
Dumpvars is not presenting the netmask, also not 'getks'.
Performing the same steps with 2.0.11-2 everything works.
Accessing system objects that were created with 2.0.11-2
after upgrade to 2.2.1-1 leads to the subnet still beeing
accessable.
Any ideas on what I could possibly be missing here?
Did not spot a related ticket on github yet, got aware of
this from a report from another cobbler user.
cheers, Christian
12 years, 4 months
RE: Cobbler 2.2.1 not removing/signing puppet certs
by Jim Goddard
First of all, I apologize for not clipping out the digest in my previous email.
I have also noticed that the /var/log/cobbler/install.log has no entries since the Dec. 16th, which I am fairly certain corresponds to the day I upgraded to the 2.2.1 version of cobbler.
Still no idea why the cobbler triggers aren't firing.
Jim Goddard
Date: Fri, 6 Jan 2012 10:50:46 -0800
From: "Jim Goddard" <jgoddard(a)gmi-mr.com>
To: <cobbler(a)lists.fedorahosted.org>
Subject: Cobbler 2.2.1 not removing/signing puppet certs
Message-ID:
<DB1915EFB5825B46AD2448D173B62A170585AE29(a)CORP-MSG-01.CORP.GMI.LCL>
Content-Type: text/plain; charset="iso-8859-1"
Hello all,
I am not positive that this happened in concurrence with my upgrading cobbler to 2.2.1 from the el5 2.0.11 version, but I haven't changed that much aside from that.
I was having problems before in that cobbler seemed to fire off the command to sign new puppet certs way too late, which I had never successfully figured out why. Now it is not ever firing that task.
Now, however, cobbler is not removing old certs either. I have the following puppet related settings in my /etc/cobbler/settings file, and cobbler sync doesn't generate any warnings.
puppet_auto_setup: 1
sign_puppet_certs_automatically: 1
puppetca_path: "/usr/sbin/puppetca"
remove_old_puppet_certs_automatically: 1
my puppet master is on the same machine as my cobbler server. I don't see anything in the logs after december 16th indicating that cobbler has kicked off the puppetca tasks:
Fri Dec 16 15:35:09 2011 - INFO | running: ['/usr/sbin/puppetca', '--clean', 'XXXX']
Fri Dec 16 16:14:07 2011 - INFO | running: ['/usr/sbin/puppetca', '--sign', 'XXXX']
New machines brought online get their puppet cert added to the puppet master.
Any ideas where to look into this further? The only change I have made to the puppet_register_if_enabled snippet is that I have the snippet load the correct auth.conf and puppet.conf files. Reverting that change makes no difference in the cobbler->puppet interactions.
Thanks,
Jim Goddard
12 years, 4 months
Cobbler 2.2.1 not removing/signing puppet certs
by Jim Goddard
Hello all,
I am not positive that this happened in concurrence with my upgrading cobbler to 2.2.1 from the el5 2.0.11 version, but I haven't changed that much aside from that.
I was having problems before in that cobbler seemed to fire off the command to sign new puppet certs way too late, which I had never successfully figured out why. Now it is not ever firing that task.
Now, however, cobbler is not removing old certs either. I have the following puppet related settings in my /etc/cobbler/settings file, and cobbler sync doesn't generate any warnings.
puppet_auto_setup: 1
sign_puppet_certs_automatically: 1
puppetca_path: "/usr/sbin/puppetca"
remove_old_puppet_certs_automatically: 1
my puppet master is on the same machine as my cobbler server. I don't see anything in the logs after december 16th indicating that cobbler has kicked off the puppetca tasks:
Fri Dec 16 15:35:09 2011 - INFO | running: ['/usr/sbin/puppetca', '--clean', 'XXXX']
Fri Dec 16 16:14:07 2011 - INFO | running: ['/usr/sbin/puppetca', '--sign', 'XXXX']
New machines brought online get their puppet cert added to the puppet master.
Any ideas where to look into this further? The only change I have made to the puppet_register_if_enabled snippet is that I have the snippet load the correct auth.conf and puppet.conf files. Reverting that change makes no difference in the cobbler->puppet interactions.
Thanks,
Jim Goddard
-----Original Message-----
From: cobbler-bounces(a)lists.fedorahosted.org on behalf of cobbler-request(a)lists.fedorahosted.org
Sent: Fri 1/6/2012 4:00 AM
To: cobbler(a)lists.fedorahosted.org
Subject: cobbler Digest, Vol 43, Issue 3
Send cobbler mailing list submissions to
cobbler(a)lists.fedorahosted.org
To subscribe or unsubscribe via the World Wide Web, visit
https://fedorahosted.org/mailman/listinfo/cobbler
or, via email, send a message with subject or body 'help' to
cobbler-request(a)lists.fedorahosted.org
You can reach the person managing the list at
cobbler-owner(a)lists.fedorahosted.org
When replying, please edit your Subject line so it is more specific
than "Re: Contents of cobbler digest..."
Today's Topics:
1. cobbler 2.2 and empty subnet (Christian Horn)
2. Re: How does one set up dnsmasq/DHCP to play nice with
Windows DNS/DHCP (Ger Apeldoorn)
3. Re: How does one set up dnsmasq/DHCP to play nice with
Windows DNS/DHCP (David Lee)
----------------------------------------------------------------------
Message: 1
Date: Thu, 5 Jan 2012 15:15:30 +0100
From: Christian Horn <chorn(a)fluxcoil.net>
To: cobbler mailing list <cobbler(a)lists.fedorahosted.org>
Subject: cobbler 2.2 and empty subnet
Message-ID: <20120105141530.GC2540(a)fluxcoil.net>
Content-Type: text/plain; charset=us-ascii
Hi,
using a RHEL6.2 and the current cobble in epel, 2.0.11-2,
a system object can be added, and is handled properly.
Using cobbler 2.2.1-1 from testing leads to this:
# cobbler system add --name=tester6 --profile=rhel6u2-x86_64
# cobbler system edit --name=tester6 --hostname=tester6 \
--interface=eth0 --ip-address=10.0.100.6 \
--subnet=255.255.255.0 --mac=00:ff:06:00:00:01 --static=1
# cobbler system dumpvars --name=tester6 |grep subnet_
subnet_eth0 :
# cobbler system report --name=tester6 |grep Subnet
Subnet Mask : 255.255.255.0
Dumpvars is not presenting the netmask, also not 'getks'.
Performing the same steps with 2.0.11-2 everything works.
Accessing system objects that were created with 2.0.11-2
after upgrade to 2.2.1-1 leads to the subnet still beeing
accessable.
Any ideas on what I could possibly be missing here?
Did not spot a related ticket on github yet, got aware of
this from a report from another cobbler user.
cheers, Christian
------------------------------
Message: 2
Date: Fri, 06 Jan 2012 07:33:30 +0100
From: Ger Apeldoorn <g.apeldoorn(a)fundeon.nl>
To: cobbler mailing list <cobbler(a)lists.fedorahosted.org>
Subject: Re: How does one set up dnsmasq/DHCP to play nice with
Windows DNS/DHCP
Message-ID: <4F0695BA.1060900(a)fundeon.nl>
Content-Type: text/plain; charset=UTF-8; format=flowed
>
> you should be able to run dhcpd on your own subnet w/o causing
> problems to any other dhcpd servers on other subnets that service you.
> to be safe, you can also disable any dhcp helpers on your switch ports
> that service you subnet.
This is only true if your network is on a seperate vlan or physical
network. Multiple subnets can coexist on a single vlan, in which case
you would surely cause issues on the LAN.
You could also configure a dhcp server to respond only to known macadresses.
Ger
------------------------------
Message: 3
Date: Fri, 06 Jan 2012 10:14:24 +0000
From: David Lee <David.Lee(a)ecmwf.int>
To: cobbler mailing list <cobbler(a)lists.fedorahosted.org>
Subject: Re: How does one set up dnsmasq/DHCP to play nice with
Windows DNS/DHCP
Message-ID: <4F06C980.2020604(a)ecmwf.int>
Content-Type: text/plain; charset=UTF-8; format=flowed
Ger Apeldoorn wrote:
>>
>> you should be able to run dhcpd on your own subnet w/o causing
>> problems to any other dhcpd servers on other subnets that service you.
>> to be safe, you can also disable any dhcp helpers on your switch ports
>> that service you subnet.
>
> This is only true if your network is on a seperate vlan or physical
> network. Multiple subnets can coexist on a single vlan, in which case
> you would surely cause issues on the LAN.
>
> You could also configure a dhcp server to respond only to known
> macadresses.
> [...]
For the future, we definitely need to ensure that the various aspects of
this are addressed in the documentation, as both principle and example.
The documentation is moving to a wiki, much of whose content originated
from the existing big man page. The main page regarding DHCP is:
https://github.com/cobbler/cobbler/wiki/DHCP%20management
I don't see any mention there of the considerations raised in this thread.
Dan, Pete, and Ger (as technical contributors to the thread): If you
don't wish to update the wiki yourselves, I'm happy to try to
consolidate any suggestions you may make or send to me.
-- David Lee
------------------------------
_______________________________________________
cobbler mailing list
cobbler(a)lists.fedorahosted.org
https://fedorahosted.org/mailman/listinfo/cobbler
End of cobbler Digest, Vol 43, Issue 3
**************************************
12 years, 4 months
snippet without LVM (CentOS 6)
by Frank
Merry Christmas to you All:
Someone on this list helped me over a year ago to set up Cobbler ...
now version 2.0.11.
I do have the following question:
As it is set up now, Cobbler installs CentOS 5 or 6 on servers using
LVM, Cluster Logical Volume Manager (see here:
http://www.centos.org/docs/5/html/Cluster_Logical_Volume_Manager/).
That has created lots of problems for me when e.g. I try to replace
one of two disks in a server (loosing data of one of the disks). It
is really complicated. It would therefore be VERY helpful if one of
you could be so very kind to post a full sample "partition_select"
snippet for a server with two disks with a standard setup -- not LVM.
Here is how then now looks presently in my Cobbler install:
(1)
/var/lib/cobbler/snippets/partition_select
---------
%include /tmp/partinfo
%pre
# Determine how many drives we have
set \$(list-harddrives)
let numd=\$#/2
d1=\$1
d2=\$3
#set $vgname = "VolGroup00"
#set $vgbackupname = "VolGroup01"
if [ ! -z \$d1 ]
then
cat << EOF > /tmp/partinfo
part /boot --fstype ext4 --size=140 --ondisk=\$d1
part pv.1 --size=1024 --grow --ondisk=\$d1
part pv.2 --size=1024 --grow --ondisk=\$d2
volgroup $vgname pv.1
volgroup $vgbackupname pv.2
logvol /tmp --fstype ext4 --name=tmp --vgname=$vgname --size=4096
logvol swap --fstype swap --name=swap --vgname=$vgname --size=1024
--grow --maxsize=4096
logvol /usr --fstype ext4 --name=usr --vgname=$vgname --size=18432
logvol / --fstype ext4 --name=root --vgname=$vgname --size=20480 --grow
logvol /backup --fstype ext4 --name=backup --vgname=$vgbackupname
--size=20480 --grow
EOF
else
cat << EOF > /tmp/partinfo
part /boot --fstype ext4 --size=140 --ondisk=\$d1
part pv.1 --size=1024 --grow --ondisk=\$d1
volgroup $vgname pv.1
logvol /tmp --fstype ext4 --name=tmp --vgname=$vgname --size=4096
logvol swap --fstype swap --name=swap --vgname=$vgname --size=1024
--grow --maxsize=4096
logvol /usr --fstype ext4 --name=usr --vgname=$vgname --size=18432
logvol / --fstype ext4 --name=root --vgname=$vgname --size=20480 --grow
EOF
fi
---------
And the kickstarts file for CentOS 6 looks like this:
(2)
/var/lib/cobbler/kickstarts/centos6.ks
---------
url --url=$tree
lang en_US.UTF-8
keyboard us
reboot
skipx
text
network --device eth0 --bootproto dhcp
rootpw --iscrypted $default_password_crypted
firewall --disabled
authconfig --enableshadow --enablemd5
selinux --disabled
timezone Europe/Berlin
zerombr
bootloader --location=mbr --append="rhgb quiet"
clearpart --all --initlabel
$SNIPPET('partition_select6')
%pre
$kickstart_start
$SNIPPET('log_ks_pre')
$SNIPPET('pre_install_network_config')
$SNIPPET('pre_anamon')
%packages --nobase
coreutils
yum
rpm
e2fsprogs
lvm2
sysstat
ntp
openssh-server
openssh-clients
wget
authconfig
sudo
man
man-pages
nano
rdate
%post
$SNIPPET('post_install_network_config')
$SNIPPET('post_install_kernel_options')
$SNIPPET('post_anamon')
$SNIPPET('log_ks_post')
$SNIPPET('cobbler_register')
$kickstart_done
----------
THANK YOU !
Frank
12 years, 4 months