snippet keep_ssh_host_keys not running in background
by Detlev Habicht
Hi all,
i am testing the snippet keep_ssh_host_keys.
Well, i am a beginner in cobbler and maybe my problem is a bash problem, but
when i use keep_ssh_host_keys i have two problems:
When i use keep_ssh_host_keys as snippet, i see while kickstart is running:
Unknown command: <!DOCTYPE
Why?
Well, than i put the whole keep_ssh_host_keys script in my kickstart file
and now the script is running until the last part (i can see, the keys are
saved in /tmp):
# Loop until the corresponding rpm is installed
if [ "$keys_found" = "yes" ]; then
while : ; do
sleep 10
if [ -d /mnt/sysimage$SEARCHDIR ] ; then
cp -af /tmp/$TEMPDIR/${PATTERN}* /mnt/sysimage$SEARCHDIR
logger "keys copied to newly installed system"
break
fi
done &
fi
The kickstart process stopped and i think the loop is running, but not
in the background, so the installation process cannot proceed.
Why?
Thanx for any help!
Detlev
--
Detlev | Institut fuer Mikroelektronische Systeme
Habicht | D-30167 Hannover +49 511 76219662 habicht(a)ims.uni-hannover.de
--------+-------- Handy +49 172 5415752 ---------------------------
12 years, 5 months
Moving tickets over to github, closing Trac tickets
by Michael DeHaan
It seems we have a fair amount of tickets left in Fedora Hosted.
Naturally I'd like to squash as many of these as I can… but…
Since some are opened on old versions, and it would take a tremendous amount of time to validate all of them, I'm planning on closing basically everything in Trac with a note that if the issue still affects you, to open a new issue in github
(with a link to where to go).
This is happening because there's a lot in there that is over a year old, and also, it spreads out the work of validation whether something is still a problem among a very large range of people.
This will probably have the added benefit of making lots of folks sign up for github accounts :)
--Michael
12 years, 5 months
static network configuration
by Kelly Goedert
Hi,
on my kickstart template I have this
# Static network configuration.
d-i netcfg/get_nameservers string 10.1.1.249
d-i netcfg/get_ipaddress string $static_ip
d-i netcfg/get_netmask string 255.0.0.0
d-i netcfg/get_gateway string 10.1.1.11
d-i netcfg/confirm_static boolean true
The static_ip value I have set on system -> kickstart metadata like
static_ip=10.1.1.8
But this config has no effect on the installation. What is wrong on this
configuration? I am trying to install ubuntu 11.10
Thanks
Kelly
12 years, 5 months
keyboard layouts - configuration
by Kelly Goedert
Hi,
on my preseed file for automatic installation I'm trying to set my keyboard
layout like this
d-i console-setup/ask_detect boolean false
d-i keyboard-configuration/layoutcode string br
d-i keyboard-configuration/modelcode abnt2
But my keyboard layout is still set as us. An the first time the machine
runs, I get an incomplete language support error. The distro I am trying to
install is a ubuntu one.
Can anybody help?
Thanks
Kelly
12 years, 5 months
cobbler documentation
by David Lee
I'd like to suggest that we review, for the long term, the aims and
objectives of the cobbler documentation and how it should be structured
to meet them.
We currently have a huge man page, whose structure is inconsistent with
the usual structure of UN*X man pages. (As one trivial example the
section "See Also" would normally be at the end, but we have it half-way
down the first page of a 17-page (troff-style) document.)
The man page also tries to serve multiple purposes, with:
(a) a lot of reference-like detail
(b) several examples of cookboook/workflow
(c) several "advanced topics", but no apparent pattern.
We also have the wiki (recently moved from fedorahosted to github) which
seems to have lots of very useful user-oriented information on
particular topics, with cookbook and workflow items.
We could discuss the details of any of those. But I suggest that, for
the moment, we concentrate on the bigger picture, avoiding detail except
by way of example.
Usually in UN*X documentation, small topics such as the "cat" command
use the man page mainly reference information about the details, and
have sufficient examples to cover cookbook and workflow purposes. The
topic is small enough for the "man" page to be all-purpose.
Big topics ("make", "rcs", "pam" etc.) tend to limit the man page to
reference information about command options. Separately from the man
page, the fuller information, such as user guides and substantial
examples, is in other, non-man, documentation. These big topics
generally avoid overloading their man page.
One other make-like technique sometimes used for big topics is to have
related man pages, not necessarily directly connected to particular
commands. For instance "rcs" also has an "rcsintro" man page (but no
"rcsintro" command). Going further, the "PAM" subsystem has a man page
for each PAM module, even though there are no PAM commands (including
"pam" itself).
From this information perspective, Cobbler must surely classify as a
big topic (like "make") rather than a small topic (like "cat").
So I'd like to suggest that we review the structure of Cobbler
documentation.
Whatever happens, it can be argued that the man page called "cobbler"
should end up considerably smaller, providing a summary and outline of
what cobbler does, but avoiding the geek temptation to include every
last detail of every last advanced topic.
The details of particular subcommands ("distro", "system", etc.) could
be farmed out to additional man pages, perhaps called "cobbler-distro",
etc. But these, too, should probably avoid the "include every last
detail" temptation.
To complement this, in these days of Internet connectivity being normal,
we can make full use of the wiki, suitably and liberally referenced from
the man page. While the main "cobbler" man page could perhaps give some
deliberately simple "workflow" examples, we should probably make much
fuller use of the wiki for providing fuller examples of this nature.
Almost certainly the wiki, not the man page, should be the prime
repository for "advanced topics".
I realise that Cobbler, like much open source software, is supported by
volunteer effort, so that time and effort are severely limited. That's
why my suggestions above stay with the existing technology and
documentation structures. It is also why I haven't suggested anything
totally different, such as a writing a reference manual (as is the case
with, for instance, "make") as that would be a major effort.
Thoughts?
--
: David Lee
: ECMWF (Data Handling System)
: Shinfield Park
: Reading RG2 9AX
: Berkshire
:
: tel: +44-118-9499 362
: email: david.lee(a)ecmwf.int
12 years, 5 months
Cobbler repositories --priorities problem
by Will McDonald
Hi List.
I wonder if anyone's seen this problem or can explain what's going on. I've
updated a couple of repository's priorities using:
[root@buildhost ~]# cobbler repo edit --name=$repo --priorities=N
When I check the repo that I've altered it LOOKS like the edit has
succeeded, but when I look at the .repo Cobbler's dynamically building,
it's still pegged at '99'
[root@buildhost ~]# cobbler profile report
--name=centos56-x86_64-oradb11202-wls1034 | grep Repo
Repos : ['centos56-repo-x86_64']
[root@buildhost ~]# cobbler repo report --name=centos56-repo-x86_64 | grep
Prio
Priority : 1
[root@buildhost ~]# wget -q "
http://192.168.0.102/cblr/svc/op/yum/profile/centos56-x86_64-oradb11202-w..."
-O /tmp/test.repo
[root@buildhost ~]# cat /tmp/test.repo
[centos56-repo-x86_64]
name=centos56-repo-x86_64
baseurl=http://192.168.0.102/cobbler/repo_mirror/centos56-repo-x86_64
enabled=1
priority=99
gpgcheck=0
This problem seems to manifest across all repos and associated profiles.
The value in the JSON for each repo is being updated, it's just the
dynamically built repo that isn't.
I've tried 'cobbler sync|reposync' and 'service cobblerd restart' the
problem remains the same. If it's of any bearing, the repos are hardlinked
mirrors of the corresponding kickstart trees.
I've attached a full 'cobbler report' too in case that's of any use. Am I
just doing something fundamentally wrong? This is cobbler-2.0.11-2.el5 from
EPEL on CentOS 5.7.
12 years, 5 months
centos 5.7 partition frustrations
by Jim Goddard
I have seen a couple messages here and around the net around getting the cobbler partition selection to work, but I am having a devil of a time.
I have a kickstart that I have converted to use in cobbler. I have slightly modified the pre_partition_select snippet that comes with cobbler 2.0.11. It is working fine under centos 6, but I cannot get it to work on centos 5. Under centos 5 kickstarting it appeared that the /tmp/partinfo file is never created or imported.
I saw some a tip about the python path possibly not getting set within anaconda, so I changed the 'set \$(list-harddrives)' line to 'set \$(PYTHONPATH=/usr/lib/booty list-harddrives)', which didn't help.
When kickstarting centos 5, anaconda never seems to get the PART= lines and prompts me to decide on partitions. Needless to say this invalidates the whole unattended setup paradigm :).
With centos 6, it just seemed to work.
As of today, I am getting a different set of errors. Now I am getting all sorts of weird errors on centos 5 & 6.
When I run cobbler validateks, all my EL6 profiles and systems check out fine, but all the 5 ones error with:
checking url: http://10.74.67.42/cblr/svc/op/ks/profile/centos5_32
running: /usr/bin/ksvalidator "http://10.74.67.42/cblr/svc/op/ks/profile/centos5_32"
received on stdout: The following problem occurred on line 52 of the kickstart file:
Unknown command: set
The following problem occurred on line 53 of the kickstart file:
Unknown command: let
The following problem occurred on line 54 of the kickstart file:
etc, about 50 of these per kickstart. When I actually tried to kickstart a test machine with that profile, anaconda bailed at the same line, with the same error "Unknown command: set"
I have two questions. 1) is there any way to get better visibility into the kickstart/anaconda process? It is very frustrating to try and debug this process by manually working through the kickstart (when possible) and then copying the log files over to another machine to look at, not to mention the 10+ minutes these servers take to reboot.
2) Are people generally using something else for partition selection? Our environment doesn't really have that many variables around partitioning. I set a ksmeta "isvirt" so that vms get a different partition layout than physical servers, and I need a good way to select the first /dev/cciss/ drive for HP servers, and ideally a way to exclude SAN volumes so I don't accidentally reformat a SAN database when rebuilding a box (which has happened, luckily a SAN volume that was due to be reclaimed anyway and was unneeded). I saw some messages on the redhat-kickstart list about using /proc/partitions to do this perhaps?
Sorry about the length of this email, trying to be thorough. For anybody still reading, here is the snippet, followed by the relevant portions of my kickstart templates.
# partition details calculation
# Determine how many drives we have
set \$(PYTHONPATH=/usr/lib/booty list-harddrives)
let numd=\$#/2
d1=\$1
d2=\$3
# want to know exactly what the list-harddrives outputs
# from the stub code it is a sorted list of <name> <size> of all drives.
# echo \$(list-harddrives)
# Get the ksmeta variable, if passed in. If this is a virtual, we want to set up the partitions correctly.
#set $isvirt = $getVar('isvirt', None)
# Determine architecture-specific partitioning needs
EFI_PART=""
PPC_PREP_PART=""
BOOT_PART=""
ROOT_PART=""
SWAP_PART=""
VAR_PART=""
case \$(uname -m) in
ia64)
EFI_PART="part /boot/efi --fstype vfat --size 200 --recommended --ondisk=\$d1"
;;
ppc*)
PPC_PREP_PART="part None --fstype 'PPC PReP Boot' --size 8"
BOOT_PART="part /boot --fstype ext3 --size 200 --recommended --ondisk=\$d1"
;;
*)
BOOT_PART="part /boot --fstype ext3 --size 200 --recommended --ondisk=\$d1"
;;
esac
# Now set our partitions based on whether this is a vm or not
#if $isvirt == "true"
ROOT_PART="part / --asprimary --fstype ext3 --fsoptions=\"defaults,noatime\" --size=5000 --ondisk=\$d1 --recommended"
SWAP_PART="part swap --asprimary --fstype swap --size=1000 --ondisk=\$d1 --recommended"
VAR_PART="part /var --asprimary --fstype ext3 --fsoptions=\"defaults,noatime\" --size=8000 --grow --ondisk=\$d1"
#else
ROOT_PART="part / --asprimary --fstype ext3 --fsoptions=\"defaults,noatime\" --size=12000 --grow --ondisk=\$d1 --recommended"
SWAP_PART="part swap --asprimary --fstype swap --size=8192 --ondisk=\$d1 --recommended"
#end if
cat <<EOF > /tmp/partinfo
\$EFI_PART
\$PPC_PREP_PART
\$BOOT_PART
\$ROOT_PART
\$VAR_PART
\$SWAP_PART
# This should be the last line of pre_partition_info_gmi cobbler snippet
EOF
# End pre_partition_select_gmi /tmp/partinfo and /root/partinfo should hold instructions
Kickstart template:
# kickstart template for el5 OSes
# partinfo not currently working.
#platform=x86, AMD64, or Intel EM64T
# System authorization information
auth --useshadow --enablemd5
# System bootloader configuration
bootloader --location=mbr
# Partition clearing information
clearpart --all --initlabel
# Use text mode install
text
# Firewall configuration
firewall --disable
# Run the Setup Agent on first boot
firstboot --disable
# System keyboard
keyboard us
# System language
lang en_US
# Use network installation
url --url=$tree
# If any cobbler repo definitions were referenced in the kickstart profile, include them here.
$yum_repo_stanza
# Network information
$SNIPPET('network_config')
# Reboot after installation
reboot
#Root password
rootpw --iscrypted $default_password_crypted
# SELinux configuration
selinux --disabled
# Do not configure the X Window System
skipx
# System timezone
timezone --utc America/Vancouver
# Install OS instead of upgrade
install
# Clear the Master Boot Record
zerombr
# Allow anaconda to partition the system as needed
$SNIPPET('pre_partition_select_gmi')
%pre
$SNIPPET('log_ks_pre')
$kickstart_start
$SNIPPET('pre_install_network_config')
$SNIPPET('main_partition_select')
# Enable installation monitoring
#$SNIPPET('pre_anamon')
%end
%packages --excludedocs
$SNIPPET('func_install_if_enabled')
$SNIPPET('puppet_install_if_enabled')
$SNIPPET('koan_install_if_virt')
@legacy-software-support
@system-tools
vim-enhanced
%post
$SNIPPET('log_ks_post')
# Start yum configuration
$yum_config_stanza
# End yum configuration
$SNIPPET('post_install_kernel_options')
$SNIPPET('post_install_network_config')
$SNIPPET('func_register_if_enabled')
# $SNIPPET('download_config_files')
$SNIPPET('koan_environment')
$SNIPPET('redhat_register')
$SNIPPET('cobbler_register')
# Enable post-install boot notification
# $SNIPPET('post_anamon')
# Start final steps
$SNIPPET('post_housecleaning_gmi')
$SNIPPET('puppet_register_if_enabled')
$SNIPPET('post_puppet_initial_runs')
$kickstart_done
# End final steps
Pretty much the only differences between my centos 5 & 6 kickstart templates is the removal of the %end tags:
diff gmi_base.ks gmi_base_centos5.ks
1,3c1,2
< # kickstart template for Fedora 8 and later.
< # (includes %end blocks)
< # do not use with earlier distros
---
> # kickstart template for el5 OSes
> # partinfo not currently working.
44c43
<
---
> $SNIPPET('pre_partition_select_gmi')
49c48,49
< $SNIPPET('pre_partition_select_gmi')
---
> $SNIPPET('main_partition_select')
>
54,55c54
< $SNIPPET('main_partition_select')
< %packages --excludedocs --ignoremissing
---
> %packages --excludedocs
142d140
< %end
165d162
< %end
Thanks so much for any assistance with this.
Jim Goddard
12 years, 5 months
[ Cobbler ] Problems after updating cobbler 1.6 to cobber 2.0.11 and cobbler-web
by Peter Mumenthaler
Hi All,
I have following problems after updating cobbler from version 1.6 to
version 2.0.11-2 running on CentOS 5.7 and running httpd 2.2.3-53
I installed cobbler and cobbler-web. If i connect to the cobbler web
site i get constant info messages in httpd/error.log as follows:
[Fri Dec 16 21:49:55 2011] [notice] child pid 12290 exit signal
Segmentation fault (11)
[Fri Dec 16 21:49:56 2011] [notice] child pid 12293 exit signal
Segmentation fault (11)
[Fri Dec 16 21:49:56 2011] [notice] child pid 12294 exit signal
Segmentation fault (11)
Also i get an url not found error if i try to view a kickstart file in
on the cobbler web page.
url /cblr/svc/op/ks/system/foobarsystem not found
cobbler sync runs without any failures and cobbler check gives me
following output:
1 : service named is not running
2 : change 'disable' to 'no' in /etc/xinetd.d/rsync
3 : since iptables may be running, ensure 69, 80, and 25151 are unblocked
4 : debmirror package is not installed, it will be required to manage
debian deployments and repositories
5 : One or more repos referenced by profile objects is no longer defined
in cobbler: fedora.13.i386.os, fedora.13.i386.puzzle,
fedora.13.i386.updates, fedora.13.x86_64.os, fedora.13.x86_64,
fedora.13.x86_64.updates, fedora.13.i386.os, fedora.13.i386.,
fedora.13.i386.updates, fedora.13.x86_64.os, fedora.13.x86_64.,
fedora.13.x86_64.updates, fedora.13.x86_64.os, fedora.13.x86_64.,
fedora.13.x86_64.updates
6 : fencing tools were not found, and are required to use the (optional)
power management features. install cman or fence-agents to use them
Restart cobblerd and then run 'cobbler sync' to apply changes.
cobbler.conf and cobbler_web.conf are configured as the come by default
with the rpm package.
Anyone an idea?
Thanks for any help.
Kind regards
peter
--
Peter Mumenthaler
Linux System-Ingenieur
Puzzle ITC GmbH
www.puzzle.ch
Telefon +41 31 370 22 00
Direkt +41 31 370 22 34
Fax +41 31 370 22 01
Werfen Sie einen Blick in unseren Blog:
<http://www.puzzle.ch/blog>
12 years, 5 months
redhat "rescue mode"?
by David Lee
(Using cobbler 2.0.11 with RHEL 5.6.)
I may have missed the documentation on this: if so, simply point me in
the right direction!
During the last few months we have developed a cobbler installation for
installing new client systems, and are happy with it. Indeed, most of
our boxes don't have internal DVD drives, so cobbler has been essential.
But one thing we would like to be able to do occasionally is to boot a
client box into a "rescue mode". The most recent example was when we
had some suspect memory on a disk controller.
Is there some sort of recipe somewhere for running a "rescue mode" boot
of a cobbler client system?
--
: David Lee
: ECMWF (Data Handling System)
: Shinfield Park
: Reading RG2 9AX
: Berkshire
:
: tel: +44-118-9499 362
: email: david.lee(a)ecmwf.int
12 years, 5 months
Issues starting cobblerd after an upgrade to 2.2.1-1.el5
by Phil Dougherty
Hello,
I am trying to upgrade cobbler to the 2.2.1-1.el5 version from epel-testing because I would like to have support to import ubuntu ISO's.
When I try to start cobbler I am getting the following error message. Could this be related to my crusty old python 2.4 version perhaps?
[root@cobbler1 cobbler]# /etc/init.d/cobblerd start
Starting cobbler daemon: Traceback (most recent call last):
File "/usr/bin/cobblerd", line 76, in main
api = cobbler_api.BootAPI(is_cobblerd=True)
File "/usr/lib/python2.4/site-packages/cobbler/api.py", line 130, in __init__
self.deserialize()
File "/usr/lib/python2.4/site-packages/cobbler/api.py", line 873, in deserialize
return self._config.deserialize()
File "/usr/lib/python2.4/site-packages/cobbler/config.py", line 266, in deserialize
raise CX("serializer: error loading collection %s. Check /etc/cobbler/modules.conf" % item.collection_type())
CX: 'serializer: error loading collection distro. Check /etc/cobbler/modules.conf'
[ OK ]
[root@cobbler1 cobbler]# ps auxw|grep cobbler
avahi 1428 0.0 0.0 22508 1132 ? Ss Dec06 0:00 avahi-daemon: running [cobbler1.local]
root 11866 0.0 0.0 61176 748 pts/0 S+ 13:45 0:00 grep cobbler
Operating system is CentOS 5.5 64bit.
The only post that I have found that looks to be related is this one. https://fedorahosted.org/pipermail/cobbler-devel/2011-July/002275.html
Is there another version that is perhaps more stable that includes the support I am looking for?
I appreciate any help that can be provided.
--
Phil Dougherty
Sent with Sparrow (http://bit.ly/sigsprw)
12 years, 5 months