RHEL/CentOS Kickstart Workflow – Optimization Tips

I'm working in an environment that consists of CentOS systems located in multiple data centers and office locations. I've been using a reasonable kickstart with a solid post-install script, but want to streamline and expand the process. We have our own yum repo and I have control over the range of hardware in use (it's usually HP ProLiant hardware or VMWare) Let's say I build or redeploy one server every week. Following the steps needed for hardware preparation, I begin the OS setup. The issues I have now are:

1). A nice way of booting systems into the network installer. I've been using the 10mb boot.iso/netinstall.iso available in CentOS distribution. I mount this either via VMWare vCenter (for VMs) or HP ILO (usually over SSH, pointing to the URL of the ISO) to kick off the installations. I can't use PXE in most cases because the data center locations do not have DHCP available. Is there a clean way to get the boot.iso onto a USB key, for instance?

2). Sometimes I need to set a static IP for the kickstart. Is there a better way to input that info other than typing it on the boot: command line at the installer prompt?

3). A good method for customizing the kickstart. Right now, I have most of what I need, but sometimes need to alter the partition scheme to account for different RAID setups (e.g. two SmartArray controllers). Right now, I'm basically creating a separate kickstart file for each system (or group of system's) settings. I end up with a directory full of old *.cfg files. I know there's a more elegant way of doing this. PHP perhaps?

4). Help with the install method. As-is, the full build from prompt to completion takes 5-10 minutes. I've noticed that the install method matters, depending on location. HTTP works well local to the yum server, but is poor over our high-bandwidth low-latency connections to other sites. NFS works better in those cases. I'd appreciate any nifty tricks used to preload settings or system parameters before the kickstart process.

5). The post-install script works well. It pulls additional packages, sets some initial settings, trims the services list down and populates a few users/passwords/SSH keys. I'd like to get this environment running with a proper configuration management (I'm used to CFEngine, but think that Puppet or Bcfg2 are probably a better choice nowadays). Do I pare-down the post-install script and move those functions into CFEngine? That's probably a different question, though.

Here's a an example of the kickstart file I've been using:

install
text
url --url http://yum.abc.com/5.6/os/x86_64/
#nfs --server=yum.abc.com --dir=/yum/5/os/x86_64
lang en_US.UTF-8
langsupport --default=en_US.UTF-8 en_US.UTF-8
keyboard us
network --device eth0 --bootproto dhcp
rootpw --iscrypted $encryptedpassword
firewall --disabled
selinux --disabled
authconfig --enableshadow --enablemd5
timezone --utc America/Chicago
bootloader --location=mbr
clearpart --all --initlabel
part /boot --fstype ext3 --size=200
part /usr --fstype ext3 --size=8192 --asprimary
part / --fstype ext3 --size=12288 --asprimary
part /var --fstype ext3 --size=4096
part swap --size=8192
part /tmp --fstype ext3 --size=2048
#part /opt --fstype ext3 --size=100 --grow
#part /scratch --fstype ext3 --size=61440

%packages
@ admin-tools
@ editors
@ system-tools
@ network-server
@ mail-server
@ server-cfg
@ development-tools
yum-fastestmirror
rpm-devel
e2fsprogs
grub
kernel-devel
net-snmp-utils
-subversion
-xdelta

%post
exec < /dev/tty3 > /dev/tty3
chvt 3
echo
echo "################################"
echo "# Running Post Configuration   #"
echo "################################"
(
/usr/bin/wget http://ks.abc.com/post/post-install-abc.sh -O /root/post-install-abc.sh
/bin/chmod 755 /root/post-install-abc.sh
/root/post-install-abc.sh
) 2>&1 | /usr/bin/tee /var/log/post-install.log
chvt 1