On Wed, 22 Sep 2010 12:27:50 -0400, Joseph Boyer Jr. <jboyer(a)liquidnet.com>
wrote:
I have recently configured cobbler for ESX/ESXi 4.x installs. Yes, it
was
a
real pain in the arse since vmware has done away with anaconda and
now
uses
its own installer weasel. The main issue I had with PXE installs is
that
after PXE and during the initialization of the VMware boot kernel the
second DHCP request will fail because the install is using vswif0, which
has a dynamically generated MAC, hence DHCP is impossible. The simple
solution was to append the IP/NETMASK/GATEWAY with PXE config to by-pass
the second DHCP and having to play with IPEND crap as well.
Here is what my pxe config looks like
default linux
prompt 0
timeout 1
label linux
kernel /images/vmware4_64_402/vmlinuz-esx402
append ksdevice=F4:CE:46:B3:1E:B0 noipv6 netmask=255.255.255.0
syslog=10.0.254.169:25150 ip=10.0.254.3 mem=800M text kssendmac
gateway=10.0.254.1
initrd=/images/vmware4_64_402/initrd-esx402.img
ks=http://10.0.254.169/cblr/svc/op/ks/system/ny0eesx01
This works beautifully, I have VMware admins adding the
ip/gateway/network
to the kernel options for the system.
ESXi is actually it's own beast, you're just doing a straight up ESX 4.x
install (we do that as well here, and ran into the random MAC issue
ourselves). The solution to the random mac problem is, don't use vswif0 :)
Use vmnic0 instead, which is what we do and it works great with the static
IP config.
--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.