PXE booting a scripted ESXi 5.x Install

As of late I’ve been working on an automated deployment solution for ESXi.  Seeing as I will be deploying 50 + identical installations in the near future one thing I don’t want to have to do is follow documentation and screenshots telling me where and when to click.  So, enter the scripted install…and hey, why not throw PXE into the mix as well.

The official VMware documentation does have section dealing with PXE booting a scripted install, but it seems to assume that the VMware installation is going to be the only application included in your PXE boot and TFTP directories.  Now I don’t know about you, but the VMware installation is not the only thing that I want to boot from my PXE server.  Most people will have a menu built with various installation and utilities….so this is a few things I’ve learned over the past week that are not in the documentation…

So first off is the actual menu item in the the pxelinux config file.  Below is the actual KERNEL and APPEND items from the VMware documentation (which won’t work unless you have extracted everything into the root of your TFTP server.  As you can see there is no path preceding mboot, thus assuming it’s in the root.

Since we are placing our install files inside a utils folder, we need to change this to read as follows.

Also you can see that I have added the ks=location of ks.cfg command to boot.cfg as well.  This is what allows us to do the scripted installation.  I’m not going to delve into all of the ks options that are out there, if you want to you can go to the documentation and check it out for yourself.  The point of this post is just to get you up and running with a scripted pxe install when not residing in the root directory.

We are almost done now, just one more step.  The problem of pointing to the install outside of the root directory is solved, however the actual boot.cfg (located in the root of the install files) also assumes it is in the root directory, so we now have to go and tweak that a bit.  Again, the boot.cfg in its unmodified state is below…

So there are a few minor tweaks we need to do with this.  First off, we need to add the prefix label and define where our installation files are actually sitting.  Secondly, we need to remove every single slash that is preceding every module.  This will allow our installation to load and boot as directed from a subfolder within the root directory of our TFTP server.  The modified boot.cfg is shown below…

So there you go, as long as your pxe environment is set up properly you should be able to just boot into the pxe menu, select your ESXi installation label and install ESXi as normal or if you have added the directive for a kickstart installation the scripted install will begin.  If you have any thoughts, questions, comments, concerns or even suggestions on how to do this a better way please let me know in the comments below 😉