I recently had a need to deploy a t2.micro instance on EC2 running Centos. Unfortunately, there are no official Centos AMIs available that will run on the newer HVM instance types.
The AWS marketplace has several 3rd party Centos AMIs that support HVM. I used one of these as a basis for the new install.
Centos has the ability to boot up into a VNC server from which a network install can be done. Using this facility, I was able to create the custom install as follows:
- Launch HVM Centos instance (using AMI from marketplace)
- Insert into grub a new entry which will boot the VNC server
- Install new grub and reboot
- Connect via VNC to TCP port
5901and proceed with normal Centos install
- ssh to new install and comment out the mac address
- Create an image from the instance
The grub entry I used is as follows:
title Centos Install (PXE) root (hd0,0) kernel /boot/vmlinuz.cent.pxe vnc vncpassword=xxxxxxxxxxxx headless ip=dhcp ksdevice=eth0 method=http://mirror.centos.org/centos-6/6.5/os/x86_64/ lang=en_US keymap=us xen_blkfront.sda_is_xvda=1 initrd /boot/initrd.img.cent.pxe
This needs to go before your existing boot entry. Replace
vncpassword value with your own. Note the
xen_blkfront.sda_is_xvda=1 - this is required so the Centos installer can map the correct device name for your block device.
To apply the new config, run the following commands:
# grub grub> device (hd0) /dev/xvda grub> root (hd0,0) grub> setup (hd0) grub> quit
Thanks to this post on Sentris.com for the PXE boot idea.