[ubuntu-cloud] Upgrading Old Intrepid EC2 instance

Scott Moser smoser at ubuntu.com
Fri Oct 1 20:05:09 BST 2010

On Wed, 29 Sep 2010, Simon Detheridge wrote:

> Hello,
> I am running some heavily modified instances based on the original canonical Intrepid images from quite a while ago. (Yes, I know they are long overdue for maintenance)
> I have converted my instance-store backed images to ebs-backed images, and can boot them up quite happily from snapshots. What I was hoping to do was:
> - Upgrade to Jaunty, snapshot disk, create new image based on that snapshot with updated AKI/ARI ids, start back up
> - Repeat for Karmic
> - Repeat for Lucid
> I fell at the first hurdle. do-release-upgrade completed without error. Snapshotted the disk and attempted to launch a new instance with kernel aki-adb69ed9 and ramdisk ari-a9b69edd (both 2.6.28-12-xen-i386)
> After launch, nothing happens. No output to the console that I can see with elasticfox, and ssh is not running.
> My questions: Is what I'm trying to do even possible, and if not why? --

Well, anything is possible.  It just might take alot of time.  I
personally have never tired Intrepid -> Jaunty (which wasn't ever really
supported) -> Karmic -> Lucid on ec2.

I don't know why you wouldn't be getting any console output when you launc
the instance, but I would suspect that something in /etc/fstab isn't
right.  Getting console output is really going to be required for any idea
of what might be going wrong.  I would also suggest jumping to a karmic
kernel if possible, skipping the jaunty one.

> I don't really want to have to go through the process of setting up the
> instances from scratch again. It will take a long time.

What you're asking *can* be done... your approach is sound, but its just
undocumented waters on ec2 and that old.  once you were to karmic, i would
not expect too many hurdles.


More information about the Ubuntu-cloud mailing list