Hey Scott, Yes, you simply place the resulting directory in the root of an EC2 Volume. Unmount the volume from the host instance, snapshot it, and convert the snapshot to an AMI.
When you convert the snapshot to an AMI, make sure you use the PV-GRUB AKI (can be found here: http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/UserProvidedKernels.html). You can see if your unikernel is working by checking the EC2 instance's system logs. I've been running my unikernels on PV AMIs + Instances. I haven't tried HVM yet, let me know if you do. On Wed, Jan 27, 2016 at 8:46 AM Scott Weiss <[email protected]> wrote: > untar to where? and then i should reboot the instance? how do i get the > instance to boot from the unikernel? > > On Wed, Jan 27, 2016 at 8:44 AM, Antti Kantee <[email protected]> wrote: > >> On 27/01/16 13:28, Scott Weiss wrote: >> >>> I have tried creating an AMI from my rump kernel binary according to >>> >>> https://www.freelists.org/post/rumpkernel-users/Amazon-EC2-support-now-in-Rumprun >>> , >>> without success. I posted an issue to github detailing the commands i am >>> running: https://github.com/rumpkernel/rumprun/issues/68 >>> >>> I received the reply that I should just follow the instructions: Place >>> contents of "rumprun-testprog.bin.ec2dir" to EC2 volume and boot. >>> I'm unclear what these instructions indicate specifically. Does this >>> mean I >>> should attach an empty EBS volume to a running instance, place the >>> contents >>> of rumprun-testprog.bin.ec2dir in that volume, then reboot the instance? >>> >> >> From the post you quote, those instructions mean: >> >> "Tar up the directory created above and copy the tarball to the Linux >> instance" >> >> And then do the obvious thing at the "+ tar" step. (where tar = untar) >> >> >
