EC2 instance terminated before starting up. No console log
I have created a new AMI based on an existing Fedora Core 13 AMI. When I try to launch an instance with new image, the instance gets terminated even before launching. The status of the instance changes from "pending" to "terminated." I tried to get the system log with Amazon Console, but the log is empty.
How can I find out the cause of instance termina开发者_开发知识库tion?
Edit: Please check my answer below.
The problem occurred due to a bug in AMI bundling utility, which in turn was caused by changes in the output of openssl in newer version. The solution is described in EC2 forum. My keyword searches didn't show up this answer. I stumbled on this answer by searching for only message available for the instance. The message was
Client.InstanceInitiatedShutdown: Instance initiated shutdown
The fix is a simple change in the ruby script for image bundling. The details are at the forum post linked above.
What you describe either sounds like an issue with the AMI or a capacity issue of some sort. E.g., I noticed that I cannot start one of the high mem instances in us-east-1b
, it had a similar effect.
As for the AWS console, I'm not aware of an error log of some sort, I'd try to start an instance from the shell using the ec2 cli tools and see if more errors are returned:
ec2-run-instances YOURAMI -n 1 -g group -k key
The above should return an instance identifier (i-xxxxxx
) and pending.
Then use ec2-describe-instances i-xxxxxx
to see the status of an instance. You could also try to skip ec2-run-instances
and to use the describe command right away. Should work just as well.
精彩评论