0

I have an auto-scaling group which was working fine. I made a copy of the launch configuration and changed an additional attached volume (not the root volume which is in the AMI) from Magnetic/standard to SSD/gp2. I changed the auto-scaling group to use the new launch configuration and executed a scaling policy to "Add 1 instances".

Then EC2 went into a loop starting , stopping, and terminating instances. EAch instance shows a state transition reason of "Client.InstanceInitiatedShutdown: Instance initiated shutdown". The system log of each instance is empty.

I created a copy of the copied launch configuration and changed the additional volume back to Magnetic/standard and updated the auto-scaling group, but it was still looping. I executed a scaling policy to "Remove 1 instance" to get it to stop.

What did I do wrong? The AMI is fine as I have another auto-scaling group using the same AMI. The auto-scaling group was working fine until I made this change. (I changed the instance type earlier in the day with no problems.)

Jayen
  • 1,857
  • 4
  • 16
  • 28

1 Answers1

1

I once faced the same situation, issue was that we ran out of gp2 volumes in our account. Check if you have sufficient gp2 disks in your account.

Gaurav Pundir
  • 1,496
  • 12
  • 14
  • I didn't know there was such a limit. That's good to know, but that's probably not our issue as we have ~10 volumes. I ended up creating an identical launch configuration with the CLI and it worked fine, so I would assume the web interface changed some hidden field. – Jayen Jan 26 '16 at 00:13