Description of problem:

Right now we are using Restart=on-failure in docker.service file and that leads to multiple restarts of docker if docker fails to start due to say some configuration error. For example, say lvm thin pool specified by docker-storage-setup is not there.

This also has side affect of filling logs of messages like.

 systemd[1]: docker.service failed.
 systemd[1]: start request repeated too quickly for docker.service
 systemd[1]: Failed to start Docker Application Container Engine.

And actual error message might get lost or be deep buried in the stack.

So use Restart=on-abnormal instead. That should make sure docker restarts if docker fails abnormally (like core dump). 

Version-Release number of  selected component (if applicable):


How reproducible:


Steps to Reproduce:
1.
2.
3.

Actual results:


Expected results:


Additional info: