Images fail intermittently on aws

[bsb808]

A student anecdotally reported that they would loose connection with their cloudsim image periodically. Until today, I had not seen this behavior - and it will be hard to debug since it sounds like it is infrequent and we don't yet have any suspicions for causes or correlations.

Something similar may have happened to me. The image was running for a few hours and now I cannot connect via novnc.

Here is the output of the REST status call...

curl -X GET -H "Private-Token: ${TOKEN}" https://staging-cloudsim-nps.ignitionrobotics.org/1.0/simulations/${GROUPID}
{"created_at":"0001-01-01T00:00:00Z","owner":"","name":"npstest","groupid":"6038eff4-95e0-4f06-8d41-e6809f904da6","status":"Removing instance.","image":"tfoote/test_novnc:main","args":"","uri" REMOVED"}

I'll leave the machine running for a bit in case some one would like to debug and can look more deeply at the status of the machine.

[bsb808]

More evidence that might help with debugging.

• For a couple of hours, the reported status remained "Removing instance."
• I think issued a stop command using the REST API. The status goes to "Stopping" and then "Removing docker image (pod)."
• This seems to suggest that the image was in a state were it was still running, to some extent, but was inaccessible via novnc.