I am running Persistence tests. Here are my concerns.
- I see that Agent shows as a ‘delayed start’ after the powershell install script is run and it isn't started, and never starts up on its own. I think that is the essence of the bug. If we do intend the delayed start, whats the benefit / reason for this?
And 2) whats the mechanism for when it is intended / expected to be started?
- it doesn't start even if Agent is started (via starting Agent on command line), not sure if that was expected to start it tho... whats the spec here?
Tested on 7.9 BC 2 with Agent/Endpoint against cloud stack and win 7 x64 vSphere host
steps:
- deploy stack 7.9 to cloud
- transfer agent/Endpoint to host (and manually put Endpoint into Agent download if needed)
- follow steps in 7.X guide for starting Agent https://www.elastic.co/guide/en/ingest-management/7.8/run-elastic-agent.html
which means to me, to:
-- unzip Agent
-- call powershell script
-- call enrollment command as output by Fleet
-- start agent via: ./elastic-agent run
If this isn't right and/or needs to be updated in docs or we have a bug, please lets follow thru asap. I know we were just making changes to the .rpm / .deb installs, not sure if those need further changes to support persistence (or the guide / steps need more details or not)
screenshot

I am running Persistence tests. Here are my concerns.
And 2) whats the mechanism for when it is intended / expected to be started?
Tested on 7.9 BC 2 with Agent/Endpoint against cloud stack and win 7 x64 vSphere host
steps:
which means to me, to:
-- unzip Agent
-- call powershell script
-- call enrollment command as output by Fleet
-- start agent via: ./elastic-agent run
If this isn't right and/or needs to be updated in docs or we have a bug, please lets follow thru asap. I know we were just making changes to the .rpm / .deb installs, not sure if those need further changes to support persistence (or the guide / steps need more details or not)
screenshot
