AMP update hangs unless all instances are manually stopped, and service must be manually restarted each time

[SlimeProgenitor]

Bug Report

System Information

• Operating System Windows Server 2016 Standard (windows updates current as of time of reporting)
• AMP Release "Halimede" v2.4.1.2, built 21/12/2022 17:57 (has been observed on multiple mainline release over the last year, but this is the one I just updated to today from v2.4.0.6)
• Mainline

I confirm:

• [x] that I have searched for an existing bug report for this issue.
• [x] that I am using the latest available version of AMP.
• [x] that my operating system is up-to-date.

Symptoms

When attempting to update the main AMP instance, the process always hangs unless I manually stop all game server instances first. When I start the update from the main instance, the AMP Service Handler process will close, but AMP Core Runtimes for the game instances will keep running, and then the AMP Service will fail to restart (presumably due to the AMP Core Runtime instances still being active) At this point, I resolve the issue by closing all AMP Core Runtime processes and manually starting the AMP service in the services list, which will launch the AMP Service Handler again. if I attempt to start the AMP service from the services list before closing all the AMP Core Runtime processes, it will briefly launch the service handler before stopping again. After logging into the manager again, it usually still shows the server as out of date and prompts me to install the update. Performing the update (and resolution process) a second time usually successfully applies the update.

Reproduction

• install an old version of AMP or wait for a new update to be pushed
• open the manager in your browser, log in, have one or more instances running, and click the button that says "update"
• click the button that says "update now" in the warning that pops up telling you the instance will be offline for the duration of the update
• observe the AMP Service Handler closing and AMP Core Runtime processes idling in task manager

[southnode]

This is a known problem with the Windows version of AMP wherein if ADS goes down for any reason (such as an update) it requires all AMP instances to be turned off before ADS will cleanly start again.

The reason being that the particular port AMP uses doesn't seem to be let go of when the service stops. This is not intended behaviour and other Windows-based control panels do not have this problem. Further development on the Windows side will be required to remove this problem and to bring the Windows installations to parity with the Linux ones.

A note that this is still a problem in 2.4.3.0 - the current mainline release as of time of writing.

[stricq]

I had this same problem with the latest 2.4.3.2 update on Linux. The instances refused to update until they were completely shutdown manually. Even then it took several attempts at manually shutting the instances down before they finally shutdown.

[IceOfWraith]

This should be solved in the latest release. Please respond if it's still an issue and it can be looked into further.