Closed luchfirearmland closed 1 month ago
I'm having a similar problem here, but I can't see any running processes running on port 5080 to kill
Check also if you did not use some similar programs which can use a docker etc.
Like I have if I run my local WP tool and try to start laravel docker it will not work. After closing a program I can run a laravel docker etc.
After a few runs, I've noticed it usually happens when something unexpected makes the deploy fail besides the expected internal errors. For eg. network failure or SIGINT.
The script doesn't gracefully exit by cleaning up the docker registry and every other changes it made
Strange... I have the cleanup set here:
Which calls:
Is there any improvements you see here?
I found the culprit. The tunnel_pid was not set until the end of the script. This could cause it to fail and never kill the tunnel.
I separated the tunnel and registry clean up and it should work a lot better.
I have this on this commit and it will be made available in the next release of Spin https://github.com/serversideup/spin/commit/7e20189d45c8ff13abdc14cd72ad4426c372f6aa
Current Behavior
So the issue started when I tried to run "spin deploy" so I can deploy my changes. As you can see in the image below, the deploy process was halted because it can't establish an SSH tunnel.
At first I thought my SSH is not correctly configured and so I tried running this command:
ssh -p 22 deploy@18.***.***.**
I thought it wouldn't work but it did. Which means that I can do an SSH connection but port forwarding fails for some reason.And so I tried to ask the error on Gemini. It told me that yes the SSH connection is indeed successful but it can't do the port forwarding because the port 5080 might be occupied by some process. I then followed the commands to check it out if indeed there's something running on that port, and it sure did!
I then proceeded to kill that process by its ID:
sudo kill 377159
.Afterwards, I tried running
spin deploy
once again and it now works!Expected Behavior
I expected
spin deploy
to run without fail. It failed due to port 5080 being occupied at the moment.Steps To Reproduce
I actually don't know how it happened. But it might be because I abruptly stopped the already running
spin deploy
so it didn't finish closing the connection.Environment
Anything else?
No response