Gitpod: Ports close after a time

[joshleblanc]

Describe the bug
Working on a project with an exposed port. At some time, I start seeing a screen telling me the port is not exposed, and to check my .gitpod.yml file.

To Reproduce
Unsure. Happens seemingly randomly. Restarting the container fixes it, however.

Expected behavior
Once a port is exposed, it should stay exposed. I suppose.

[svenefftinge]

Interesting, haven't seen that or heard from anyone.
Can you share a project with which it happens?

[jankeromnes]

Actually, I think the wording of that error screen is wrong. The problem isn't that the port is not exposed, but that nothing is currently running on the exposed port.

The solution is not to check your .gitpod.yml file (as the error screen currently says), but to make sure the server is actually running. We should fix this error screen (and also, the error screen should not redirect away, and it should auto-refresh every 5s to check if the requested app or server is back online).

[meysholdt]

I very much like Jan's proposal.

As a workaround, it's already possible to open the "Ports" view in Gitpod and see of some program listens on that port.

[joshleblanc]

The server is actually running. Restarting the server doesn't fix the problem. I should have put that in my original post.

I'll take some screenshots next time it happens.

[joshleblanc]

This about sums it up

[svenefftinge]

Can we reproduce it with this project?
https://github.com/HorizonShadow/hyperstack-test

[joshleblanc]

Go for it. It’s inconsistent. In the screenshot, the web process is clearly running, and the URL isn’t working. I think some of the times I’ve experienced it, the web process had just crashed.

Could well just be hyperstack and not gitpod

On Apr 6, 2019, at 5:52 PM, Sven Efftinge [e-mail] wrote:

Can we reproduce it with this project?
https://github.com/HorizonShadow/hyperstack-test

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub, or mute the thread.

[joshleblanc]

‘foreman start’ to run

On Apr 6, 2019, at 5:52 PM, Sven Efftinge [e-mail] wrote:

Can we reproduce it with this project?
https://github.com/HorizonShadow/hyperstack-test

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub, or mute the thread.

[jankeromnes]

Could it be that "web process just crashed" followed by "reload preview" showed you the error page once, and that since the error page is redirected away, even when the server resumes and you reload the error page, it stays the same?

[geropl]

@HorizonShadow Could it be that it is a problem with foreman? Do not know anything about it, but had similar problems here: A process behind it would not receive packages at all.

[joshleblanc]

@geropl It could be, however with my monkey patches, the hot-reloader + web server both work behind foreman. (Well, most of the time). I have noticed one or the other will hang sometimes, as described in that issue.

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.