Containerized execution hangs because of update checks

[jason-weirather]

Describe the bug
While executing comfy-cli in a docker container, execution hangs when the code calls comfy_cli.update's check_for_updates function.

To Reproduce

1. Install comfy-cli in a docker container
2. Disable tracking in comfy-cli since its intended to run in a container without phoning out
3. Use comfy launch to spin up a server in the background (--listen in the args should be set to 0.0.0.0 in docker)

Expected behavior
Upon running comfy launch the server should start up and run, indicating a background server is up and return to a prompt.

Additional context
I have succeeded to get comfy-cli to work in a container by modifying the code to not check for updates in commands that have nothing to do with updating.

My Dockerfile that does succeed with the modified comfy-cli

https://github.com/jason-weirather/comfyui-image-api/commits/main/Dockerfile

My branch of the fork where I've gotten it working by removing a few calls checking for updates from the comfy launch and comfy env commands.

https://github.com/jason-weirather/comfy-cli/tree/fixhangbug/comfy_cli

[robinjhuang]

Can you share the error message you see? Does it halt the comfy launch?

[jason-weirather]

Can you share the error message you see? Does it halt the comfy launch?

Theres no error message. It just hangs and halts the launch if it cannot reach out to talk to pypi.

[robinjhuang]

How about we make the request timeout 3 seconds Sent via Superhuman iOS ***@***.***>

…

On Fri, Sep 13 2024 at 9:39 PM, Jason L Weirather ***@***.***> wrote: Can you share the error message you see? Does it halt the comfy launch? Theres no error message. It just hangs and halts the launch if it cannot reach out to talk to pypi. — Reply to this email directly, view it on GitHub <#175 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABLJKASKKAZ2SENDMVRP3F3ZWLMJTAVCNFSM6AAAAABN2YIW3GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGNBYHA2TSNBTGY> . You are receiving this because you commented.Message ID: ***@***.***>

[jason-weirather]

How about we make the request timeout 3 seconds

I think it sounds great solution to the hang bug. Should I make a PR request that implements this?

In the future, I think it would be good to add a command line option for an offline mode that could save that time and skip bonking pypi. A few seconds if you are spinning up dockers as like a lambda function thing is still not great, but thats something I could raise in a feature request rather than bug fix.

[robinjhuang]

Please make a PR. Having an offline mode makes sense as well! Please make another issue for that and we will get to it. Sent via Superhuman iOS ***@***.***>

…

On Sat, Sep 14 2024 at 12:01 AM, Jason L Weirather ***@***.***> wrote: How about we make the request timeout 3 seconds I think it sounds great solution to the hang bug. Should I make a PR request that implements this? In the future, I think it would be good to add a command line option for an offline mode that could save that time and skip bonking pypi. A few seconds if you are spinning up dockers as like a lambda function thing is still not great, but thats something I could raise in a feature request rather than bug fix. — Reply to this email directly, view it on GitHub <#175 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABLJKAX2VEGMK75NCYQ5CKTZWL46NAVCNFSM6AAAAABN2YIW3GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGNBZGE3DMMJUGQ> . You are receiving this because you commented.Message ID: ***@***.***>