Program crashes after running for a few days about sneakpeek

[xlanor]

Are you opening a file and not closing it properly anywhere?

[fterh]

Nope. I suspect it ran out of memory or something, because this's happened before, and it's always after a few days.

@yleong suggested running it as a system service I think, so this might be low priority if we're able to rewrite how the program is run.

[xlanor]

If you are planning to do a refactor, I strongly suggest that the files be regrouped, minimally, into an src/ drive and tests/ drive respectively for source code and test files.

[yleong]

The Reddit API was returning a HTTP response status of 503 Service Unavailable when praw is trying to call response = self._authenticator._post(url, **data) to get new submissions.
The relevant line is prawcore.exceptions.ResponseException: received 503 HTTP response. (source here) The rest of it looks to be this exception being thrown up the stack until it hits some sort of stdin handler.
I think we can catch this response here and retry the request after waiting for a bit, since 503 sometimes indicates an overloaded server. This can be done in addition to running it under pm2 or systemd to ensure higher uptime.

Regarding the code refactor, I am not familiar with Python project structure, but what @xlanor suggested makes sense.

[yleong]

Are there more crash logs available?

[fterh]

Yup!

nohup.txt

[yleong]

Ricemedia handler also threw exception. I have opened #19 to track that separately.

[fterh]

@yleong I checked the PRs and I saw you merged one which fixes this. Can I close this issue and prepare to create a new release?

[yleong]

Sure. If the program still crashes maybe we can reopen.

[fterh]

Okay, closing this for now. @yleong @xlanor I'm pushing v0.5.0-beta into production, I've run manual tests and it looks good.