Startup errors crashing in log statement about chanjo2 HOT 17 CLOSED

You can store cases and samples in MySQL if you want, but it's optional really. Yes, MongoDB is not used at all here so yes, it should be removed

OK I understand! OK I'll remove the dependency.

from chanjo2.

I think this will fix it. Closing for now. Reopening if it comes back.

from chanjo2.

It looks like this update fixed the error - the logs look clean now.

We still have the issue with Chanjo2 not running after every weekend though 🤔 But seems more and more likely that this is related to our setup rather than Chanjo2.

Anyway, I think this issue is resolved. I'll continue digging. If it still seems related to Chanjo2 internals, I'll be back with a separate issue.

from chanjo2.

Hi @Jakob37, we never saw this error but I guess it's because we don't use uvicorn as a runner, like you guys do. We don't have to have uvicorn handling the logging. I'll open a PR with a change

from chanjo2.

Hi @Jakob37, we never saw this error but I guess it's because we don't use uvicorn as a runner, like you guys do. We don't have to have uvicorn handling the logging. I'll open a PR with a change

OK thanks!

from chanjo2.

What's the code you use to start the server? Are you logging to file?

from chanjo2.

Also, are you passing a log config file when you start the app?

from chanjo2.

What's the code you use to start the server? Are you logging to file?

Here is the command we use to start Chanjo2:

command: ["uvicorn", "chanjo2.main:app", "--host=0.0.0.0", "--port=8000", "--workers=2", "--log-level=info", "--root-path=/chanjo2"]

Also, are you passing a log config file when you start the app?

Hmm, looks like not, not intentionally at least.

from chanjo2.

Hmm, looks like not, not intentionally at least.

We don't have a log config file either. It's weird, because if the app works fine with the command you mentioned I don't understand why you get that error over weekends. I'm not sure how to fix the logging since I don't understand what's really going on.
Looks like the app is restarted in the wrong way by some process because you get a call to lifespan. Chanjo2 is working with a MySQL database so it shouldn't be affected by the MongoDB backups. Please try to investigate a bit more what can be trying to restart chanjo2 on your side.

from chanjo2.

Hmm, looks like not, not intentionally at least.

We don't have a log config file either. It's weird, because if the app works fine with the command you mentioned I don't understand why you get that error over weekends. I'm not sure how to fix the logging since I don't understand what's really going on. Looks like the app is restarted in the wrong way by some process because you get a call to lifespan. Chanjo2 is working with a MySQL database so it shouldn't be affected by the MongoDB backups. Please try to investigate a bit more what can be trying to restart chanjo2 on your side.

OK, I'll investigate a bit more!

from chanjo2.

OK, I'll investigate a bit more!

One thing that comes to my mind is that perhaps database and logging initiation should be present on a app startup function and not on a lifespan one (which gets executed also when the app stops or reboots)- I'll fix accordingly, and perhaps this will solve your issue

from chanjo2.

Chanjo2 is working with a MySQL database so it shouldn't be affected by the MongoDB backups.

Hmm, aha. I had misunderstood here. I thought the SQL database was used only for intervals, but I see now that it also stores the cases and samples.

I currently have the mongodb container as a dependency for Chanjo2. Maybe that is part of what is happening here. I'll remove it.

from chanjo2.

You can store cases and samples in MySQL if you want, but it's optional really. Yes, MongoDB is not used at all here so yes, it should be removed

from chanjo2.

Looks like it came back, without the mongodb dependency and with the updated version running the log statement from within the lifetime method.

I guess there are two parts here.

(1) Why is Chanjo2 restarting (must be something in our setup)
(2) Why does it crash on logging

I'll see if I can configure uvicorn somehow such that it doesn't crash on logging.

from chanjo2.

Question @northwestwitch , in the logger.py, specifically this part:

if LOG.handlers:
    LOG.handlers[0].setFormatter(console_formatter)
else:
    logging.basicConfig()

If the else-statement is hit for some reason, logging.basicConfig() is called. This does not seem to update the LOG instance, which later is used. Could this cause issues?

Anyway, I'll try to see if I can reproduce the error on my laptop. I think with a quicker cycle, I should be able to pinpoint the error.

from chanjo2.

Ok! I'll look into it

from chanjo2.

Anyway, I think this issue is resolved. I'll continue digging. If it still seems related to Chanjo2 internals, I'll be back with a separate issue.

Please do that! Thanks for reporting and sorry if we couldn't pinpoint the issue better

from chanjo2.