failed to login(login failed: start failed: 400 - Unknown error)

[kidtangerine]

I recently switched from using Arlo to Pyaarlo for doing a simple snapshot of Arlo cameras for timelapse videos. Everything worked as expected for 24 hours, and now multi hosts local and cloud are throwing the error:

failed to login(login failed: start failed: 400 - Unknown error)

2025-03-22 00:05:36,904 - urllib3.connectionpool - DEBUG - https://ocapi-app.arlo.com:443 "POST /api/startAuth HTTP/1.1" 200 None
2025-03-22 00:05:36,906 - pyaarlo - DEBUG - backend: request-body=
{'meta': {'code': 400, 'error': 1110, 'message': 'Unknown error'}}
2025-03-22 00:05:36,906 - pyaarlo - DEBUG - backend: request-end=200
2025-03-22 00:05:36,907 - pyaarlo - WARNING - error in new response={'meta': {'code': 400, 'error': 1110, 'message': 'Unknown error'}}
2025-03-22 00:05:36,907 - pyaarlo - ERROR - login failed: start failed: 400 - Unknown error
2025-03-22 00:05:36,907 - pyaarlo - DEBUG - backend: login failed, trying another ecdh_curve
2025-03-22 00:05:36,907 - pyaarlo - DEBUG - backend: failed to log in
2025-03-22 00:05:36,908 - pyaarlo - DEBUG - media-downloader: not starting downloader
failed to login(login failed: start failed: 400 - Unknown error)

Any Ideas?

[Jeff-Rand]

That is pretty much what I've been dealing with since Arlo changed the backend. I usually get locked out for 24 hours. After that period, I can resume. Until this is fixed, I've stopped trying to continuously monitor the cameras for motion detection and just log in once every half hour. This has helped but I still see the login error occasionally. Unfortunately, I'm not real familiar with the login/authentication code so I've been unable to determine what is failing.

[twrecked]

I've seen the same thing and I'm looking at how to mitigate it. I'm working on a branch that refactors the authentication.

But can either of you give me an idea of what your work flow is, ie, do you login a lot or leave it running etc?

[kidtangerine]

I have a cron job every 15 mins to get a snapshot of a camera. So login a lot.

Maybe this is the wrong approach too.

[twrecked]

For now you might need to back that off to every hour.

The quick fix I made to work around the Arlo change is probably too login happy. I'm looking at it.

[Jeff-Rand]

I login and setup a callback to monitor camera attribute changes using camera.add_attr_callback('*', attribute_changed). When the callback receives an attribute of presignedLastImageData, I know that a recording has begun. I wait for the recording to end and then I call
station.build_ratls() station.build_media_library() to download the video from the base station. I then idle until the next video is generated. Up until recently, I would stay logged in for a very long time (weeks or months).

Right now, I'm no longer monitoring the camera attribute changes. I just log in every half hour and retrieve any new videos. This works, sort of...

[twrecked]

@Jeff-Rand you can probably get yours to work, that's how it's used in Home Assistant and mine has been up for a couple of weeks now.

Can you post your settings? Obviously without the sensitive stuff.

[kidtangerine]

Is it possible to stay logged in, and just pull the snapshot every 15 minutes?

[twrecked]

@kidtangerine

Yes. That's how Home Assistant works. You can login and then get a snapshot then wait 15 minutes and get another one. The code should stay logged in in the background.

[kidtangerine]

@twrecked

Cool, so do a while loop with 900 seconds sleep command?

[Jeff-Rand]

I'm running version 0.8.0.18 that I downloaded on Feb. 28th. I can try the latest version and see how it goes.

This is the command I was using to log in:
arlo = pyaarlo.PyArlo( username=USERNAME,password=PASSWORD, tfa_type='SMS',tfa_source='console',user_agent='linux', verbose_debug=False,backend='sse',reconnect_every=110, library_days=1,save_media_to='F:/arlo/METADATA/${Y}${m}${d}/${SN}/${SN}_${Y}${m}${d}_${H}${M}${S}')

[twrecked]

@Jeff-Rand Try adding save_session=True to get it to reuse the authentication credentials.

@kidtangerine Pretty much. (Not knowing your code) But start like you do now and put the camera snapshot code into a loop with a sleep.

[kidtangerine]

Cool thanks

[kidtangerine]

When keeping the same login, by the script continuously running, how often should I expect to have to login in again, or redo MFA?

[kidtangerine]

Running a while loop indefinitely, and taking a snapshot every 10 minutes; after about 2 hours camera.get_snapshot starts returning the same cached image.

When I kill the script and immediately try to login again I get a cloudflare 403 error. I have to delete the aarlo folder with the cookie and other info in it and I get prompted for MFA again and go another 2 hours.

I've added my code.

gitHub-getSnapshot.py.txt