search

Hivemapper / hdc_firmware

The firmware that runs the OpenDashcam
https://github.com/hivemapper/odc
Apache License 2.0
0 stars 3 forks source link

HDC stops recording randomly #208

Closed jayalberts closed 4 months ago

jayalberts commented 5 months ago

Camera will be going fine for hours, then stop recording. Three lights on. Camera view in app still updates as normal. Only way to save cam and continue hivemapping is to reboot. After reboot, no new data is sent to phone related to the sections driven while recording had stopped.

Reported on Discord by community testers

jayalberts commented 5 months ago

Mazar investigating logs, might not be related to changes he made.

@jayalberts double check if USB plugged in and exact timestamps.

jayalberts commented 5 months ago

All times PST

Eddie stopped recording 1/19 approx 11:20-11:40 and 1/18 approx 10:40-11:00. (No USB)

Koen stopped recording on 1/20 sometime before 2:30. (No USB)

Gator stopped recording on 1/20 sometime probably around 9:00-10:00. (no USB)

Both Gator and Koen reported app crashing when trying to send DB. Seems to point to some issue with camera DB perhaps?

mazar-farran commented 5 months ago

@jayalberts Did Koen and Gator use a USB?

mazar-farran commented 5 months ago

Also, to be 100% sure, all three testers reported the camera view was still showing images?

Masaya-RT commented 5 months ago

Today I drove on 3.9.1 for about an hour, with USB attached.

Caveat - Didn't clock several hours like the community testers

jayalberts commented 5 months ago

@mazar-farran

Koen: confirmed 3 lights on dashcam and image preview on app Gator: confirmed 3 lights on dashcam and image preview on app Eddie: confirmed 3 lights on dashcam but he doesn't use dashcam tab (only map) so can't determine if preview was showing

mazar-farran commented 5 months ago

@punov image preview on app indicates that camera bridge is unlikely to be the problem. I found some gaps in the DashcamPackedFrameKm updates on Mode (attached in slack). But those times also do not seem to all line up with the reported recording times. Could you take a look at this? Thank you

great-periwinkle-ice gap between 12:32 and 14:04 UTC hopeful-brown-bluebird gap between 19:20 and 19:40 UTC (as reported) (logs show "TRAVELLED TOO FAR" and "Not enough frames to add yet" messages). sharp-champagne-pig gap between 20:11 and 22:42 UTC, another smaller one between 17:02 and 17:30 UTC

jayalberts commented 5 months ago

Another user (Dragon666 on Discord) to investigate with same issue.

This happened on 1/26 around 1pm PST. All 3 lights, no USB.

idealistic-grass-crow

jayalberts commented 5 months ago

Mazar investigated further, no smoking gun.

jayalberts commented 5 months ago

No common cause, all slightly different. At least one user with reliability of GNSS logger.

jayalberts commented 4 months ago

This might have been related to db issues which were resolved in 4.2. Closing out