[Support]: Sync recordings option not working

[rstanislav]

Checklist

• I have updated to the latest available Frigate version.
• I have cleared the cache of my browser.
• I have tried a different browser to see if it is related to my browser.
• I have tried reproducing the issue in incognito mode to rule out problems with any third party extensions or plugins I have installed.

Describe the problem you are having

When option sync_recordings is enabled after start abort message in log appears, and no recordings deleted. (all of them should be deleted)
025-02-28 22:37:54.485843364 [2025-02-28 22:37:54] frigate.record.util INFO : Deleting 261000 recording DB entries with missing files

2025-02-28 22:37:54.667605538 [2025-02-28 22:37:54] frigate.record.util WARNING : Deleting 0.996605% of recordings DB entries, could be due to configuration error. Aborting...

Steps to reproduce

1. enable sync_recordings in config
2. restart

Version

0.15.0-cea210d

In which browser(s) are you experiencing the issue with?

No response

Frigate config file

record:
  sync_recordings: True
  enabled: true
  retain:
    days: 15
    mode: motion
  alerts:
    retain:
      days: 60
      mode: motion
    pre_capture: 5
    post_capture: 10
  detections:
    retain:
      days: 90
      mode: motion
    pre_capture: 5
    post_capture: 10
snapshots:
  enabled: true
  bounding_box: true

docker-compose file or Docker CLI command

..

Relevant Frigate log output

025-02-28 22:37:54.485843364  [2025-02-28 22:37:54] frigate.record.util            INFO    : Deleting 261000 recording DB entries with missing files

2025-02-28 22:37:54.667605538  [2025-02-28 22:37:54] frigate.record.util            WARNING : Deleting 0.996605% of recordings DB entries, could be due to configuration error. Aborting...

Relevant go2rtc log output

Empty

Operating system

Debian

Install method

Docker Compose

Network connection

Wired

Camera make and model

Any

Screenshots of the Frigate UI's System metrics pages

..

Any other information that may be helpful

No response

[NickM-27]

this is not a bug and is working as expected. Sync_recordings is meant to fix variations in files, not completely delete the database as this usually indicates an error like a missing volume mount

[rstanislav]

There is no mention in doc's about this, its expected to delete all files, because apart from deleting whole DB there is currently no way to manually delete all files(for example in situations where user upgrades HDD to get more storage without data migration). I dont understand tbh how situation where 'some' files are missing can happen, because if hdd failed for example they all will be corrupted to some extent. If its for situations where user 'wants' to delete 'some' files with no traces="tampering"(maybe its not the owner of the system?) - its even more strange, because any surveillance system is supposed to not allow such actions, because surveillance is about security. Frigate is great project, but this bug is strange(if its a 'feature', i think doc's should mention threshold in % when it will work).

[NickM-27]

because apart from deleting whole DB there is currently no way to manually delete all files(for example in situations where user upgrades HDD to get more storage without data migration).

in that case you should just delete the entire DB

I dont understand tbh how situation where 'some' files are missing can happen,

it is very simple, maybe you are migrating your docker compose and forget to map the recordings directory, or maybe one of multiple HDDs in a pool does not connect / fails, but data recovery is possible. Or perhaps the database becomes corrupted but a backup exists (since the sync is bidirectional)

there are many cases where this can happen and this is a safety feature.

I will put up a PR to update the docs to include information about this scenario

[rstanislav]

Thank you for your response! However, I still believe this could potentially compromise security. At present, Frigate doesn’t support creating a “pool” of hard drives. The only file system I know of that offers something similar is Unraid (when the directory is mounted). In a classic RAID setup—say RAID 0 with two drives—if one drive fails, all data is lost. If the system owner discovers that all recordings have disappeared, it can only be due to a misconfiguration or a breach, and in both cases, an investigation is necessary to determine what happened. If certain events vanish silently, it opens the possibility for an intruder to tamper with the data without the owner ever noticing.

[NickM-27]

At present, Frigate doesn’t support creating a “pool” of hard drives

Right but it uses docker so that is irrelevant, any number of solutions for this can and are used.

The only file system I know of that offers something similar is Unraid

Unraid just uses fuse.fs which is an open Linux file system. There is also mergerfs, among others. This is a fairly common setup, not even considering that a lot of Frigate users run unraid.

Also, the drive doesn't even need to fail. By default docker won't stop even if a drive goes offline, so even something simple like a power cable failure or accidentally unplugging a drive would have the potential to have recordings deleted that didn't need to be.