Disk Full/Over-Allocated after power failure

TheWaterbug

Known around here
Oct 20, 2017
1,074
2,046
Palos Verdes
BI 5.9.48/x64. OS and db on an internal SSD. New stored to 8500 GB of a 4-disk, 9.08 TB external Storage Space/JBOD over USB with daily compact/reindex:

1723025929337.png
1723025856812.png

Normally this works fine, and my storage is nearly "full" (as reported by BI) with a buffer of ~400 GB of real empty disk space (as reported by Windows Explorer), and new clips replacing old continuously.

Until I have a power failure, that is. If I have a power failure, my BI box restarts automatically, but my storage box does not. I have to physically visit the machine and press the power button on the storage box.

After that, I will often have a symptom where BI fails to store new clips due to drives being over allocated (as reported by BI) and the Storage Space being full (<<<1 GB free) as reported by Windows Explorer. I forgot to take screen captures of these before fixing the problem (see below), but the BI logs during this time show events like:

Code:
2     8/6/2024 8:48:21.943 AM    BackDoor                Clip: Disk full
2     8/6/2024 8:48:21.970 AM    TrailDown               Clip: Disk full
2     8/6/2024 8:48:22.024 AM    EVSouth                 Clip: Disk full
2     8/6/2024 8:48:22.232 AM    Park                    Clip: Disk full
2     8/6/2024 8:48:22.364 AM    MailboxIn               Clip: Disk full
2     8/6/2024 8:48:22.424 AM    Banana                  Clip: Disk full
2     8/6/2024 8:48:22.525 AM    Bee                     Clip: Disk full
0     8/6/2024 8:48:23.836 AM    New                     Delete: over quota 821.1/8500GB, 3.75MB free
0     8/6/2024 8:48:23.832 AM    Clips                   DeleteFile OK: V:\BlueIris\New\Garage.20240805_101616.bvr
0     8/6/2024 8:48:23.833 AM    New                     Delete: 1 items 2.25GB
0     8/6/2024 8:48:23.833 AM    Stored                  Delete: nothing to do [0.00/0.00GB, 2.25GB free]
0     8/6/2024 8:48:23.844 AM    Alerts                  Delete: nothing to do [0.00/0.00GB, 2.25GB free]

followed shortly again by:

Code:
2     8/6/2024 8:51:04.662 AM    UGate                   Clip: Disk full
2     8/6/2024 8:51:04.663 AM    Banana                  Clip: Disk full
2     8/6/2024 8:51:04.677 AM    LowerStair              Clip: Disk full
2     8/6/2024 8:51:04.674 AM    CombLapse               Clip: Disk full
2     8/6/2024 8:51:04.681 AM    SwarmIn                 Clip: Disk full
2     8/6/2024 8:51:04.682 AM    SwarmOut                Clip: Disk full

and new video doesn't get saved:

1723027144895.png

None of the database maintenance options will fix this; the only fix (that I've found) has been to manually delete the oldest ~1 TB of files in V:\BlueIris\New\ and the repair the database.

Anyone know why this happens and how this can be prevented/fixed within BI?

One hypothesis is that, when BI is running while the storage box is off, it knows it needs to delete old files from ~8 days ago, so it tries and fails, but removes those clips from the db. Then, since they've been removed from the db, BI never tries to delete them again once the storage box comes back online, and the drives fill up, but BI doesn't know why, so it can't fix it. Plausible?
 

Attachments

  • 1723026801488.png
    1723026801488.png
    803.1 KB · Views: 1
BI 5.9.48/x64. OS and db on an internal SSD. New stored to 8500 GB of a 4-disk, 9.08 TB external Storage Space/JBOD over USB with daily compact/reindex:

Normally this works fine,........

Until I have a power failure, that is. If I have a power failure, my BI box restarts automatically, but my storage box does not.
Sounds to me like a UPS would solve (mostly) your issue...I have my BI server and POE switch on one, I recommend everyone have one (or more) for BI or NVR. :cool:
 
Sounds to me like a UPS would solve (mostly) your issue...I have my BI server and POE switch on one, I recommend everyone have one (or more) for BI or NVR. :cool:
Yeah, long overdue. I just purchased this basic $80 UPS for my BI server and disk box. My Kill-A-Watt shows that these two consume about 100 W, steady state, and APC estimates this will give me 23 minutes of run-time. This will handle most of our local outages, which tend to be in the << 5 minute range, and will allow for graceful shutdown when necessary.

Thanks for the suggestion!
 
As an Amazon Associate IPCamTalk earns from qualifying purchases.
  • Like
Reactions: TonyR