After a few weeks of working fine, now getting Fatal error: Uncaught Error: Call to a member function fetchArray()...

[chezlong]

Installed BirdNET-Pi a few weeks ago on a Raspberry Pi 3 Model B Plus Rev 1.3 with a USB mic (nothing special), and was very pleased with the results. However, I recently see this on the Overview page- Fatal error: Uncaught Error: Call to a member function fetchArray() on false in .../BirdNET-Pi/scripts/overview.php:304 Stack trace: #0 .../BirdNET-Pi/homepage/views.php(154): include() #1 {main} thrown in .../BirdNET-Pi/scripts/overview.php on line 304

The only change I made recently was to tick the box for the get internet date/time. I see code in the area of the error is related to date, so wondering if this could be the reason. I reverted to manual input, but still getting same error. My PHP and SQL knowledge is non-existent!

I found a similar error query in the original mcguire github discussion, tried the suggestions there, but didnt fix it.
Anyone found anything similar?

[EDIT] added a bit of debug code to the overview.php before line 304 to check the $result = $statement->execute(); line, and $result is indeed false.

[philphauler]

That stack trace means the prepared statement never ran successfully: $statement->execute() returned false, so $result is boolean false and calling fetchArray() on it fatals.

In BirdNET-Pi this almost always points at SQLite, not the “internet date/time” checkbox itself (that flag only changes how timestamps are written once the DB is healthy).

What to check (in order)

1. Is the DB file still there and readable?

   ls -l ~/BirdNET-Pi/scripts/birds.db
   # or wherever your install put it — common paths:
   # /home/pi/BirdNET-Pi/scripts/birds.db
   # /var/www/html or the path printed in overview.php near the SQLite open

   If the file is missing, zero bytes, or owned by root while PHP runs as www-data/nginx, execute() fails every time.

2. Open the DB and run a probe query

   sqlite3 /path/to/birds.db "PRAGMA integrity_check; SELECT count(*) FROM detections LIMIT 1;"

   • integrity_check → ok and a number = DB is fine, likely path/permissions in PHP.
   • error / “database disk image is malformed” = corruption (power loss on Pi SD cards is common after weeks).

3. If corrupted: restore or rebuild

   • Prefer a backup under BirdNET-Pi / scripts if you have one.
   • Otherwise stop the services, move the bad DB aside, and re-run the project’s install/update script so a fresh schema is created (you’ll lose detection history unless a backup exists).

4. Disk full / SD dying

   df -h
   dmesg | tail -50

   A full root partition makes SQLite fail writes and later reads in weird ways.

5. Make the error readable (temporary debug, remove after)
   Right before execute(), log:

   if ($result === false) {
     error_log(print_r($db->lastErrorMsg(), true));
   }

   lastErrorMsg() usually names the real cause (unable to open database file, locked, no such table, etc.).

About the date/time toggle

Reverting it was the right first step, but once execute() is already failing, the overview page will keep fataling until the DB open/query path works again — so the checkbox is a red herring unless it somehow pointed SQLite at a bad path (uncommon).

If you paste the output of PRAGMA integrity_check and ls -l on birds.db, we can say restore-vs-permissions in one shot.

[chezlong]

Thanks so much for the quick reply-

1. Database exists, and non zero.
2. Integrity check resulted in Error: stepping, database disk image is malformed (11) - I will rebuild when I get a chance.

Interestingly, I also checked 4) and the dmesg tail was fill of [ 1554.398554] __vm_enough_memory: pid: 1784, comm: streamlit, bytes: 17180053504 not enough memory for the allocation, so maybe a full root partition too?

[colinpeddle]

Fairly certain this was the error I saw when I was out of storage.