Frequent DB disk write access even after background jobs disabled

[social-sparrow]

I have observed frequent disk usage access to sqlite db files on my self hosted instance with background jobs disabled.

Log: MODIFY <docker data>/linkding/data/tasks.sqlite3-wal

I have analyzed the db access scenario by enabling the Debug logs, I observed that, when UI/API access in not used, then also noticed on writes on the disk, which are accessed periodically at ~30S interval.

For my local testing, tweaked the below interval values, this results into nearly ZERO disk write when UI/API was not accessed.

HUEY = {
    "huey_class": "huey.SqliteHuey",
    "consumer": {
        "scheduler_interval": 10,  #--> changed to 200s
        "periodic": True,
        "check_worker_health": True,
        "health_check_interval": 10, #--> changed to 100s
    },
}

I would like to understand the intervals and is it possible to reduce the disk usage when the app inactive ?
Background tasks are absolutely okay but disk writes at periodically in very short intervals could be avoided.

[sissbruecker]

When the LD_DISABLE_BACKGROUND_TASKS variable is set correctly it should prevent the background task queue from even starting:

linkding/bootstrap.sh

Lines 30 to 33 in 83092cc

	# Start background task processor using supervisord, unless explicitly disabled
	if [ "$LD_DISABLE_BACKGROUND_TASKS" != "True" ]; then
	supervisord -c supervisord.conf
	fi

From some testing I can observe that the database is not touched again after the initial startup. As another indicator, there should be no background_tasks.log file in the container. Are you sure you have applied the variable correctly?

[social-sparrow]

Okay, Looks like environment variables (ex: LD_DISABLE_BACKGROUND_TASKS) are case sensitive. In my environment, I have used "true", which leads to this observation. After your comment, changed to "True" and issue resolved.

It might be better if the environment variable checks were case-insensitive.

Followup request: I would like to have background tasks enabled but for longer duration for periodic interval (ex: hours, daily, weekly, monthly). can we have this option configurable ?

[sissbruecker]

Not sure if I want to make that configurable out of the box. Specifically the way how HTML snapshots are scheduled is a bit odd at the moment: it relies on some repeating task running every minute to check if there are any missing snapshots, and then does like 5 of them sequentially. That would require some further testing how that would be impacted.

For now a workaround could be to create a custom.py file and mount that into the container at /etc/linkding/bookmarks/settings/custom.py. In that file you can override any of the base settings, for example with:

HUEY = {
    "huey_class": "huey.SqliteHuey",
    "filename": os.path.join(BASE_DIR, "data", "tasks.sqlite3"),
    "immediate": False,
    "results": False,
    "store_none": False,
    "utc": True,
    "consumer": {
        "workers": 2,
        "worker_type": "thread",
        "initial_delay": 5,
        "backoff": 1.15,
        "max_delay": 10,
        "scheduler_interval": 200, # changed to 200s
        "periodic": True,
        "check_worker_health": True,
        "health_check_interval": 200, # changed to 200s
    },
}

[social-sparrow]

Looks like Huey is not happy with the large interval.

  File "/etc/linkding/.venv/lib/python3.13/site-packages/huey/consumer_options.py", line 135, in validate
    raise ValueError('The scheduler must run at least once per '
                     'minute, and at most once per second (1-60).')
ValueError: The scheduler must run at least once per minute, and at most once per second (1-60).

Maximum possible interval is 60 Seconds

2025-12-17 19:58:04,213 INFO Huey consumer started with 2 thread, PID 30 at 2025-12-17 14:28:04.213964
2025-12-17 19:58:04,214 INFO Scheduler runs every 60 second(s).
2025-12-17 19:58:04,214 INFO Periodic tasks are enabled.

Every execution leads to disk write access.

2025-12-17 19:58:04,215 INFO Enqueueing periodic task bookmarks.services.tasks._schedule_html_snapshots_task: 4d9a4f2b-ecc7-426b-8f1a-4c0c94792f72.
2025-12-17 19:58:04,217 INFO Executing bookmarks.services.tasks._schedule_html_snapshots_task: 7eaaaea8-cdaa-43b4-bf9e-d6b862a185cc
2025-12-17 **19:58:04**,223 INFO bookmarks.services.tasks._schedule_html_snapshots_task: 7eaaaea8-cdaa-43b4-bf9e-d6b862a185cc executed in 0.006s
...
2025-12-17 **19:59:04**,214 INFO Enqueueing periodic task bookmarks.services.tasks._schedule_html_snapshots_task: 3075f4b6-cbf0-4ed0-843b-8fdcaf507228.
...
2025-12-17 **20:00:04**,214 INFO Enqueueing periodic task bookmarks.services.tasks._schedule_html_snapshots_task: 6bcc6ad9-10c1-48d1-ac5a-e6633a0b4c70.
..