Skip to content

Run VariantGrid

Dave Lawrence edited this page Aug 7, 2026 · 3 revisions

If you are running a server or VM, we recommend the services set up in Install (Server) so components will automatically start with the system.

Starting as a developer

Activate the venv first (see Install Python venv), then:

cd ${VARIANTGRID_DIR}
source .venv/bin/activate
python3 manage.py runserver

Then connect to http://localhost:8000

Anything asynchronous — VCF import, annotation, liftover, ClinVar export — runs on Celery, so you also need a worker. RabbitMQ (the broker) and Redis (result backend / cache) are installed by scripts/install/ubuntu_install_dependencies.sh and run as system services.

For development, one worker covering every queue is usually enough:

celery --app variantgrid worker \
    --queues=analysis_workers,annotation_workers,db_workers,web_workers,scheduling_single_worker,variant_id_single_worker \
    --loglevel=INFO

For scheduled tasks (hourly syncs etc) also run beat:

celery --app variantgrid beat --loglevel=INFO

The queues and what routes to them are defined in variantgrid/settings/components/celery_settings.py. On a server each queue gets its own worker process — see Install (Server).

Checking it's working

  • Settings -> Server Status (/variantopedia/server_status) - Celery workers should be green
  • Annotation (top menu) - look for anything red (annotation that needs upgrading)

If something isn't running, see Troubleshooting.

Clone this wiki locally