Running tests locally¶

Pytest and Labgrid on the lab host. Deploy and services: Routine operations.

First time: validate the lab¶

Concepts: Place = reservable DUT (serial, power, TFTP, SSH). Exporter = publishes places. Coordinator = reservations. labgrid-client / pytest acquire and run.

Steps: (1) Deploy configs via Ansible (see Routine operations - Ansible). (2) sudo systemctl start labgrid-coordinator labgrid-exporter if not running. (3) labgrid-client places. (4) Firmware on TFTP: TFTP / dnsmasq. (5) Reserve before pytest: labgrid-client lock (with LG_PLACE), run tests, labgrid-client unlock.

labgrid-dev and TFTP symlinks¶

TFTP dirs belong to labgrid-dev. For Labgrid to create symlinks when running tests, connect to the host as labgrid-dev. Paths and ownership: host-config, TFTP / dnsmasq.

Single-node test (one DUT)¶

# OpenWrt vanilla
LG_PLACE=labgrid-fcefyn-belkin_rt3200_1 \
LG_IMAGE=/srv/tftp/firmwares/belkin_rt3200/openwrt/openwrt-23.05.5-vanilla-mediatek-mt7622-linksys_e8450-ubi-initramfs-recovery.itb \
uv run pytest tests/test_base.py tests/test_lan.py -v

# LibreMesh (fixture moves DUT VLAN to 200 automatically)
LG_PLACE=labgrid-fcefyn-belkin_rt3200_1 \
LG_IMAGE=/srv/tftp/firmwares/belkin_rt3200/libremesh/lime-24.10.5-mediatek-mt7622-linksys_e8450-initramfs-kernel.bin \
uv run pytest tests/test_libremesh.py tests/test_lan.py -v

Multi-node LibreMesh test¶

The libremesh-tests fixture moves DUT VLANs to 200 automatically via switch-vlan invoked from conftest_vlan.py (SSH to the lab host if LG_PROXY is set). Design background (Labgrid strategies, TFTP_SERVER_IP, parallel boot subprocesses): Labgrid mesh strategy and orchestration.

LG_MESH_PLACES="labgrid-fcefyn-openwrt_one,labgrid-fcefyn-bananapi_bpi-r4,labgrid-fcefyn-librerouter_1,labgrid-fcefyn-belkin_rt3200_2" \
LG_IMAGE_MAP="labgrid-fcefyn-openwrt_one=/srv/tftp/firmwares/openwrt_one/libremesh/lime-24.10.5-mediatek-filogic-openwrt_one-initramfs.itb,labgrid-fcefyn-bananapi_bpi-r4=/srv/tftp/firmwares/bananapi_bpi-r4/libremesh/lime-24.10.5-mediatek-filogic-bananapi_bpi-r4-initramfs-recovery.itb,labgrid-fcefyn-librerouter_1=/srv/tftp/firmwares/librerouter_librerouter-v1/libremesh/lime-24.10.5-ath79-generic-librerouter_librerouter-v1-initramfs-kernel.bin,labgrid-fcefyn-belkin_rt3200_2=/srv/tftp/firmwares/belkin_rt3200/libremesh/lime-24.10.5-mediatek-mt7622-linksys_e8450-initramfs-kernel.bin" \
LG_MESH_KEEP_POWERED=1 \
uv run pytest tests/test_mesh.py -v

LG_MESH_KEEP_POWERED=1 leaves DUTs powered after the test (VLANs are still restored to isolated). On the host, manual SSH to a node still on VLAN 200: sudo labgrid-bound-connect vlan200 <mesh_ssh_ip> 22. From a developer machine, the local Host dut-* aliases do not work for VLAN 200; use the nested ProxyCommand from SSH access to DUTs - Remote developer.

Running from a developer machine¶

A developer can run tests from their own machine using LG_PROXY to reach the lab. LG_IMAGE points to a local file on the machine; Labgrid uploads it to the lab server automatically (see TFTP staging). Full guide: Developer quickstart.

# On the developer machine (not the lab host)
cd ~/pi/libremesh-tests
export LG_PLACE=labgrid-fcefyn-bananapi_bpi-r4
export LG_PROXY=labgrid-fcefyn
export LG_IMAGE=$HOME/builds/lime-24.10.5-mediatek-filogic-bananapi_bpi-r4-initramfs-recovery.itb
export LG_ENV=targets/bananapi_bpi-r4.yaml
uv run pytest tests/test_base.py -v --log-cli-level=INFO

The image is uploaded via SCP to /var/cache/labgrid/<user>/<sha256>/ on the lab host, and a symlink is created in the DUT's TFTP directory. Subsequent runs with the same image skip the upload (hash match).

Requirements: the developer's GitHub username must be in labnet.yaml access: list (Ansible fetches SSH keys from GitHub - see openwrt-tests onboarding), and the lab host must be reachable via LG_PROXY.

The test suite locates labnet.yaml via LABNET_PATH, OPENWRT_TESTS_DIR/labnet.yaml, or ../openwrt-tests/labnet.yaml (sibling of libremesh-tests). See libremesh-tests CONTRIBUTING_LAB.

Using CI-built firmware in tests¶

The build-and-test-libremesh.yml workflow builds a firmware artifact and runs flash_and_test automatically. If you want to run the tests manually with a firmware built by CI:

Go to GitHub → Actions → Build LibreMesh and Test on DUT → select a run
Download the artifact firmware-<dut>-<lime_ref>-<sha>
Extract the .bin or .itb file and point LG_IMAGE to it:

LG_PLACE=labgrid-fcefyn-belkin_rt3200_1 \
LG_IMAGE=/path/to/openwrt-23.05.5-mediatek-mt7622-linksys_e8450-squashfs-sysupgrade.bin \
uv run pytest tests/ -v

Artifacts are kept for 7 days. For automated end-to-end runs, trigger the workflow directly — it handles the full build → flash → test pipeline without manual steps.

Remote coordinator access (openwrt-tests)¶

# The coordinator runs locally on the lab host (loopback :20408).
# GitHub-hosted openwrt-tests runners reach it via LG_PROXY (SSH tunnel through the upstream VM).
# LG_PROXY is set in the CI runner env.
export LG_PROXY=ssh://labgrid-coordinator:51818

# List available places
labgrid-client places

# Reserve a specific FCEFYN place
labgrid-client -p labgrid-fcefyn-openwrt_one reserve --wait --token mytoken

Manual console or SSH access follows the same rule: use only places that are free or already reserved by you, never a place currently held by CI or another remote developer.