I tried to add a custom module to the dashboard and its failed. Now sparkbox crashed overnight (I assume during its auto-update) and throws this error while booting: [INFO] Starting SparkBox... [INFO] Active modules: audiobookshelf backup cloudflared core dashboard files media monitoring openwa pihole shelfarr speedtest stirling-pdf unpackerr vaultwarden [INFO] Prepared 1 module config dir(s) for PUID=1000 PGID=10 [INFO] Fetching images for module: openwa unable to get image 'https://github.com/rmyndharis/openwa.git': Error response from daemon: invalid reference format [WARN] pull attempt 2/3 (previous attempt failed) — retrying in 10s unable to get image 'https://github.com/rmyndharis/openwa.git': Error response from daemon: invalid reference format [WARN] pull attempt 3/3 (previous attempt failed) — retrying in 15s unable to get image 'https://github.com/rmyndharis/openwa.git': Error response from daemon: invalid reference format [ERROR] Failed to pull images for module 'openwa' after 3 attempts. [ERROR] Likely causes: no internet access from this host, Docker Hub rate limit, or insufficient disk space. [ERROR] If the pull output above ends in 'tls: bad record MAC' or 'unexpected EOF', the download is being corrupted on your network path (cable, extender, switch, router or ISP) — not by SparkBox or Docker. Try a different cable straight into the router; do not reinstall. [ERROR] Try: docker pull <image:tag manually, or check 'docker info' for disk pressure. [ERROR] Image prefetch failed for: openwa. Aborting before compose up to give a clean error.
9 replies
Chris wrote:
That error points to the custom module's image value, not an automatic update: Docker was given a GitHub repository URL where it expected a container image reference, so image fetching stops before SparkBox can bring the stack up. Please reply with your SparkBox version, which host/platform this is (for example, UGREEN/NAS, Linux, or Windows with WSL Docker Engine vs Docker Desktop), and only the custom module's image value as configured—not a full config or any tokens. Please don't reinstall or reset anything; that isn't needed to diagnose this and your files/config should be preserved.
EineBeBoP wrote:
SparkBox v1.6.808, ugreen nas. i dont know what you're asking for about the custom modules value. Id be happy to disable the custom module and not install it.
Chris wrote:
Thanks—that’s enough, and sorry my last question wasn’t clear. You don’t need to find or share an image value. The startup output points to the added custom module: Docker is being given a GitHub source URL where it expects a container image name, so SparkBox stops before it can bring the stack up. I understand you’d rather disable that module and not install it. How did you add it—through a SparkBox screen, or by editing a file? Please don’t remove or reinstall SparkBox or change other module settings; once I know which route, I’ll give you the safe, exact disable step.
EineBeBoP wrote:
through the dashboard adding this to the command box on the settings page. x-sparkbox: id: my-app openwa title: OpenWA description: A self-hosted WhatsApp HTTP API for developers who want to own their stack ram: "~200MB" what it needs; the memory check counts it services: my-app: image: https://github.com/rmyndharis/OpenWA.git containername: sb-openwa must start with sb- restart: unless-stopped ports: - "2785:2785" a host port no shipped app uses volumes: - ./config:/config keep state beside the compose file
Chris wrote:
Thanks, that explains what happened. This is coming from the custom module you added, not an automatic update: its image: value is a GitHub repository URL, but Docker needs a container image reference there, so Docker rejects it and SparkBox stops before bringing the stack up. I understand you want to remove that module and leave the rest alone. I can’t safely guess the dashboard’s removal control; please tell me the exact button or option labels shown with that command box (no need to paste the YAML again). Until we confirm the right control, please don’t reinstall or reset anything.
EineBeBoP wrote:
Im no longer able to access the command box as the dashboard wont load.
Chris wrote:
I’m sorry—I asked you to use the dashboard to remove the app even though it isn’t loading. The startup output shows SparkBox stopping while it tries to fetch the app: its image entry is a GitHub repository URL, not a Docker image reference, so startup aborts before the dashboard comes up. From an SSH session to your UGREEN NAS, run sudo sparkbox custom remove openwa and confirm when prompted; this stops and unlinks that custom app while keeping its folder under custom/.removed/. Then try the dashboard again and let me know if it still won’t load. I’ve corrected what I know so this doesn’t send the next person back to an unavailable dashboard control.
EineBeBoP wrote:
This has fixed it. Would you forward this to Tom so that similar issues for other users don't lead to a dashboard that wont load on the next "sparkbox up" command run?
Chris wrote:
I'm glad removing the custom app restored the dashboard, and thanks for flagging the broader recovery gap. I’ve recorded that a bad custom-app image can stop startup before the dashboard is available, so it can be reviewed and addressed for other users too. Your existing files were kept by the removal step.