Ugreen updated the nas after restart cloudflare did not start on its own

Posted by franciscus

I turned my nas on and cloudflare did not start, I had to do it on the homepage, button "resume" . Ugreen just updated the nas to a newer version and after the restart the same cloudflare did not start, so I had to press the button "resume" again on the homepage(version 1) On the sparkbox version 3 there is no such button. But why does cloudflare start on its own? My question: status=running restart=on-failure started=2026-09-28T15:42:34.570304222Z SparkBox AI replied: So pressing Homepage’s Resume starts it manually; it doesn’t fix the reboot behaviour. It’s running now, and its logs show successful Cloudflare connections. sparkbox doctor also passes, but that checks the current state—not why this container has that restart policy. I don’t have a documented SparkBox fix for changing it, so I won’t guess or suggest editing generated config that SparkBox may overwrite. Use Share to demox so Tom can check why cloudflared isn’t being started automatically after the UGREEN reboot. Current checks: - sb-syncthing: Up 12 minutes (healthy) - sb-wg-easy: Up 12 minutes (healthy) - sb-vaultwarden: Up 12 minutes (healthy) - sb-unpackerr: Up 12 minutes (healthy) - sb-tomsparkcal: Up 12 minutes (healthy) - sb-stirling-pdf: Up 12 minutes (healthy) - sb-paperless: Up 12 minutes (healthy) - sb-paperless-redis: Up 12 minutes (healthy) - sb-paperless-db: Up 12 minutes (healthy) - sb-navidrome: Up 12 minutes (healthy) - sb-uptime-kuma: Up 12 minutes (healthy) - sb-sonarr: Up 11 minutes (healthy) - sb-radarr: Up 11 minutes (healthy) - sb-prowlarr: Up 11 minutes (healthy) - sb-qbittorrent: Up 11 minutes (healthy) - sb-sabnzbd: Up 11 minutes (healthy) - sb-flaresolverr: Up 11 minutes (healthy) - sb-bazarr: Up 12 mi …(cut here — the full chat went to Tom) What I still need help with: (describe what didn't work or what's still broken)

1 replies

Chris wrote:

You found a real bug, and thanks for the clear report. Every other SparkBox app is set to come back whenever the box starts. The Cloudflare Tunnel was set to come back only after a crash, and a normal shutdown (switching the NAS off, or UGOS restarting it for an update) is not a crash, so it stayed stopped until you pressed Resume. The newer dashboard also read the stopped tunnel as a finished setup step, which is why it showed no Resume button. SparkBox 1.6.809 fixes it: the tunnel now starts with the rest of your apps after every restart. Your tunnel token and hostname routes stay exactly as they are. One check: update SparkBox (Updates page, SparkBox Core, Update SparkBox, or sudo sparkbox upgrade), then run sudo docker inspect -f '{{.HostConfig.RestartPolicy.Name}}' sb-cloudflared. It should print unless-stopped. From then on a NAS restart should bring the tunnel back without Resume. If it ever doesn't, reply here with what that command prints.