Tailscale is continuously crashing and starting every 2-3 seconds. Is this a concern?
6 replies
tomspark wrote:
Good catch — a service bouncing every 2-3 seconds isn't normal, so let's sort it. The usual cause is a Tailscale admin-console setting: go to https://login.tailscale.com/admin/settings/features and make sure BOTH "HTTPS Certificates" and "Tailscale Serve" are turned on. If either is off, the container can't finish starting and keeps exiting and restarting — exactly what you're seeing. Flip both on, then run sudo sparkbox restart tailscale. If it's still looping after that, grab the reason with: sudo docker logs sb-tailscale --tail 30 and paste the last few lines — that'll tell us if it's the auth key instead. What does it show?
hairychinchilla69 wrote:
The entire media stack is now doing what tailscale was doing last night. I dont know what I did to break stuff.
tomspark wrote:
Deep breath — you didn't break anything permanent, this is recoverable. When the whole media stack cycles together like that, it's almost always the one VPN container they all run through restarting and dragging the rest with it. First, bring everything back up cleanly in the right order: sudo sparkbox up Give it 2-3 minutes to settle. If it's STILL cycling after that, paste the top of: docker ps -a so I can see which container is actually restarting (I suspect the VPN piece), and we'll fix that one thing.
hairychinchilla69 wrote:
0806c7c0eff5 ghcr.io/wg-easy/wg-easy:14 "docker-entrypoint.s…" 30 minutes ago Up 9 minutes (healthy) 0.0.0.0:51820-51820/udp, [::]:51820-51820/udp, 0.0.0.0:51821-51821/tcp, [::]:51821-51821/tcp sb-wg-easy 0273211a89ac tailscale/tailscale:v1.96.5 "/usr/local/bin/cont…" 30 minutes ago Up 13 seconds (healthy) sb-tailscale 463828a91718 lscr.io/linuxserver/speedtest-tracker:1.14.1 "/init" 30 minutes ago Up 9 minutes (healthy) 443/tcp, 0.0.0.0:8765-80/tcp, [::]:8765-80/tcp sb-speedtest 4c4de617d2ed pihole/pihole:2026.05.0 "start.sh" 30 minutes ago Up 9 minutes (healthy) 67/udp, 0.0.0.0:53-53/tcp, 0.0.0.0:53-53/udp, [::]:53-53/tcp, [::]:53-53/udp, 123/udp, 443/tc
tomspark wrote:
Found it — and it's all one thing, not a pile of separate breakages. Your box's SSD filled up (you spotted it on the dashboard), and the media apps can't run on a full disk, so they loop and show unhealthy — while Pi-hole and Tailscale survive because they barely write anything. Nothing's actually broken. I've put the step-by-step on your other thread ("I'm having issues") so we keep it in one place — short version: free up some disk space and the whole stack recovers on its own. No reinstall needed.