Sprite is cold and won't start up

Questions / Help
1

I was in the middle of using my sprite when it ‘went cold’. It now says it’s cold and trying to go to the sprite URL gives me a 502 error “The Sprite isn’t answering”.

  • sprite exec fails with read tcp …->169.155.48.226:443: i/o timeout
  • sprite checkpoint list returns 500 overlay manager not configured
  • Stuck cold, last_running_at: null, wake attempts time out
  • Restarting the sprite via the dashboard does nothing

Second time this has happened. After the first time it did come back but only after a restart.

2

The logs show cgroup monitors for juicefs and litestream — the filesystem and its SQLite replication. Boot is stuck acquiring a DB lease, and sprite checkpoint list fails with overlay manager not configured.

Starting system boot sequence using unified service manager
All system services registered with unified manager
Starting API server addr=0.0.0.0:7778
Acquiring database lease...
Attempting lease PUT  ifMatch=250c93d8d454c96b0f9c76ddfc35358e
services API socket server started
Successfully joined admin channel
Sprite assignment queued until DB is ready   ← last line, then nothing
4

Hi there - I’m seeing one sprite under your org that seems to be up and running. Can you confirm that you’re set here? Or if you’re still seeing issues, what’s the sprite name? Happy to dig in further.

5

I left it for a couple of days and nothing changed so I ended up deleting it and creating a new one

6

Gotcha, sorry for the trouble here! We’re currently working on overhauling the Sprites storage stack with Sprites Block Device (SBD), the new storage layer, which should make the Sprites filesystem faster and more reliable. That should be landing in the not-too-distant future, and it’ll hopefully prevent these sorts of issues going forward.