My sprite util-01 (org raymond-yee) persistently returns 502 on any connection:
% sprite exec -o raymond-yee -s util-01 -- echo hello
Error: failed to start sprite command: failed to connect: websocket: bad handshake (HTTP 502)
All 5 other sprites in the same org work fine. API (sprite ls) works. Retried multiple times — same result. Appears to be a stuck VM state. Is there a way to force-restart, or does this need server-side help?
This is very common for me as well. I wish they were more reliable or Fly would post some statment of what to expect with 502.. Is this growing pains, or something developers should account for.
I’ve been getting this as well. For what it’s worth, I was just able to revive a sprite that had seemingly gotten into this broken state by hitting its http service url a few times. It gave me a 502 the first couple of times, but then it woke back up and console/exec started working as well.
Interestingly, the sprite didn’t actually have any http services registered…the http endpoint returned an empty response body with a 502 status code when the sprite was “dead”, but once it woke back up the endpoint produces a 502 with a plain text “Bad Gateway” response body (since there was no service to respond).
I don’t have any other sprites in this broken state to test against, so I can’t say whether this workaround will bring back broken sprites consistently, though.
Thanks for your tip. I tried waking up my poor sprite by accessing the HTTP endpoint several times but to no avail. Still hoping to hear from a fly.io sys admin.
Same issue, sprite was working fine earlier and now returning 502. https://api.sprites.dev/v1/sprites returns that the sprite is cold but logs, services start, services stop all return 502
I’ve got multiple sprites experiencing this as well. Hitting the sprite URL hasn’t helped unfortunately.
I can now get in now after upgrading to the latest version of sprite.
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.