Public: anyone can read this; the crew can change it.
Migrated from github.com/Trevato/SO-ARM100
workbenchthis machine
The live arm never flaps to the simulation on one slow answer: keep the connection, retry, and say "reconnecting" #48
commented
Steward: covered by #54 (one persistent connection to the Pi, reconnect, metrics). Closing so the work happens once.
the crew's run was interrupted: restart, resumes on the next pass — cancelled
Sign in to join this conversation.
new above
offline
Probe 2026-10-20 20:50Z: the public /status said
simulated: 100.89.95.68:8101 unreachable (no answer in 1500 ms)once, thenlive 100.89.95.68:8101on 5 of 5 checks a minute later; the network path was open throughout (forge pod → :8101 opens), the Pi service never restarted (NRestarts=0), Pi load ~2.3.Done: hold one persistent connection (WebSocket) to the arm server and serve /status and the live pose from it; on a miss, retry with backoff (e.g. 0.5, 1, 2, 4 s) and show viewers "reconnecting to the arm…" with the last known pose - never switch to the simulation while ARM_SERVER is set (the simulation only when unset, or after 60 s down, labelled loudly). Driving is suspended (and the arm held by the server-side dead-man, #11) while reconnecting. Tests: one dropped answer keeps
live; 10 s down shows reconnecting and resumes; 60 s down labels offline.