HydraIssues

Problem

The HydraBody scheduled task on Windows bodies (schtasks /query /tn HydraBody) is configured with trigger At system start up only and no on-failure restart policy. If hydrabody crashes for any reason, it stays down until the machine is physically rebooted or an operator manually runs schtasks /run /tn HydraBody.

Evidence

This has now caused two incidents on boom-pickle-38 (rupelmonde):

• 2026-05-07 (issue #138): hydrabody crashed at 2026-05-06 05:24 with exit code 2. Stayed down until manually restarted. Visitor-facing symptom: Failed to find application rupelmonde-castle-viewer.
• 2026-05-11: hydrabody found not running again (tasklist /fi "IMAGENAME eq hydrabody.exe" returned no results). Manually restarted via schtasks /run /tn HydraBody. Same body, 4 days later.

Expected fix

Add an on-failure restart trigger to the HydraBody scheduled task so it auto-recovers from crashes without operator intervention. Options:

1. Configure the scheduled task with a restart-on-failure action (Task Scheduler UI: Settings → "If the task fails, restart every X minutes").
2. Switch to a Windows Service with sc failure recovery policy (sc failure HydraBody reset= 60 actions= restart/5000).
3. Add a watchdog loop in the hydrabody install recipe that restarts the process on non-zero exit.

Affected nodes

• boom-pickle-38 (node-74f9fbf2), rupelmonde — confirmed recurring
• Likely affects all Windows bodies with the same scheduled task setup

Related

• Issue #138 (closed) — first occurrence, documented the manual fix