openvidu/openvidu-deployment/pro/ha
Piwccle 5d0a4ac300 openvidu-deployment: Media Node health watchdog: withdraw the unhealthy mark when the replacement request fails
The replacement script marks the node unhealthy (/run/openvidu-unhealthy) before it asks
the cloud for a new node, so that the node then leaves without waiting for its sessions.
When the request failed (a cloud API error, a missing permission) the mark stayed: the
node remained a member and, if LiveKit recovered, a later normal scale-in would drop its
sessions instead of draining them. The mark is now withdrawn on any failed exit of the
request.
2026-10-16 13:02:16 +02:00
..
aws openvidu-deployment: Media Node health watchdog: withdraw the unhealthy mark when the replacement request fails 2026-10-16 13:02:16 +02:00
azure openvidu-deployment: Media Node health watchdog: withdraw the unhealthy mark when the replacement request fails 2026-10-16 13:02:16 +02:00
gcp openvidu-deployment: Media Node health watchdog: withdraw the unhealthy mark when the replacement request fails 2026-10-16 13:02:16 +02:00
install_ov_master_node.sh openvidu-deployment: add LIVEKIT_EGRESS_SERVER_PRO_IMAGE env var 2026-10-08 13:05:07 +02:00
install_ov_media_node.sh openvidu-deployment: add LIVEKIT_EGRESS_SERVER_PRO_IMAGE env var 2026-10-08 13:05:07 +02:00