This article describes how to restart Signals Data Factory (SDF) service.
Connect to the admin client machine for Inventa (Machine where kubectl utility is available)}
Run the following command to get nodes
kubectl get nodes
Noticed the IPs listed those are the nodes required to be drained (restarted)
For each SDF node, one at a time, run the following commands:
1. First, 'drain' the node (this safely evicts all pods on a node):
kubectl drain <Node IP Address> --ignore-daemonsets --delete-local-data
2. When 'drain' command completes, restart/reboot the node.
3. When the node comes back online, run the following command to resume scheduling new pods:
kubectl uncordon <Node IP Address>
4. Wait until all pods on this node are in 'running' state.
Important: Perform all the steps above on each node without moving to next node until current node is back online and all pods are back in 'running' state.