The original intent of the shutdown lock was to prevent the hub from restarting a node to clear up out of sync curator locks.
We have fixed the underlying lock issue, and the hub no longer auto-restarts.
In addition, since prod clusters can have 4+ nodes, the impact of two nodes restarting at the same time is minimized.
The original intent of the shutdown lock was to prevent the hub from restarting a node to clear up out of sync curator locks.
We have fixed the underlying lock issue, and the hub no longer auto-restarts. In addition, since prod clusters can have 4+ nodes, the impact of two nodes restarting at the same time is minimized.