op-node: persisted sequencer state is not loaded after restart

[zhwrd]

Bug Description
The OP_NODE_RPC_ADMIN_STATE flag is used to persist the sequencer state so that sequencer maintains its active state after unexpected restart. This behaviour seems to be broken.

Steps to Reproduce

1. Set OP_NODE_RPC_ADMIN_STATE to a persistent path
2. Make admin_startSequencer call to start the sequencer.
3. Observe that OP_NODE_RPC_ADMIN_STATE/state is updated with {"sequencerStarted":true}
4. Restart op-node
5. Observe that the op-node is started with sequencing disabled and that OP_NODE_RPC_ADMIN_STATE/state is updated to {"sequencerStarted":false}

Expected behavior
op-node should load and apply the state stored in OP_NODE_RPC_ADMIN_STATE after restarts

Environment Information:
op-node v1.9.5 (although i think this has been an issue since at least a few 1.9 releases back)