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

**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)

Start in the op-node startup path that handles OP_NODE_RPC_ADMIN_STATE and compare it with the admin_startSequencer persistence behavior. Reproduce the restart steps from the issue; done means a persisted {"sequencerStarted":true} state is loaded after restart and sequencing remains enabled instead of being overwritten as false.

Written by the indexing model from the issue text.