Op-Conductor Localnet (hemilabs#267)

* prove out op-conductor in localnet

set up 3 sequencers, each controlled by op-conductors p2p'd via raft consensus

update localnet tests CI to kill 1 op-node, another should take over sequencing

decrease the initial btc blocks because 3000 is excessive (this will need testing over time)

updated l2 block time to 2 seconds, instead of 1, it's easier to follow this way IMO

* test fixes

* remove references of op-geth-l1 in docker-compose

.github/workflows/localnet-test.yml

@@ -35,9 +35,15 @@ jobs:
       - name: "Download and verify dependencies"
         run: make deps
 
+      - name: "build localnet"
+        run: docker compose -f ./e2e/docker-compose.yml build
+
       - name: "run localnet"
         run: docker compose -f ./e2e/docker-compose.yml up -d
 
+      - name: "kill an op-node after a minute"
+        run: sleep 60 && docker compose -f ./e2e/docker-compose.yml down op-node
+
       - name: "get localnet stats"
         working-directory: ./e2e/monitor
         # XXX should this be a make command?

e2e/deploy-config.json

@@ -1,7 +1,7 @@
 {
     "l1ChainID": 1337,
     "l2ChainID": 901,
-    "l2BlockTime": 1,
+    "l2BlockTime": 2,
     "maxSequencerDrift": 300,
     "sequencerWindowSize": 200,
     "channelTimeout": 120,