Nodes are stuck when one node are restarted and other node are new restarted in 4 nodes.

(just to keep recording) may be I am wrong.

4 nodes are working for consensus and then I restarted one node of 4 nodes and I removed other node's storage and start the node. but each of nodes have different block height and no more block height...

- 4 nodes worked (4/4)
- node1 was stopped (1/4)
- node2 was stopped and removed the storage of it (this node's height is started from 1) (1/4)
- node2 was started (1/4)
- node1 was started (0/4)

```
$http localhost:2821 | jq .block.height
69
anarch@booker ~
$http localhost:2822 | jq .block.height
69
anarch@booker ~
$http localhost:2823 | jq .block.height
65
anarch@booker ~
$http localhost:2824 | jq .block.height
1
```

Just this log messages are logging.
```
DBUG[11-08|21:40:24] failed to handle message module=noderunner node=GCDC.WBKY message="{\"H\":{\"hash\":\"JE4ossZ97StJpCwxnBzK5gPJkgAi7VUNthy6aVHVJxK8\",\"signature\":\"p6WKo2AsLrpKS7X1cyDHFrgtPRCWqAJ2BzVFKRSoiDN4refKQ1F4LcJY8yB2MXu4PBZ7EgpmEckXyjhJEGLLSKa\",\"proposer_signature\":\"5m4PmzDuW3bBmPyE9X9ofAR8DsKCedjv9as4DzB4QCvyTh3xUsvSBdiVgKL5ScxWnGjLTUaVRwZEM9zXjBLipodZ\"},\"B\":{\"confirmed\":\"2018-11-08T21:40:24.634450000+09:00\",\"proposed\":{\"confirmed\":\"2018-11-08T21:40:24.634023000+09:00\",\"proposer\":\"GCDCXYUTLFOZSRQ4K6DTZ3R7ZJMD6CHVL3ZM7KGOG52NOTNHWBKYPCIO\",\"voting_basis\":{\"round\":0,\"height\":65,\"block-hash\":\"ERHVMBGYrwdWjvjD1e7k3BYSWKQ7D9b8hWHyhgFdC7Tq\",\"total-txs\":65,\"total-ops\":130},\"transactions\":[],\"proposer_transaction\":{\"H\":{\"version\":\"\",\"created\":\"2018-11-08T21:40:24.633329000+09:00\",\"signature\":\"24pyjkMzVLh8QnjFYHir558S7kJEr1sGnBS7VobEdCRJxysXMBjTeuMjP9ABjXK5yxELp9vK2xH4M9bMDZg78gHP\"},\"B\":{\"source\":\"GDTEPFWEITKFHSUO44NQABY2XHRBBH2UBVGJ2ZJPDREIOL2F6RAEBJE4\",\"fee\":\"0\",\"sequence_id\":0,\"operations\":[{\"H\":{\"type\":\"collect-tx-fee\"},\"B\":{\"target\":\"GDYIHSHMDXJ4MXE35N4IMNC2X3Q3F665C5EX2JWHHCUW2PCFVXIFEE2C\",\"amount\":\"0\",\"txs\":0,\"block-height\":65,\"block-hash\":\"ERHVMBGYrwdWjvjD1e7k3BYSWKQ7D9b8hWHyhgFdC7Tq\",\"total-txs\":65,\"total-ops\":0}},{\"H\":{\"type\":\"inflation\"},\"B\":{\"target\":\"GDYIHSHMDXJ4MXE35N4IMNC2X3Q3F665C5EX2JWHHCUW2PCFVXIFEE2C\",\"amount\":\"1000000000000\",\"initial_balance\":\"10000000000000000000\",\"ratio\":\"0.00000010000000000\",\"block-height\":65,\"block-hash\":\"ERHVMBGYrwdWjvjD1e7k3BYSWKQ7D9b8hWHyhgFdC7Tq\",\"total-txs\":65,\"total-ops\":0}}]}}},\"source\":\"GDTEPFWEITKFHSUO44NQABY2XHRBBH2UBVGJ2ZJPDREIOL2F6RAEBJE4\",\"state\":3,\"vote\":\"EXPIRED\",\"reason\":null}}" error="{\"code\":142,\"message\":\"ballot already finished\",\"data\":{}}" caller=node_runner.go:427
```

The failure is logged from node_runner.go:427; start by reproducing the four-node stop, storage removal, and restart sequence while recording the HTTP block heights. Inspect the “ballot already finished” handling and compare the node logs. Done means the restarted nodes resume consensus and converge on the same block height.

Written by the indexing model from the issue text.