etcd: Cannot restart after shutdown / Ctrl+C (etcdmain: database file does not match with snapshot)

I have a cluster of 3 and I’m testing failover scenarios right now. I simulate power outages (shutting of a VM) and service failures (Ctrl+C on etcd). As of now, shutting off an instance has terrible consequences, as it cannot recover by simply rebooting a machine / restarting the service:

$ etcd
// ...runs...
<Press Ctrl+C>

$ etcd
2016-07-03 02:05:21.753663 I | flags: recognized and used environment variable ETCD_ADVERTISE_CLIENT_URLS=http://rly4:2379
2016-07-03 02:05:21.754254 I | flags: recognized and used environment variable ETCD_DATA_DIR=/var/lib/etcd/data
2016-07-03 02:05:21.754550 I | flags: recognized and used environment variable ETCD_DEBUG=true
2016-07-03 02:05:21.754805 I | flags: recognized and used environment variable ETCD_INITIAL_ADVERTISE_PEER_URLS=http://rly4:2380
2016-07-03 02:05:21.755066 I | flags: recognized and used environment variable ETCD_INITIAL_CLUSTER=rly4=http://rly4:2380,rly1=http://rly1:2380,rly5=http://rly5:2380
2016-07-03 02:05:21.755317 I | flags: recognized and used environment variable ETCD_INITIAL_CLUSTER_STATE=existing
2016-07-03 02:05:21.755585 I | flags: recognized and used environment variable ETCD_INITIAL_CLUSTER_TOKEN=rly-cluster-1
2016-07-03 02:05:21.755833 I | flags: recognized and used environment variable ETCD_LISTEN_CLIENT_URLS=http://rly4:2379,http://127.0.0.1:2379
2016-07-03 02:05:21.756038 I | flags: recognized and used environment variable ETCD_LISTEN_PEER_URLS=http://rly4:2380
2016-07-03 02:05:21.756276 I | flags: recognized and used environment variable ETCD_NAME=rly4
2016-07-03 02:05:21.756525 I | flags: recognized and used environment variable ETCD_WAL_DIR=/var/lib/etcd/wal
2016-07-03 02:05:21.756784 I | etcdmain: etcd Version: 3.0.0
2016-07-03 02:05:21.757038 I | etcdmain: Git SHA: 6f48bda
2016-07-03 02:05:21.757261 I | etcdmain: Go Version: go1.6.2
2016-07-03 02:05:21.757478 I | etcdmain: Go OS/Arch: linux/amd64
2016-07-03 02:05:21.757696 I | etcdmain: setting maximum number of CPUs to 1, total number of available CPUs is 1
2016-07-03 02:05:21.757950 N | etcdmain: the server is already initialized as member before, starting as etcd member...
2016-07-03 02:05:21.761023 I | etcdmain: listening for peers on http://rly4:2380
2016-07-03 02:05:21.761065 I | etcdmain: listening for client requests on 127.0.0.1:2379
2016-07-03 02:05:21.763682 I | etcdmain: listening for client requests on rly4:2379
2016-07-03 02:05:21.804445 I | etcdserver: recovered store from snapshot at index 141365
2016-07-03 02:05:21.804844 I | etcdserver: name = rly4
2016-07-03 02:05:21.805118 I | etcdserver: data dir = /var/lib/etcd/data
2016-07-03 02:05:21.805389 I | etcdserver: member dir = /var/lib/etcd/data/member
2016-07-03 02:05:21.805661 I | etcdserver: dedicated WAL dir = /var/lib/etcd/wal
2016-07-03 02:05:21.805891 I | etcdserver: heartbeat = 100ms
2016-07-03 02:05:21.806169 I | etcdserver: election = 1000ms
2016-07-03 02:05:21.806428 I | etcdserver: snapshot count = 10000
2016-07-03 02:05:21.806696 I | etcdserver: advertise client URLs = http://rly4:2379
2016-07-03 02:05:21.807377 I | etcdserver: restarting member 89889c482441ec6e in cluster d0ce954dc5f082d0 at commit index 141549
2016-07-03 02:05:21.807753 I | raft: 89889c482441ec6e became follower at term 1966
2016-07-03 02:05:21.808041 I | raft: newRaft 89889c482441ec6e [peers: [89889c482441ec6e,d27323ab9b295f50,e790e6b697b3c219], term: 1966, commit: 141549, applied: 141365, lastindex: 141549, lastterm: 1966]
2016-07-03 02:05:21.808406 I | membership: added member d27323ab9b295f50 [http://rly5:2380] to cluster d0ce954dc5f082d0 from store
2016-07-03 02:05:21.808751 I | membership: added member e790e6b697b3c219 [http://rly1:2380] to cluster d0ce954dc5f082d0 from store
2016-07-03 02:05:21.809021 I | membership: added member 89889c482441ec6e [http://rly4:2380] to cluster d0ce954dc5f082d0 from store
2016-07-03 02:05:21.809295 I | membership: set the cluster version to 3.0 from store
2016-07-03 02:05:21.818080 I | etcdmain: stopping listening for client requests on rly4:2379
2016-07-03 02:05:21.818511 I | etcdmain: stopping listening for client requests on rly4:2379
2016-07-03 02:05:21.818707 I | etcdmain: stopping listening for peers on http://rly4:2380
2016-07-03 02:05:21.818894 C | etcdmain: database file (/var/lib/etcd/data/member/snap/db index 140014) does not match with snapshot (index 141365).

$ find
./data
./data/member
./data/member/snap
./data/member/snap/00000000000007ae-0000000000022835.snap
./data/member/snap/db
./wal
./wal/0000000000000000-0000000000000000.wal
./wal/0.tmp

The only way to recover from this right now is to delete the data+wal dir, remove the member, re-add it and then restart the daemon. However, I fear that if all fail at the same time (power failure), I will lose all data.