Cluster fail to form after all nodes are restarted

[vasily-kirichenko]

• There is cluster of two nodes, both have port = 0.
• Run them, everything is OK:
• Restart both nodes

They are trying to connect to two nodes, one or two of which are old, dead ones:

Node 1:

[WARNING][25.01.2018 9:59:03][Thread 0008][[akka://test/system/cluster/core/daemon/joinSeedNodeProcess-1#1727576353]] Couldn't join seed nodes after [27] attempts, will try again. seed-nodes=[akka.tcp://test@kirichenko-xp:55403,akka.tcp://test@kirichenko-xp:55408,akka.tcp://test@kirichenko-xp:55708]

Node 2

[WARNING][25.01.2018 9:59:30][Thread 0006][[akka://test/system/cluster/core/daemon/joinSeedNodeProcess-1#670706284]] Couldn't join seed nodes after [33] attempts, will try again. seed-nodes=[akka.tcp://test@kirichenko-xp:55403,ak ka.tcp://test@kirichenko-xp:55408]

This is happening despite the old nodes are successfully removed from Consul:

[vasily-kirichenko]

The only way to form the cluster again is to wait until all nodes disappear from Consul and start them.