Behaviour of 2 nodes claster

Hi guys.
I have 2 nodes claster.
Which behaviour does Couchbase have If one node goes down?
As I think in such situation claster can’t work correctly… But is it possible to connect to last available node?
(I try to use libcouchbase library to do it, but I got network error).

And one more question. When one node goes down I see, that another node goes to pending state after some time.
When and why it happens?

If one node of a two node cluster goes down and you have replication, you should be able to manually failover, but not automatically. What version are you running at the cluster side?

@ingenthr, we use 2.5.1 version.
And I’m interesting in how Couchbase works if I won’t do a manual rebalance.
Is it a correct behaviour then last available node goes to pending after some time and It im[possible to connect to it?