after node is removed forcibly from cluster logs are spammed by side-transport messages

[aliher1911]

Describe the problem

Whenever node crashes irrecoverably and removed from the cluster closed timestamp side transport will pick up node id's from range descriptors and will keep trying to talk to them. This will continue possibly forever or till each node restart after all range descriptors are fixed.
This would show up in the logs as:

I220419 10:56:23.265470 1461 kv/kvserver/closedts/sidetransport/sender.go:785 ⋮ [n1,ctstream=3] 871  side-transport failed to connect to n3: failed to resolve n3: unable to look up descriptor for n3
I220419 10:56:23.271968 1460 kv/kvserver/closedts/sidetransport/sender.go:785 ⋮ [n1,ctstream=2] 872  side-transport failed to connect to n2: failed to resolve n2: unable to look up descriptor for n2
...
I220419 10:56:33.271807 1461 kv/kvserver/closedts/sidetransport/sender.go:785 ⋮ [n1,ctstream=3] 878  side-transport failed to connect to n3: failed to resolve n3: unable to look up descriptor for n3
I220419 10:56:33.277426 1460 kv/kvserver/closedts/sidetransport/sender.go:785 ⋮ [n1,ctstream=2] 879  side-transport failed to connect to n2: failed to resolve n2: unable to look up descriptor for n2
...
I220419 10:56:43.277381 1461 kv/kvserver/closedts/sidetransport/sender.go:785 ⋮ [n1,ctstream=3] 882  side-transport failed to connect to n3: failed to resolve n3: unable to look up descriptor for n3
I220419 10:56:43.281789 1460 kv/kvserver/closedts/sidetransport/sender.go:785 ⋮ [n1,ctstream=2] 883  side-transport failed to connect to n2: failed to resolve n2: unable to look up descriptor for n2

To Reproduce

Run a cluster with 5 nodes. Kill all of the nodes and perform recovery from quorum loss using nodes 1, 4, 5. Restart "fixed" nodes and observe the messages in logs.

Jira issue: CRDB-15826