FYI to those interested, we contacted support and after some digging they fouind a stale smf table entry blocking new nodes trying to joing the cluster.
Although Nodes 3 and 4 had the same version of ONTAP as the existing cluster they would not join because there was a stale entry in the smf tables.
xxxxxxx::*> debug smdb table cluster_version_replicated show
uuid generation major minor version-string date ontapi-major ontapi-minor is-image-same state
------------------------------------ ---------- ----- ----- -------------------------------------------------- ------------------------ ------------ ------------ ------------- -----
3101b6df-7cec-11e5-8e37-00a0985f3fc6 8 3 0 NetApp Release 8.3P1: Tue Apr 07 16:05:35 PDT 2015 Tue Apr 07 12:05:35 2015 1 30 true none
57c64277-7cec-11e5-8e37-00a0985f3fc6 9 5 0 NetApp Release 9.5P5: Fri Jun 14 15:33:34 UTC 2019 Fri Jun 14 11:33:34 2019 1 150 true none
833939e5-7cd5-11e5-b363-396932647d67 9 5 0 NetApp Release 9.5P5: Fri Jun 14 15:33:34 UTC 2019 Fri Jun 14 11:33:34 2019 1 150 true none
f4270dd4-7cd3-11e5-a735-a570cc7c464a 9 5 0 NetApp Release 9.5P5: Fri Jun 14 15:33:34 UTC 2019 Fri Jun 14 11:33:34 2019 1 150 true none
4 entries were displayed.
There should be an entry for each node in the Cluster plus 1 entry for the Cluster.
We removed the 8.3P1 entry:
::*> debug smdb table cluster_version_replicated delete -uuid 3101b6df-7cec-11e5-8e37-00a0985f3fc6
Note: Do not edit the smf tables without guidance from NetApp Support.