bootstrap join consumes unnecessary time on passwordless checks for the existing cluster

[zzhou1]

Could bootstrap join skips "Checking passwordless" among the existing 3-node cluster, when the 4th new node joins the cluster? It's noticeable time consuming when the cluster is large, and is not a good enough user experience. Example as below,

hanode4: # crm cluster join -yc hanode1                                        
[...]
INFO: Checking passwordless for hacluster between hanode1 and hanode2
INFO: Checking passwordless for hacluster between hanode1 and hanode3
INFO: Checking passwordless for hacluster between hanode2 and hanode1
INFO: Checking passwordless for hacluster between hanode2 and hanode3
INFO: Checking passwordless for hacluster between hanode3 and hanode1
INFO: Checking passwordless for hacluster between hanode3 and hanode2
[...]

[nicholasyang2022]

I think it is a workaround for certain use case related to -N option used in bootstrap, which does not swap public keys correctly. I will try to find the root cause.

[nicholasyang2022]

Fixed in #1261.