Handle unknown errors while provisining cluster nodes

[yussufsh]

The cluster nodes would fail with API errors shown below but the nodes are getting created in the backend.

Error: failed to provision unknown error (status 504): {resp:0xc000cf87e0} 

  on modules/4_nodes/nodes.tf line 92, in resource "ibm_pi_instance" "master":
  92: resource "ibm_pi_instance" "master" {

Next apply throws an error:

Error: failed to provision {"description":"bad request: invalid name server name already exists for cloud-instance","error":"bad request"}

  on modules/4_nodes/nodes.tf line 92, in resource "ibm_pi_instance" "master":
  92: resource "ibm_pi_instance" "master" {

Need to add the check for such cases where created resource can be imported again. This will fix the issues where user do not need to go and delete the instances from the UI.