CORTX-34284: Do not count cortx-consul-client pods in status check (#380

)

status-cortx-cloud.sh was checking the total count of consul pods
against the sum of expected server pods (max(3, #worker-nodes))
and expected client pods.  However, the expected number is based
on the number of worker nodes not tainted w/ NoSchedule.  This
does not work as expected when a node that is normally a worker
node is tainted with NoSchedule.

To fix I have stopped checking for the expected number of
cortx-consul-client pods.  This is not a useful check anyway.

Signed-off-by: Walter Lopatka <walter.lopatka@seagate.com>

Signed-off-by: Walter Lopatka <walter.lopatka@seagate.com>
Co-authored-by: Rick Osowski <rosowski@gmail.com>

k8_cortx_cloud/status-cortx-cloud.sh

@@ -983,7 +983,7 @@ else
 fi
 
 # Check Pods
-num_items=$(( num_replicas + num_worker_nodes ))
+num_items="${num_replicas}"
 count=0
 msg_info "| Checking Pods |"
 while IFS= read -r line; do
@@ -997,7 +997,7 @@ while IFS= read -r line; do
         msg_passed
         count=$((count+1))
     fi
-done < <(kubectl get pods --namespace="${namespace}" --selector=${consul_selector} --no-headers)
+done < <(kubectl get pods --namespace="${namespace}" --selector="${consul_selector}",component=server --no-headers)
 
 if [[ ${num_items} -eq ${count} ]]; then
     msg_overall_passed