Since the single namespace change we are not able to startup 44-19 containers without a couple restarts.
-PullPolicy: Always
+PullPolicy: IfNotPresent

Then we can leverage the already pulled containers vi the prepull script

I found the issue with onap master - likely not related to OOM-722 but we take twice as long to startup - usually 20 min on a 16vCore/128G vm - now taking 40m - so I increased the wait time and run the cd build every 2 hours - doing a test on the side on a 96core/384G vm to find the new short time - By checking "top" I see a lot more python processes pushing all 16 cores to 100% even 30 min in - so this may be this issue (we are CPU bound now - not RAM bound)

ubuntu@ip-172-31-15-14:~/oom/kubernetes/oneclick$ kubectl get pods --all-namespaces | grep 0/
onap aaf-6c64db8fdd-9h6fh 0/1 Running 0 34m
onap consul-agent-56499f9db5-lcskz 0/1 CrashLoopBackOff 9 40m
onap consul-server-695997745c-7jhj9 0/1 CrashLoopBackOff 10 40m
onap vfc-ztevnfmdriver-fcf4ddf68-tmwwn 0/1 ErrImagePull 0 34m
onap vnfsdk-refrepo-55f544c5f5-dg6g6 0/1 ErrImagePull 0 34m

Another issue on one of my VMs' a stuck nodejs instance
40703 root 20 0 1361184 77612 2604 S 1223 0.1 11422:21 nodjs -c /bin/config.json