I am following the instructions from here: https://cloud.google.com/kubernetes-engine/docs/quickstart
When running the gcloud container clusters create [CLUSTER_NAME] command, it's been running for well over 15 minutes. (Yesterday, a custom cluster that I was creating with much large machine images started up consistently in under 3 minutes.) Is this variation normal? Is is even normal that a Kubernetes cluster creation may take well over 15 minutes (and counting)? How can one tell if "something is going wrong"? I don't see any logs in the command line, and when I run:
gcloud container clusters list
all I see is the word "PROVISIONING":
NAME LOCATION MASTER_VERSION MASTER_IP MACHINE_TYPE NODE_VERSION NUM_NODES STATUS deepsearchcluster us-central1-a 1.10.9-gke.5 35.239.29.229 n1-standard-2 1.10.9-gke.5 1 RUNNING deepsearchgitlab us-central1-a 1.9.7-gke.11 35.225.135.57 n1-standard-4 1.9.7-gke.7 * 1 RUNNING demo us-central1-a 1.10.9-gke.5 n1-standard-1 1.10.9-gke.5 3 PROVISIONING
UPDATE: It finally started after > 28 minutes....this is crazy....Is this "normal" variation? Did it take a long time because the "default" cluster uses tiny (ie 1 vCPU) nodes with little RAM? Any suggestions on how I can minimize this problem?
The problem appears to be because of creating multiple node pools all from a single command separated by "&&" operand and then running it all on Windows. (It may also occurr on Linux). Basically, because there were multiple commands and the command line is interactive, it was waiting for "user input" before moving on to the next command....Not sure if "--quiet" would help here, but will investigate.