Merge pull request #49344 from shyamjvs/master-disk-size

Automatic merge from submit-queue

Auto-calculate master disk and root disk sizes in GCE

@gmarek PR https://github.com/kubernetes/kubernetes/pull/49282 didn't fix the issue because MASTER_DISK_SIZE was defaulting to 20GB in config-test.sh before being calculated inside get-master-disk-size() where you use pre-existing value if any.

It should be fixed by this now.

cluster/gce/config-common.sh

@@ -39,6 +39,32 @@ function get-master-size {
   echo "${suggested_master_size}"
 }
 
+# Vars assumed:
+#   NUM_NODES
+function get-master-root-disk-size() {
+  local suggested_master_root_disk_size="20GB"
+  if [[ "${NUM_NODES}" -gt "1000" ]]; then
+    suggested_master_root_disk_size="50GB"
+  fi
+  if [[ "${NUM_NODES}" -gt "2000" ]]; then
+    suggested_master_root_disk_size="100GB"
+  fi
+  echo "${suggested_master_root_disk_size}"
+}
+
+# Vars assumed:
+#   NUM_NODES
+function get-master-disk-size() {
+  local suggested_master_disk_size="20GB"
+  if [[ "${NUM_NODES}" -gt "1000" ]]; then
+    suggested_master_disk_size="100GB"
+  fi
+  if [[ "${NUM_NODES}" -gt "2000" ]]; then
+    suggested_master_disk_size="200GB"
+  fi
+  echo "${suggested_master_disk_size}"
+}
+
 function get-node-ip-range {
   if [[ -n "${NODE_IP_RANGE:-}" ]]; then
     >&2 echo "Using user provided NODE_IP_RANGE: ${NODE_IP_RANGE}"

cluster/gce/config-default.sh

@@ -31,7 +31,8 @@ NODE_SIZE=${NODE_SIZE:-n1-standard-2}
 NUM_NODES=${NUM_NODES:-3}
 MASTER_SIZE=${MASTER_SIZE:-n1-standard-$(get-master-size)}
 MASTER_DISK_TYPE=pd-ssd
-MASTER_DISK_SIZE=${MASTER_DISK_SIZE:-20GB}
+MASTER_DISK_SIZE=${MASTER_DISK_SIZE:-$(get-master-disk-size)}
+MASTER_ROOT_DISK_SIZE=${MASTER_ROOT_DISK_SIZE:-$(get-master-root-disk-size)}
 NODE_DISK_TYPE=${NODE_DISK_TYPE:-pd-standard}
 NODE_DISK_SIZE=${NODE_DISK_SIZE:-100GB}
 NODE_LOCAL_SSDS=${NODE_LOCAL_SSDS:-0}

cluster/gce/config-test.sh

@@ -31,7 +31,8 @@ NODE_SIZE=${NODE_SIZE:-n1-standard-2}
 NUM_NODES=${NUM_NODES:-3}
 MASTER_SIZE=${MASTER_SIZE:-n1-standard-$(get-master-size)}
 MASTER_DISK_TYPE=pd-ssd
-MASTER_DISK_SIZE=${MASTER_DISK_SIZE:-20GB}
+MASTER_DISK_SIZE=${MASTER_DISK_SIZE:-$(get-master-disk-size)}
+MASTER_ROOT_DISK_SIZE=${MASTER_ROOT_DISK_SIZE:-$(get-master-root-disk-size)}
 NODE_DISK_TYPE=${NODE_DISK_TYPE:-pd-standard}
 NODE_DISK_SIZE=${NODE_DISK_SIZE:-100GB}
 NODE_LOCAL_SSDS=${NODE_LOCAL_SSDS:-0}

cluster/gce/gci/master-helper.sh

@@ -116,7 +116,7 @@ function create-master-instance-internal() {
     --scopes "storage-ro,compute-rw,monitoring,logging-write" \
     --metadata-from-file "${metadata}" \
     --disk "${disk}" \
-    --boot-disk-size "${MASTER_ROOT_DISK_SIZE:-10}" \
+    --boot-disk-size "${MASTER_ROOT_DISK_SIZE}" \
     ${preemptible_master} \
     ${network}
 }

cluster/gce/util.sh

@@ -949,35 +949,6 @@ function delete-subnetworks() {
   fi
 }
 
-# Assumes:
-#   NUM_NODES
-# Sets:
-#   MASTER_ROOT_DISK_SIZE
-function get-master-root-disk-size() {
-  if [[ "${NUM_NODES}" -le "1000" ]]; then
-    export MASTER_ROOT_DISK_SIZE="${MASTER_ROOT_DISK_SIZE:-20}"
-  elif [[ "${NUM_NODES}" -le "2000" ]]; then
-    export MASTER_ROOT_DISK_SIZE="${MASTER_ROOT_DISK_SIZE:-50}"
-  else
-    export MASTER_ROOT_DISK_SIZE="${MASTER_ROOT_DISK_SIZE:-100}"
-  fi
-}
-
-# Assumes:
-#   NUM_NODES
-# Sets:
-#   MASTER_DISK_SIZE
-function get-master-disk-size() {
-  if [[ "${NUM_NODES}" -le "1000" ]]; then
-    export MASTER_DISK_SIZE="${MASTER_DISK_SIZE:-20GB}"
-  elif [[ "${NUM_NODES}" -le "2000" ]]; then
-    export MASTER_DISK_SIZE="${MASTER_DISK_SIZE:-100GB}"
-  else
-    export MASTER_DISK_SIZE="${MASTER_DISK_SIZE:-200GB}"
-  fi
-}
-
-
 # Generates SSL certificates for etcd cluster. Uses cfssl program.
 #
 # Assumed vars:
@@ -1022,7 +993,6 @@ function create-master() {
 
   # We have to make sure the disk is created before creating the master VM, so
   # run this in the foreground.
-  get-master-disk-size
   gcloud compute disks create "${MASTER_NAME}-pd" \
     --project "${PROJECT}" \
     --zone "${ZONE}" \
@@ -1074,9 +1044,6 @@ function create-master() {
   create-certs "${MASTER_RESERVED_IP}"
   create-etcd-certs ${MASTER_NAME}
 
-  # Sets MASTER_ROOT_DISK_SIZE that is used by create-master-instance
-  get-master-root-disk-size
-
   create-master-instance "${MASTER_RESERVED_IP}" &
 }
 
@@ -1135,16 +1102,12 @@ function replicate-master() {
 
   # We have to make sure the disk is created before creating the master VM, so
   # run this in the foreground.
-  get-master-disk-size
   gcloud compute disks create "${REPLICA_NAME}-pd" \
     --project "${PROJECT}" \
     --zone "${ZONE}" \
     --type "${MASTER_DISK_TYPE}" \
     --size "${MASTER_DISK_SIZE}"
 
-  # Sets MASTER_ROOT_DISK_SIZE that is used by create-master-instance
-  get-master-root-disk-size
-
   local existing_master_replicas="$(get-all-replica-names)"
   replicate-master-instance "${EXISTING_MASTER_ZONE}" "${EXISTING_MASTER_NAME}" "${existing_master_replicas}"