Merge pull request #75 from hyder/v12

updated docs for new node pool parameter, changed default bastion shape

Author: hyder

docs/images/mixedworkload.png

@@ -87,12 +87,12 @@ If you want to use bastion host, set the parameter *create_bastion* to *true* in
 
 There are 2 additional parameters for the bastion:
 
-* package_update
-* package_upgrade
+* bastion_package_update
+* bastion_package_upgrade
 
-_package_update_ will update the apt database *if* you choose Ubuntu as the Linux distribution for the bastion host.
+_bastion_package_update_ will update the apt database *if* you choose Ubuntu as the Linux distribution for the bastion host.
 
-_package_upgrade_ will upgrade the bastion compute instance on first boot. If you choose Ubuntu for bastion host and you set _package_upgrade_ to *true*, you should also set the _package_update_ to *true*.
+_bastion_package_upgrade_ will upgrade the bastion compute instance on first boot. If you choose Ubuntu for bastion host and you set _bastion_package_upgrade_ to *true*, you should also set the _bastion_package_update_ to *true*.
 
 ****
 N.B. It is good and recommended practice to upgrade your bastion host to the latest packages to minimize the possibility of vulnerabilities. However, it will also take slightly longer before the bastion host is available.
@@ -146,13 +146,14 @@ Calico enables network policy in Kubernetes clusters across the cloud. To instal
 
 {uri-metricserver}[Kubernetes Metrics Server] can be installed by setting the parameter *install_metricserver = true* in terraform.tfvars. By default, the latest version is installed in kube-system namespace. This is required if you need to use Horizontal Pod Autoscaling.
 
-=== Scaling the number of worker nodes
+=== Scaling the node pools
 
-Set the parameter *node_pool_quantity_per_subnet* to the desired quantity.For single AD region a minimum quantity of 2 will get created. This is helpful to utilize the OCI Fault Domains in single AD regions. Refer to {uri-topology}#fault-domains[Fault Domain].
+There are 2 ways you can scale the node pools:
 
-=== Scaling the number of node pools
+* add more node pools
+* increase the number of workers in a subnet per node pool.
 
-Set the parameter *node_pools* to the desired quantity. Refer to {uri-topology}#node-pools[Nodepool].
+Set the parameter *node_pools* to the desired quantities to scale the node pools accordingly. Refer to {uri-topology}#node-pools[Nodepool].
 
 === Accessing the Kubernetes dashboard
 

docs/images/np311.png

@@ -236,7 +236,7 @@ subnets = {
 |bastion_shape
 |The shape of bastion instance.
 |
-|VM.Standard2.1
+|VM.Standard.E2.1
 
 |bastion_access
 |CIDR block in the form of a string to which ssh access to the bastion must be restricted to. *_ANYWHERE_* is equivalent to 0.0.0.0/0 and allows ssh access from anywhere.
@@ -326,9 +326,19 @@ availability_domains = {
 |LATEST
 
 |node_pools
-|The number of node pools to create. Refer to {uri-topology}[topology] for more thorough examples.
-|
-|1
+|The number, shape and quantities per subnets of node pools to create. Refer to {uri-topology}[topology] for more thorough examples.
+|e.g.
+[source]
+----
+node_pools = {
+  "np1" = ["VM.Standard2.1", 1]
+}
+----
+|----
+node_pools = {
+  "np1" = ["VM.Standard2.1", 1]
+}
+----
 
 |node_pool_name_prefix
 |A string prefixed to the node pool name.
@@ -350,16 +360,6 @@ availability_domains = {
 |
 |7.6
 
-|node_pool_node_shape
-|The shape of worker nodes to provision.
-|
-|VM.Standard2.1
-
-|node_pool_quantity_per_subnet
-|Number of worker nodes by worker subnets in a node pool. Refer to {uri-topology}[topology] for more thorough examples.
-|
-|1
-
 |nodepool_topology
 a|The number of Availability Domains the node pools should span. Use 1 for single-AD regions and 3 for multiple-AD regions. _Topology 2 is experimental and is only used in multiple-AD regions_.
 

docs/images/np312.png

@@ -193,26 +193,58 @@ A node pool is a set of hosts within a cluster that all have the same configurat
 
 Node pools enable you to create pools of machines within a cluster that have different configurations. For example, you might create one pool of nodes in a cluster as virtual machines, and another pool of nodes as bare metal machines. A cluster must have a minimum of one node pool, but a node pool need not contain any worker nodes.
 
-****
-*N.B. As of this version, all node pools have the same configuration. You can manually add node pools of different configuration (e.g. shapes) after the cluster is created.*
-****
+When using this project to create the node pools, the following is done:
 
+* a number of node pools are created. The number of node pools created is equal to the number of elements in the node_pools parameter e.g.
 
-When using this project to create the node pools, the following is done:
+----
+node_pools = {
+  "np1" = ["VM.Standard2.1", 1]
+  "np2" = ["VM.Standard2.2", 1]
+}
+----
+
+will create 2 node pools (np1 and np2) whereas
+
+----
+node_pools = {
+  "np1" = ["VM.Standard2.1", 1]
+  "np2" = ["VM.Standard2.2", 1]
+  "np3" = ["VM.Standard2.4", 1]
+}
+----
 
-* a number of node pools are created. This is controlled by the node_pools parameter. By default, this value is 1.
+will create 3 node pools (np1, np2 and np3).
 
 * the node pool names are generated by combining a prefix and the node pool number. The prefix is set by the node_pool_name_prefix parameter and has a default value of "np". The node pool names will therefore have names like np-1, np-2 and so on.
 
 * the Kubernetes version is set automatically to the same version as the cluster.
 
 * the image used is an Oracle Linux image with the version specified. You can also specify your own image OCID. However, note that these 2 are mutually exclusive i.e. either use Operating System and version *_or_* specify the OCID of your custom image.
 
-* the {uri-oci-shape}[shape] of the worker node determines the compute capacity of the worker node. By default, this is VM.Standard2.1, giving you 1 OCPU, 15GB Memory, 1 Gbps in network bandwidth and 2 VNICs.
+* the {uri-oci-shape}[shape] of the worker node determines the compute capacity of the worker node. This is controlled by the first element in the tuple for the node pool. By default, this is VM.Standard2.1, giving you 1 OCPU, 15GB Memory, 1 Gbps in network bandwidth and 2 VNICs e.g.
+
+----
+node_pools = {
+  "np1" = ["VM.Standard2.1", 1]
+  "np2" = ["BM.Standard2.52", 1]
+}
+----
+
+In the above example, workers in node pool np1 will all have a shape of VM.Standard2.1 whereas workers in node pool np2 will all have a shape of BM.Standard.2.52.
 
 * the subnets the node pool will span i.e. the subnets where the worker nodes will be created. See below for more explanation.
 
-* the number of worker nodes per subnet that will be created for this node pool. This is controlled by the node_pool_quantity_per_subnet parameter.
+* the number of worker nodes per subnet that will be created for this node pool. This is controlled by the 2nd element in the tuple for each node pool e.g.
+
+----
+node_pools = {
+  "np1" = ["VM.Standard2.1", 1]
+  "np2" = ["VM.Standard2.2", 3]
+}
+----
+
+will create a node pool (np1) with 1 worker node per subnet and a 2nd node pool (np2) with 3 worker nodes per subnet. 
 
 * the public ssh key used is the same as that used for the bastion host.
 
@@ -226,31 +258,68 @@ When using Topology 3, this ensures that the node pool spans all 3 worker subnet
 
 ==== Number of Node Pools
 
-The number of node pools created is controlled by the node_pools parameter. The diagram below shows a cluster with 1 node pool and 1 worker node per subnet using topology 3 i.e. node_pools=1, node_pool_quantity_per_subnet=1 and nodepool_topology=3.
+The number, shape and size of the node pools created is controlled by the number of entries in the node_pools parameter. The diagram below shows a cluster with 1 node pool and 1 worker node per subnet using topology 3 i.e. 
+
+----
+node_pools = {
+  "np1" = ["VM.Standard2.1", 1]
+}
+nodepool_topology=3
+----
+
 
 .1 Node Pool with 1 worker node per subnet (other details removed for convenience)
 image::images/np311.png[align="center"]
 
 {bl}
 
-You can increase the number of node pools by setting node_pools=5, node_pool_quantity_per_subnet=1 and nodepool_topology=3.
+You can increase the number of node pools by adding more entries in the node_pools e.g. 
+
+----
+node_pools = {
+  "np1" = ["VM.Standard2.1", 1]
+  "np2" = ["VM.Standard2.1", 1]
+  "np3" = ["VM.Standard2.1", 1]
+  "np4" = ["VM.Standard2.1", 1]
+  "np5" = ["VM.Standard2.1", 1]
+}
+nodepool_topology=3
+----
 
 .5 Node Pools with 1 worker node per subnet
 image::images/np351.png[align="center"]
 
 ==== Worker Nodes per subnet
 
-You can also change the number of worker nodes per subnet. For example, setting the node_pools=1 and node_pool_quantity_per_subnet=2 and nodepool_topology=3 will result in the following cluster:
+You can also change the number of worker nodes per subnet e.g.
+
+----
+node_pools = {
+  "np1" = ["VM.Standard2.1", 2]
+}
+nodepool_topology=3
+----
+
+will result in the following cluster:
 
 .1 Node Pool with 2 worker nodes per subnet
 image::images/np312.png[align="center"]
 
 {bl}
 
-Similarly, you can change both node pools and number of worker nodes per subnet:
+Similarly, you can support mixed workloads by adding node pools of different shapes and number of worker nodes per subnet:
+
+----
+node_pools = {
+  "np1" = ["VM.Standard.E2.1", 3]
+  "np2" = ["VM.Standard2.24", 2]
+  "np3" = ["BM.Standard1.36", 1]
+}
+nodepool_topology=3
+----
 
 .4 Node Pools with 2 worker nodes per subnet
-image::images/np342.png[align="center"]
+image::images/mixedworkload.png[align="center"]
 
 === Fault Domains
 

docs/images/np342.png

@@ -63,7 +63,7 @@ service_gateway_name = "sg"
 
 # bastion
 
-bastion_shape = "VM.Standard2.1"
+bastion_shape = "VM.Standard.E2.1"
 
 create_bastion = true
 
@@ -101,8 +101,7 @@ dashboard_enabled = true
 kubernetes_version = "LATEST"
 
 node_pools = {
-  "np1" = ["VM.Standard2.1", 4]
-  "np2" = ["VM.Standard2.2", 2]
+  "np1" = ["VM.Standard2.1", 1]
 }
 
 node_pool_name_prefix = "np"

docs/images/np351.png

@@ -127,7 +127,7 @@ variable "subnets" {
 # bastion
 variable "bastion_shape" {
   description = "shape of bastion instance"
-  default     = "VM.Standard2.1"
+  default     = "VM.Standard.E2.1"
 }
 
 variable "create_bastion" {