# Copyright:: Copyright (c) 2019 eGlobalTech, Inc., all rights reserved
#
# Licensed under the BSD-3 license (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License in the root of the project or at
#
# http://egt-labs.com/mu/LICENSE.html
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
module MU
class Cloud
class Azure
# A Kubernetes cluster as configured in {MU::Config::BasketofKittens::container_clusters}
class ContainerCluster < MU::Cloud::ContainerCluster
# Initialize this cloud resource object. Calling +super+ will invoke the initializer defined under {MU::Cloud}, which should set the attribtues listed in {MU::Cloud::PUBLIC_ATTRS} as well as applicable dependency shortcuts, like @vpc, for us.
# @param args [Hash]: Hash of named arguments passed via Ruby's double-splat
def initialize(**args)
super
# @mu_name = mu_name ? mu_name : @deploy.getResourceName(@config["name"])
if !mu_name.nil?
@mu_name = mu_name
@cloud_id = Id.new(cloud_desc.id) if @cloud_id
else
@mu_name ||= @deploy.getResourceName(@config["name"], max_length: 31)
end
end
# Called automatically by {MU::Deploy#createResources}
# @return [String]: The cloud provider's identifier for this GKE instance.
def create
create_update
end
# Called automatically by {MU::Deploy#createResources}
def groom
create_update
kube_conf = @deploy.deploy_dir+"/kubeconfig-#{@config['name']}"
admin_creds = MU::Cloud::Azure.containers(credentials: @config['credentials']).managed_clusters.list_cluster_admin_credentials(
@resource_group,
@mu_name
)
admin_creds.kubeconfigs.each { |kube|
next if kube.name != "clusterAdmin"
cfgfile = ""
kube.value.each { |ord|
cfgfile += ord.chr
}
File.open(kube_conf, "w"){ |k|
k.puts cfgfile
}
}
if @config['kubernetes_resources']
MU::Master.applyKubernetesResources(
@config['name'],
@config['kubernetes_resources'],
kubeconfig: kube_conf,
outputdir: @deploy.deploy_dir
)
end
MU.log %Q{How to interact with your AKS cluster\nkubectl --kubeconfig "#{kube_conf}" get events --all-namespaces\nkubectl --kubeconfig "#{kube_conf}" get all\nkubectl --kubeconfig "#{kube_conf}" create -f some_k8s_deploy.yml\nkubectl --kubeconfig "#{kube_conf}" get nodes}, MU::SUMMARY
end
# Locate and return cloud provider descriptors of this resource type
# which match the provided parameters, or all visible resources if no
# filters are specified. At minimum, implementations of +find+ must
# honor +credentials+ and +cloud_id+ arguments. We may optionally
# support other search methods, such as +tag_key+ and +tag_value+, or
# cloud-specific arguments like +project+. See also {MU::MommaCat.findStray}.
# @param args [Hash]: Hash of named arguments passed via Ruby's double-splat
# @return [Hash]: The cloud provider's complete descriptions of matching resources
def self.find(**args)
found = {}
# Azure resources are namedspaced by resource group. If we weren't
# told one, we may have to search all the ones we can see.
resource_groups = if args[:resource_group]
[args[:resource_group]]
elsif args[:cloud_id] and args[:cloud_id].is_a?(MU::Cloud::Azure::Id)
[args[:cloud_id].resource_group]
else
MU::Cloud::Azure.resources(credentials: args[:credentials]).resource_groups.list.map { |rg| rg.name }
end
if args[:cloud_id]
id_str = args[:cloud_id].is_a?(MU::Cloud::Azure::Id) ? args[:cloud_id].name : args[:cloud_id]
resource_groups.each { |rg|
resp = MU::Cloud::Azure.containers(credentials: args[:credentials]).managed_clusters.get(rg, id_str)
found[Id.new(resp.id)] = resp if resp
}
else
if args[:resource_group]
MU::Cloud::Azure.containers(credentials: args[:credentials]).managed_clusters.list_by_resource_group(args[:resource_group]).each { |cluster|
found[Id.new(cluster.id)] = cluster
}
else
MU::Cloud::Azure.containers(credentials: args[:credentials]).managed_clusters.list.each { |cluster|
found[Id.new(cluster.id)] = cluster
}
end
end
found
end
# Register a description of this cluster instance with this deployment's metadata.
def notify
base = {}
base = MU.structToHash(cloud_desc)
base["cloud_id"] = @cloud_id.name
base.merge!(@config.to_h)
base
end
# Does this resource type exist as a global (cloud-wide) artifact, or
# is it localized to a region/zone?
# @return [Boolean]
def self.isGlobal?
false
end
# Denote whether this resource implementation is experiment, ready for
# testing, or ready for production use.
def self.quality
MU::Cloud::BETA
end
# Stub method. Azure resources are cleaned up by removing the parent
# resource group.
# @return [void]
def self.cleanup(**args)
end
# Cloud-specific configuration properties.
# @param config [MU::Config]: The calling MU::Config object
# @return [Array]: List of required fields, and json-schema Hash of cloud-specific configuration parameters for this resource
def self.schema(config)
toplevel_required = []
schema = {
"flavor" => {
"enum" => ["Kubernetes", "OpenShift", "Swarm", "DC/OS"],
"description" => "The Azure container platform to deploy. Currently only +Kubernetes+ is supported.",
"default" => "Kubernetes"
},
"platform" => {
"description" => "The OS platform to deploy for workers and containers.",
"default" => "Linux",
"enum" => ["Linux", "Windows"]
},
"max_pods" => {
"type" => "integer",
"description" => "Maximum number of pods allowed on this cluster",
"default" => 30
},
"kubernetes" => {
"default" => { "version" => "1.12.8" }
},
"dns_prefix" => {
"type" => "string",
"description" => "DNS name prefix to use with the hosted Kubernetes API server FQDN. Will default to the global +appname+ value if not specified."
},
"disk_size_gb" => {
"type" => "integer",
"description" => "Size of the disk attached to each worker, specified in GB. The smallest allowed disk size is 30, the largest 1024.",
"default" => 100
},
}
[toplevel_required, schema]
end
# Cloud-specific pre-processing of {MU::Config::BasketofKittens::container_clusters}, bare and unvalidated.
# @param cluster [Hash]: The resource to process and validate
# @param configurator [MU::Config]: The overall deployment configurator of which this resource is a member
# @return [Boolean]: True if validation succeeded, False otherwise
def self.validateConfig(cluster, configurator)
ok = true
# XXX validate k8s versions (master and node)
# XXX validate image types
# MU::Cloud::Azure.container.get_project_zone_serverconfig(@config["project"], @config['availability_zone'])
cluster["dns_prefix"] ||= $myAppName # XXX woof globals wtf
cluster['region'] ||= MU::Cloud::Azure.myRegion(cluster['credentials'])
if cluster["disk_size_gb"] < 30 or cluster["disk_size_gb"] > 1024
MU.log "Azure ContainerCluster disk_size_gb must be between 30 and 1024.", MU::ERR
ok = false
end
if cluster['min_size'] and cluster['instance_count'] < cluster['min_size']
cluster['instance_count'] = cluster['min_size']
end
if cluster['max_size'] and cluster['instance_count'] < cluster['max_size']
cluster['instance_count'] = cluster['max_size']
end
cluster['instance_type'] ||= "Standard_DS2_v2" # TODO when Server is implemented, it should have a validateInstanceType method we can use here
svcacct_desc = {
"name" => cluster["name"]+"user",
"region" => cluster["region"],
"type" => "service",
"cloud" => "Azure",
"create_api_key" => true,
"credentials" => cluster["credentials"],
"roles" => [
"Azure Kubernetes Service Cluster Admin Role"
]
}
cluster['dependencies'] ||= []
cluster['dependencies'] << {
"type" => "user",
"name" => cluster["name"]+"user"
}
ok = false if !configurator.insertKitten(svcacct_desc, "users")
ok
end
private
def create_update
need_apply = false
ext_cluster = MU::Cloud::Azure.containers(credentials: @config[:credentials]).managed_clusters.get(
@resource_group,
@mu_name
)
if ext_cluster
@cloud_id = MU::Cloud::Azure::Id.new(ext_cluster.id)
end
key_obj = MU::Cloud::Azure.containers(:ContainerServiceSshPublicKey).new
key_obj.key_data = @deploy.ssh_public_key
ssh_obj = MU::Cloud::Azure.containers(:ContainerServiceSshConfiguration).new
ssh_obj.public_keys = [key_obj]
os_profile_obj = if !ext_cluster
if @config['platform'] == "Windows"
os_obj = MU::Cloud::Azure.containers(:ContainerServiceWindowsProfile, model_version: "V2019_02_01").new
os_obj.admin_username = "muadmin"
# Azure password constraints are extra-annoying
winpass = MU.generateWindowsPassword(safe_pattern: '!@#$%^&*()', retries: 150)
# TODO store this somewhere the user can get at it
os_obj.admin_password = winpass
os_obj
else
os_obj = MU::Cloud::Azure.containers(:ContainerServiceLinuxProfile).new
os_obj.admin_username = "muadmin"
os_obj.ssh = ssh_obj
os_obj
end
else
# Azure does not support updates to this parameter
@config['platform'] == "Windows" ? ext_cluster.windows_profile : ext_cluster.linux_profile
end
svc_principal_obj = MU::Cloud::Azure.containers(:ManagedClusterServicePrincipalProfile).new
# XXX this should come from a MU::Cloud::Azure::User object, but right now
# there's no way to get the 'secret' field from a user-assigned identity afaict
# For now, we'll cheat with Mu's system credentials.
creds = MU::Cloud::Azure.credConfig(@config['credentials'])
svc_principal_obj.client_id = creds["client_id"]
svc_principal_obj.secret = creds["client_secret"]
# svc_acct = @deploy.findLitterMate(type: "user", name: @config['name']+"user")
# raise MuError, "Failed to locate service account #{@config['name']}user" if !svc_acct
# svc_principal_obj.client_id = svc_acct.cloud_desc.client_id
# svc_principal_obj.secret = svc_acct.getSecret
agent_profiles = if !ext_cluster
profile_obj = MU::Cloud::Azure.containers(:ManagedClusterAgentPoolProfile).new
profile_obj.name = @deploy.getResourceName(@config["name"], max_length: 11).downcase.gsub(/[^0-9a-z]/, "")
if @config['min_size'] and @config['max_size']
# Special API features need to be enabled for scaling
MU::Cloud::Azure.ensureFeature("Microsoft.ContainerService/WindowsPreview", credentials: @config['credentials'])
MU::Cloud::Azure.ensureFeature("Microsoft.ContainerService/VMSSPreview", credentials: @config['credentials'])
profile_obj.min_count = @config['min_size']
profile_obj.max_count = @config['max_size']
profile_obj.enable_auto_scaling = true
profile_obj.type = MU::Cloud::Azure.containers(:AgentPoolType)::VirtualMachineScaleSets
# XXX if you actually try to do this:
# BadRequest: Virtual Machine Scale Set agent nodes are not allowed since feature "Microsoft.ContainerService/WindowsPreview" is not enabled.
end
profile_obj.count = @config['instance_count']
profile_obj.vm_size = @config['instance_type']
profile_obj.max_pods = @config['max_pods']
profile_obj.os_type = @config['platform']
profile_obj.os_disk_size_gb = @config['disk_size_gb']
# XXX correlate this with the one(s) we configured in @config['vpc']
# profile_obj.vnet_subnet_id = @vpc.subnets.first.cloud_desc.id # XXX has to have its own subnet for k8s apparently
[profile_obj]
else
# Azure does not support adding/removing agent profiles to a live
# cluster, but it does support changing some values on an existing
# one.
profile_obj = ext_cluster.agent_pool_profiles.first
nochange_map = {
"disk_size_gb" => :os_disk_size_gb,
"instance_type" => :vm_size,
"platform" => :os_type,
"max_pods" => :max_pods,
}
tried_to_change =[]
nochange_map.each_pair { |cfg, attribute|
if @config.has_key?(cfg) and
@config[cfg] != profile_obj.send(attribute)
tried_to_change << cfg
end
}
if @config['min_size'] and @config['max_size'] and
!profile_obj.enable_auto_scaling
tried_to_change << "enable_auto_scaling"
end
if tried_to_change.size > 0
MU.log "Changes specified to one or more immutable AKS Agent Pool parameters in cluster #{@mu_name}, ignoring.", MU::NOTICE, details: tried_to_change
end
if @config['min_size'] and @config['max_size'] and
profile_obj.enable_auto_scaling and
(
profile_obj.min_count != @config['min_size'] or
profile_obj.max_count != @config['max_size']
)
profile_obj.min_count = @config['min_size']
profile_obj.max_count = @config['max_size']
need_apply = true
end
if profile_obj.count != @config['instance_count']
profile_obj.count = @config['instance_count']
need_apply = true
end
[profile_obj]
end
cluster_obj = MU::Cloud::Azure.containers(:ManagedCluster).new
if ext_cluster
cluster_obj.dns_prefix = ext_cluster.dns_prefix
cluster_obj.location = ext_cluster.location
else
# Azure does not support updates to these parameters
cluster_obj.dns_prefix = @config['dns_prefix']
cluster_obj.location = @config['region']
end
cluster_obj.tags = @tags
cluster_obj.service_principal_profile = svc_principal_obj
if @config['platform'] == "Windows"
cluster_obj.windows_profile = os_profile_obj
else
cluster_obj.linux_profile = os_profile_obj
end
# cluster_obj.api_server_authorized_ipranges = [MU.mu_public_ip+"/32", MU.my_private_ip+"/32"] # XXX only allowed with Microsoft.ContainerService/APIServerSecurityPreview enabled
cluster_obj.agent_pool_profiles = agent_profiles
if @config['flavor'] == "Kubernetes"
cluster_obj.kubernetes_version = @config['kubernetes']['version'].to_s
if ext_cluster and @config['kubernetes']['version'] != ext_cluster.kubernetes_version
need_apply = true
end
end
# XXX it may be possible to create a new AgentPool and fall forward into it?
# API behavior suggests otherwise. Project for later.
# pool_obj = MU::Cloud::Azure.containers(:AgentPool).new
# pool_obj.count = @config['instance_count']
# pool_obj.vm_size = "Standard_DS2_v2"
if !ext_cluster
pp cluster_obj
MU.log "Creating AKS cluster #{@mu_name}", details: cluster_obj
need_apply = true
elsif need_apply
MU.log "Updating AKS cluster #{@mu_name}", MU::NOTICE, details: cluster_obj
end
if need_apply
resp = MU::Cloud::Azure.containers(credentials: @config['credentials']).managed_clusters.create_or_update(
@resource_group,
@mu_name,
cluster_obj
)
@cloud_id = Id.new(resp.id)
end
end
end #class
end #class
end
end #module