You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

282 lines
12 KiB

Added file and container image caching (#4828) * File and container image downloads are now cached localy, so that repeated vagrant up/down runs do not trigger downloading of those files. This is especially useful on laptops with kubernetes runnig locally on vm's. The total size of the cache, after an ansible run, is currently around 800MB, so bandwidth (=time) savings can be quite significant. * When download_run_once is false, the default is still not to cache, but setting download_force_cache will still enable caching. * The local cache location can be set with download_cache_dir and defaults to /tmp/kubernetes_cache * A local docker instance is no longer required to cache docker images; Images are cached to file. A local docker instance is still required, though, if you wish to download images on localhost. * Fixed a FIXME, wher the argument was that delegate_to doesn't play nice with omit. That is a correct observation and the fix is to use default(inventory_host) instead of default(omit). See ansible/ansible#26009 * Removed "Register docker images info" task from download_container and set_docker_image_facts because it was faulty and unused. * Removed redundant when:download.{container,enabled,run_once} conditions from {sync,download}_container.yml * All features of commit d6fd0d2acaec9f53e75d82db30411f96a5bf2cc9 by Timoses <timosesu@gmail.com>, merged May 1st 2019, are included in this patch. Not all code was included verbatim, but each feature of that commit was checked to be working in this patch. One notable change: The actual downloading of the kubeadm images was moved to {download,sync)_container, to enable caching. Note 1: I considered splitting this patch, but most changes that are not directly related to caching, are a pleasant by-product of implementing the caching code, so splitting would be impractical. Note 2: I have my doubts about the usefulness of the upload, download and upgrade tags in the download role. Must they remain or can they be removed? If anybody knows, then please speak up.
5 years ago
Added file and container image caching (#4828) * File and container image downloads are now cached localy, so that repeated vagrant up/down runs do not trigger downloading of those files. This is especially useful on laptops with kubernetes runnig locally on vm's. The total size of the cache, after an ansible run, is currently around 800MB, so bandwidth (=time) savings can be quite significant. * When download_run_once is false, the default is still not to cache, but setting download_force_cache will still enable caching. * The local cache location can be set with download_cache_dir and defaults to /tmp/kubernetes_cache * A local docker instance is no longer required to cache docker images; Images are cached to file. A local docker instance is still required, though, if you wish to download images on localhost. * Fixed a FIXME, wher the argument was that delegate_to doesn't play nice with omit. That is a correct observation and the fix is to use default(inventory_host) instead of default(omit). See ansible/ansible#26009 * Removed "Register docker images info" task from download_container and set_docker_image_facts because it was faulty and unused. * Removed redundant when:download.{container,enabled,run_once} conditions from {sync,download}_container.yml * All features of commit d6fd0d2acaec9f53e75d82db30411f96a5bf2cc9 by Timoses <timosesu@gmail.com>, merged May 1st 2019, are included in this patch. Not all code was included verbatim, but each feature of that commit was checked to be working in this patch. One notable change: The actual downloading of the kubeadm images was moved to {download,sync)_container, to enable caching. Note 1: I considered splitting this patch, but most changes that are not directly related to caching, are a pleasant by-product of implementing the caching code, so splitting would be impractical. Note 2: I have my doubts about the usefulness of the upload, download and upgrade tags in the download role. Must they remain or can they be removed? If anybody knows, then please speak up.
5 years ago
Added file and container image caching (#4828) * File and container image downloads are now cached localy, so that repeated vagrant up/down runs do not trigger downloading of those files. This is especially useful on laptops with kubernetes runnig locally on vm's. The total size of the cache, after an ansible run, is currently around 800MB, so bandwidth (=time) savings can be quite significant. * When download_run_once is false, the default is still not to cache, but setting download_force_cache will still enable caching. * The local cache location can be set with download_cache_dir and defaults to /tmp/kubernetes_cache * A local docker instance is no longer required to cache docker images; Images are cached to file. A local docker instance is still required, though, if you wish to download images on localhost. * Fixed a FIXME, wher the argument was that delegate_to doesn't play nice with omit. That is a correct observation and the fix is to use default(inventory_host) instead of default(omit). See ansible/ansible#26009 * Removed "Register docker images info" task from download_container and set_docker_image_facts because it was faulty and unused. * Removed redundant when:download.{container,enabled,run_once} conditions from {sync,download}_container.yml * All features of commit d6fd0d2acaec9f53e75d82db30411f96a5bf2cc9 by Timoses <timosesu@gmail.com>, merged May 1st 2019, are included in this patch. Not all code was included verbatim, but each feature of that commit was checked to be working in this patch. One notable change: The actual downloading of the kubeadm images was moved to {download,sync)_container, to enable caching. Note 1: I considered splitting this patch, but most changes that are not directly related to caching, are a pleasant by-product of implementing the caching code, so splitting would be impractical. Note 2: I have my doubts about the usefulness of the upload, download and upgrade tags in the download role. Must they remain or can they be removed? If anybody knows, then please speak up.
5 years ago
  1. # -*- mode: ruby -*-
  2. # # vi: set ft=ruby :
  3. # For help on using kubespray with vagrant, check out docs/vagrant.md
  4. require 'fileutils'
  5. Vagrant.require_version ">= 2.0.0"
  6. CONFIG = File.join(File.dirname(__FILE__), ENV['KUBESPRAY_VAGRANT_CONFIG'] || 'vagrant/config.rb')
  7. FLATCAR_URL_TEMPLATE = "https://%s.release.flatcar-linux.net/amd64-usr/current/flatcar_production_vagrant.json"
  8. FEDORA35_MIRROR = "https://download.fedoraproject.org/pub/fedora/linux/releases/35/Cloud/x86_64/images/Fedora-Cloud-Base-Vagrant-35-1.2.x86_64.vagrant-libvirt.box"
  9. # Uniq disk UUID for libvirt
  10. DISK_UUID = Time.now.utc.to_i
  11. SUPPORTED_OS = {
  12. "flatcar-stable" => {box: "flatcar-stable", user: "core", box_url: FLATCAR_URL_TEMPLATE % ["stable"]},
  13. "flatcar-beta" => {box: "flatcar-beta", user: "core", box_url: FLATCAR_URL_TEMPLATE % ["beta"]},
  14. "flatcar-alpha" => {box: "flatcar-alpha", user: "core", box_url: FLATCAR_URL_TEMPLATE % ["alpha"]},
  15. "flatcar-edge" => {box: "flatcar-edge", user: "core", box_url: FLATCAR_URL_TEMPLATE % ["edge"]},
  16. "ubuntu2004" => {box: "generic/ubuntu2004", user: "vagrant"},
  17. "ubuntu2204" => {box: "generic/ubuntu2204", user: "vagrant"},
  18. "centos" => {box: "centos/7", user: "vagrant"},
  19. "centos-bento" => {box: "bento/centos-7.6", user: "vagrant"},
  20. "centos8" => {box: "centos/8", user: "vagrant"},
  21. "centos8-bento" => {box: "bento/centos-8", user: "vagrant"},
  22. "almalinux8" => {box: "almalinux/8", user: "vagrant"},
  23. "almalinux8-bento" => {box: "bento/almalinux-8", user: "vagrant"},
  24. "rockylinux8" => {box: "generic/rocky8", user: "vagrant"},
  25. "fedora35" => {box: "fedora/35-cloud-base", user: "vagrant", box_url: FEDORA35_MIRROR},
  26. "fedora36" => {box: "fedora/36-cloud-base", user: "vagrant"},
  27. "opensuse" => {box: "opensuse/Leap-15.4.x86_64", user: "vagrant"},
  28. "opensuse-tumbleweed" => {box: "opensuse/Tumbleweed.x86_64", user: "vagrant"},
  29. "oraclelinux" => {box: "generic/oracle7", user: "vagrant"},
  30. "oraclelinux8" => {box: "generic/oracle8", user: "vagrant"},
  31. "rhel7" => {box: "generic/rhel7", user: "vagrant"},
  32. "rhel8" => {box: "generic/rhel8", user: "vagrant"},
  33. }
  34. if File.exist?(CONFIG)
  35. require CONFIG
  36. end
  37. # Defaults for config options defined in CONFIG
  38. $num_instances ||= 3
  39. $instance_name_prefix ||= "k8s"
  40. $vm_gui ||= false
  41. $vm_memory ||= 2048
  42. $vm_cpus ||= 2
  43. $shared_folders ||= {}
  44. $forwarded_ports ||= {}
  45. $subnet ||= "172.18.8"
  46. $subnet_ipv6 ||= "fd3c:b398:0698:0756"
  47. $os ||= "ubuntu2004"
  48. $network_plugin ||= "flannel"
  49. # Setting multi_networking to true will install Multus: https://github.com/k8snetworkplumbingwg/multus-cni
  50. $multi_networking ||= "False"
  51. $download_run_once ||= "True"
  52. $download_force_cache ||= "False"
  53. # The first three nodes are etcd servers
  54. $etcd_instances ||= [$num_instances, 3].min
  55. # The first two nodes are kube masters
  56. $kube_master_instances ||= [$num_instances, 2].min
  57. # All nodes are kube nodes
  58. $kube_node_instances ||= $num_instances
  59. # The following only works when using the libvirt provider
  60. $kube_node_instances_with_disks ||= false
  61. $kube_node_instances_with_disks_size ||= "20G"
  62. $kube_node_instances_with_disks_number ||= 2
  63. $override_disk_size ||= false
  64. $disk_size ||= "20GB"
  65. $local_path_provisioner_enabled ||= "False"
  66. $local_path_provisioner_claim_root ||= "/opt/local-path-provisioner/"
  67. $libvirt_nested ||= false
  68. # boolean or string (e.g. "-vvv")
  69. $ansible_verbosity ||= false
  70. $ansible_tags ||= ENV['VAGRANT_ANSIBLE_TAGS'] || ""
  71. $playbook ||= "cluster.yml"
  72. host_vars = {}
  73. # throw error if os is not supported
  74. if ! SUPPORTED_OS.key?($os)
  75. puts "Unsupported OS: #{$os}"
  76. puts "Supported OS are: #{SUPPORTED_OS.keys.join(', ')}"
  77. exit 1
  78. end
  79. $box = SUPPORTED_OS[$os][:box]
  80. # if $inventory is not set, try to use example
  81. $inventory = "inventory/sample" if ! $inventory
  82. $inventory = File.absolute_path($inventory, File.dirname(__FILE__))
  83. # if $inventory has a hosts.ini file use it, otherwise copy over
  84. # vars etc to where vagrant expects dynamic inventory to be
  85. if ! File.exist?(File.join(File.dirname($inventory), "hosts.ini"))
  86. $vagrant_ansible = File.join(File.dirname(__FILE__), ".vagrant", "provisioners", "ansible")
  87. FileUtils.mkdir_p($vagrant_ansible) if ! File.exist?($vagrant_ansible)
  88. $vagrant_inventory = File.join($vagrant_ansible,"inventory")
  89. FileUtils.rm_f($vagrant_inventory)
  90. FileUtils.ln_s($inventory, $vagrant_inventory)
  91. end
  92. if Vagrant.has_plugin?("vagrant-proxyconf")
  93. $no_proxy = ENV['NO_PROXY'] || ENV['no_proxy'] || "127.0.0.1,localhost"
  94. (1..$num_instances).each do |i|
  95. $no_proxy += ",#{$subnet}.#{i+100}"
  96. end
  97. end
  98. Vagrant.configure("2") do |config|
  99. config.vm.box = $box
  100. if SUPPORTED_OS[$os].has_key? :box_url
  101. config.vm.box_url = SUPPORTED_OS[$os][:box_url]
  102. end
  103. config.ssh.username = SUPPORTED_OS[$os][:user]
  104. # plugin conflict
  105. if Vagrant.has_plugin?("vagrant-vbguest") then
  106. config.vbguest.auto_update = false
  107. end
  108. # always use Vagrants insecure key
  109. config.ssh.insert_key = false
  110. if ($override_disk_size)
  111. unless Vagrant.has_plugin?("vagrant-disksize")
  112. system "vagrant plugin install vagrant-disksize"
  113. end
  114. config.disksize.size = $disk_size
  115. end
  116. (1..$num_instances).each do |i|
  117. config.vm.define vm_name = "%s-%01d" % [$instance_name_prefix, i] do |node|
  118. node.vm.hostname = vm_name
  119. if Vagrant.has_plugin?("vagrant-proxyconf")
  120. node.proxy.http = ENV['HTTP_PROXY'] || ENV['http_proxy'] || ""
  121. node.proxy.https = ENV['HTTPS_PROXY'] || ENV['https_proxy'] || ""
  122. node.proxy.no_proxy = $no_proxy
  123. end
  124. ["vmware_fusion", "vmware_workstation"].each do |vmware|
  125. node.vm.provider vmware do |v|
  126. v.vmx['memsize'] = $vm_memory
  127. v.vmx['numvcpus'] = $vm_cpus
  128. end
  129. end
  130. node.vm.provider :virtualbox do |vb|
  131. vb.memory = $vm_memory
  132. vb.cpus = $vm_cpus
  133. vb.gui = $vm_gui
  134. vb.linked_clone = true
  135. vb.customize ["modifyvm", :id, "--vram", "8"] # ubuntu defaults to 256 MB which is a waste of precious RAM
  136. vb.customize ["modifyvm", :id, "--audio", "none"]
  137. end
  138. node.vm.provider :libvirt do |lv|
  139. lv.nested = $libvirt_nested
  140. lv.cpu_mode = "host-model"
  141. lv.memory = $vm_memory
  142. lv.cpus = $vm_cpus
  143. lv.default_prefix = 'kubespray'
  144. # Fix kernel panic on fedora 28
  145. if $os == "fedora"
  146. lv.cpu_mode = "host-passthrough"
  147. end
  148. end
  149. if $kube_node_instances_with_disks
  150. # Libvirt
  151. driverletters = ('a'..'z').to_a
  152. node.vm.provider :libvirt do |lv|
  153. # always make /dev/sd{a/b/c} so that CI can ensure that
  154. # virtualbox and libvirt will have the same devices to use for OSDs
  155. (1..$kube_node_instances_with_disks_number).each do |d|
  156. lv.storage :file, :device => "hd#{driverletters[d]}", :path => "disk-#{i}-#{d}-#{DISK_UUID}.disk", :size => $kube_node_instances_with_disks_size, :bus => "scsi"
  157. end
  158. end
  159. end
  160. if $expose_docker_tcp
  161. node.vm.network "forwarded_port", guest: 2375, host: ($expose_docker_tcp + i - 1), auto_correct: true
  162. end
  163. $forwarded_ports.each do |guest, host|
  164. node.vm.network "forwarded_port", guest: guest, host: host, auto_correct: true
  165. end
  166. if ["rhel7","rhel8"].include? $os
  167. # Vagrant synced_folder rsync options cannot be used for RHEL boxes as Rsync package cannot
  168. # be installed until the host is registered with a valid Red Hat support subscription
  169. node.vm.synced_folder ".", "/vagrant", disabled: false
  170. $shared_folders.each do |src, dst|
  171. node.vm.synced_folder src, dst
  172. end
  173. else
  174. node.vm.synced_folder ".", "/vagrant", disabled: false, type: "rsync", rsync__args: ['--verbose', '--archive', '--delete', '-z'] , rsync__exclude: ['.git','venv']
  175. $shared_folders.each do |src, dst|
  176. node.vm.synced_folder src, dst, type: "rsync", rsync__args: ['--verbose', '--archive', '--delete', '-z']
  177. end
  178. end
  179. ip = "#{$subnet}.#{i+100}"
  180. node.vm.network :private_network, ip: ip,
  181. :libvirt__guest_ipv6 => 'yes',
  182. :libvirt__ipv6_address => "#{$subnet_ipv6}::#{i+100}",
  183. :libvirt__ipv6_prefix => "64",
  184. :libvirt__forward_mode => "none",
  185. :libvirt__dhcp_enabled => false
  186. # Disable swap for each vm
  187. node.vm.provision "shell", inline: "swapoff -a"
  188. # ubuntu2004 and ubuntu2204 have IPv6 explicitly disabled. This undoes that.
  189. if ["ubuntu2004", "ubuntu2204"].include? $os
  190. node.vm.provision "shell", inline: "rm -f /etc/modprobe.d/local.conf"
  191. node.vm.provision "shell", inline: "sed -i '/net.ipv6.conf.all.disable_ipv6/d' /etc/sysctl.d/99-sysctl.conf /etc/sysctl.conf"
  192. end
  193. # Disable firewalld on oraclelinux/redhat vms
  194. if ["oraclelinux","oraclelinux8","rhel7","rhel8"].include? $os
  195. node.vm.provision "shell", inline: "systemctl stop firewalld; systemctl disable firewalld"
  196. end
  197. host_vars[vm_name] = {
  198. "ip": ip,
  199. "flannel_interface": "eth1",
  200. "kube_network_plugin": $network_plugin,
  201. "kube_network_plugin_multus": $multi_networking,
  202. "download_run_once": $download_run_once,
  203. "download_localhost": "False",
  204. "download_cache_dir": ENV['HOME'] + "/kubespray_cache",
  205. # Make kubespray cache even when download_run_once is false
  206. "download_force_cache": $download_force_cache,
  207. # Keeping the cache on the nodes can improve provisioning speed while debugging kubespray
  208. "download_keep_remote_cache": "False",
  209. "docker_rpm_keepcache": "1",
  210. # These two settings will put kubectl and admin.config in $inventory/artifacts
  211. "kubeconfig_localhost": "True",
  212. "kubectl_localhost": "True",
  213. "local_path_provisioner_enabled": "#{$local_path_provisioner_enabled}",
  214. "local_path_provisioner_claim_root": "#{$local_path_provisioner_claim_root}",
  215. "ansible_ssh_user": SUPPORTED_OS[$os][:user]
  216. }
  217. # Only execute the Ansible provisioner once, when all the machines are up and ready.
  218. # And limit the action to gathering facts, the full playbook is going to be ran by testcases_run.sh
  219. if i == $num_instances
  220. node.vm.provision "ansible" do |ansible|
  221. ansible.playbook = $playbook
  222. ansible.verbose = $ansible_verbosity
  223. $ansible_inventory_path = File.join( $inventory, "hosts.ini")
  224. if File.exist?($ansible_inventory_path)
  225. ansible.inventory_path = $ansible_inventory_path
  226. end
  227. ansible.become = true
  228. ansible.limit = "all,localhost"
  229. ansible.host_key_checking = false
  230. ansible.raw_arguments = ["--forks=#{$num_instances}", "--flush-cache", "-e ansible_become_pass=vagrant"]
  231. ansible.host_vars = host_vars
  232. if $ansible_tags != ""
  233. ansible.tags = [$ansible_tags]
  234. end
  235. ansible.groups = {
  236. "etcd" => ["#{$instance_name_prefix}-[1:#{$etcd_instances}]"],
  237. "kube_control_plane" => ["#{$instance_name_prefix}-[1:#{$kube_master_instances}]"],
  238. "kube_node" => ["#{$instance_name_prefix}-[1:#{$kube_node_instances}]"],
  239. "k8s_cluster:children" => ["kube_control_plane", "kube_node"],
  240. }
  241. end
  242. end
  243. end
  244. end
  245. end