Class: Vmpooler::PoolManager::Provider::Gce
- Inherits:
-
Base
- Object
- Base
- Vmpooler::PoolManager::Provider::Gce
- Defined in:
- lib/vmpooler/providers/gce.rb
Overview
This class represent a GCE provider to CRUD resources in a gce cloud.
Instance Attribute Summary collapse
-
#connection_pool ⇒ Object
readonly
The connection_pool method is normally used only for testing.
Instance Method Summary collapse
- #connect_to_gce ⇒ Object
- #connection ⇒ Object
-
#create_disk(pool_name, vm_name, disk_size) ⇒ Object
create_disk creates an additional disk for an existing VM.
-
#create_snapshot(pool_name, vm_name, new_snapshot_name) ⇒ Object
create_snapshot creates new snapshots with the unique name new_snapshot_name-#diskdisk.name for one vm, and one create_snapshot() there could be multiple snapshots created, one for each drive.
-
#create_vm(pool_name, new_vmname) ⇒ Object
create_vm creates a new VM with a default network from the config, a initial disk named #new_vmname-disk0 that uses the ‘template’ as its source image and labels added for vm and pool and an instance configuration for machine_type from the config and labels vm and pool having a label “pool” that match a pool config name.
-
#debug_logger(message, send_to_upstream: false) ⇒ Object
used in local dev environment, set DEBUG_FLAG=true this way the upstream vmpooler manager does not get polluted with logs.
-
#destroy_vm(pool_name, vm_name) ⇒ Object
destroy_vm deletes an existing VM instance and any disks and snapshots via the labels in gce instances, disks and snapshots are resources that can exist independent of each other inputs [String] pool_name : Name of the pool [String] vm_name : Name of the existing VM returns [boolean] true : once the operations are finished.
- #disk_name_from_source(attached_disk) ⇒ Object
- #domain(pool_name) ⇒ Object
- #ensured_gce_connection(connection_pool_object) ⇒ Object
-
#find_all_snapshots(vm_name) ⇒ Object
find all snapshots ever created for one vm, regardless of snapshot name, for example when deleting it all.
-
#find_snapshot(vm_name, snapshotname) ⇒ Object
this is used because for one vm, with the same snapshot name there could be multiple snapshots, one for each disk.
-
#generate_vm_hash(vm_object, pool_name) ⇒ Object
Return a hash of VM data Provides vmname, hostname, template, poolname, boottime, status, zone, machine_type, labels, label_fingerprint, ip information.
- #get_current_user(vm_name) ⇒ Object
-
#get_vm(pool_name, vm_name) ⇒ Object
inputs [String] pool_name : Name of the pool [String] vm_name : Name of the VM to find returns nil if VM doesn’t exist [Hastable] of the VM [String] name : The name of the resource, provided by the client when initially creating the resource [String] hostname : Specifies the hostname of the instance.
- #get_vm_ip_address(vm_name, pool_name) ⇒ Object
-
#initialize(config, logger, metrics, redis_connection_pool, name, options) ⇒ Gce
constructor
A new instance of Gce.
- #machine_type(pool_name) ⇒ Object
-
#name ⇒ Object
name of the provider class.
- #network_name ⇒ Object
-
#open_socket(host, domain = nil, timeout = 5, port = 22, &_block) ⇒ Object
This should supercede the open_socket method in the Pool Manager.
-
#project ⇒ Object
main configuration options.
-
#purge_unconfigured_resources(allowlist) ⇒ Object
Scans zones that are configured for list of resources (VM, disks, snapshots) that do not have the label.pool set to one of the configured pools.
-
#revert_snapshot(pool_name, vm_name, snapshot_name) ⇒ Object
revert_snapshot reverts an existing VM’s disks to an existing snapshot_name reverting in gce entails 1.
-
#should_be_ignored(item, allowlist) ⇒ Object
END BASE METHODS.
- #subnetwork_name(pool_name) ⇒ Object
-
#tag_vm_user(pool, vm_name) ⇒ Object
tag_vm_user This method is called once we know who is using the VM (it is running).
- #vm_ready?(pool_name, vm_name, redis) ⇒ Boolean
-
#vms_in_pool(pool_name) ⇒ Object
vms_in_pool lists all the VM names in a pool, which is based on the VMs having a label “pool” that match a pool config name.
- #wait_for_operation(project, pool_name, result, retries = 5) ⇒ Object
-
#wait_for_zone_operation(project, zone, result, retries = 5) ⇒ Object
Compute resource wait for operation to be DONE (synchronous operation).
-
#zone(pool_name) ⇒ Object
main configuration options, overridable for each pool.
Constructor Details
#initialize(config, logger, metrics, redis_connection_pool, name, options) ⇒ Gce
Returns a new instance of Gce.
18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 |
# File 'lib/vmpooler/providers/gce.rb', line 18 def initialize(config, logger, metrics, redis_connection_pool, name, ) super(config, logger, metrics, redis_connection_pool, name, ) task_limit = global_config[:config].nil? || global_config[:config]['task_limit'].nil? ? 10 : global_config[:config]['task_limit'].to_i # The default connection pool size is: # Whatever is biggest from: # - How many pools this provider services # - Maximum number of cloning tasks allowed # - Need at least 2 connections so that a pool can have inventory functions performed while cloning etc. default_connpool_size = [provided_pools.count, task_limit, 2].max connpool_size = provider_config['connection_pool_size'].nil? ? default_connpool_size : provider_config['connection_pool_size'].to_i # The default connection pool timeout should be quite large - 60 seconds connpool_timeout = provider_config['connection_pool_timeout'].nil? ? 60 : provider_config['connection_pool_timeout'].to_i logger.log('d', "[#{name}] ConnPool - Creating a connection pool of size #{connpool_size} with timeout #{connpool_timeout}") @connection_pool = Vmpooler::PoolManager::GenericConnectionPool.new( metrics: metrics, connpool_type: 'provider_connection_pool', connpool_provider: name, size: connpool_size, timeout: connpool_timeout ) do logger.log('d', "[#{name}] Connection Pool - Creating a connection object") # Need to wrap the vSphere connection object in another object. The generic connection pooler will preserve # the object reference for the connection, which means it cannot "reconnect" by creating an entirely new connection # object. Instead by wrapping it in a Hash, the Hash object reference itself never changes but the content of the # Hash can change, and is preserved across invocations. new_conn = connect_to_gce { connection: new_conn } end @redis = redis_connection_pool end |
Instance Attribute Details
#connection_pool ⇒ Object (readonly)
The connection_pool method is normally used only for testing
16 17 18 |
# File 'lib/vmpooler/providers/gce.rb', line 16 def connection_pool @connection_pool end |
Instance Method Details
#connect_to_gce ⇒ Object
649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 |
# File 'lib/vmpooler/providers/gce.rb', line 649 def connect_to_gce max_tries = global_config[:config]['max_tries'] || 3 retry_factor = global_config[:config]['retry_factor'] || 10 try = 1 begin scopes = ['https://www.googleapis.com/auth/compute', 'https://www.googleapis.com/auth/cloud-platform'] = Google::Auth.get_application_default(scopes) compute = ::Google::Apis::ComputeV1::ComputeService.new compute. = metrics.increment('connect.open') compute rescue StandardError => e # is that even a thing? metrics.increment('connect.fail') raise e if try >= max_tries sleep(try * retry_factor) try += 1 retry end end |
#connection ⇒ Object
55 56 57 58 59 |
# File 'lib/vmpooler/providers/gce.rb', line 55 def connection @connection_pool.with_metrics do |pool_object| return ensured_gce_connection(pool_object) end end |
#create_disk(pool_name, vm_name, disk_size) ⇒ Object
create_disk creates an additional disk for an existing VM. It will name the new disk #vm_name-disk#number_disk where number_disk is the next logical disk number starting with 1 when adding an additional disk to a VM with only the boot disk: #vm_name-disk0 == boot disk #vm_name-disk1 == additional disk added via create_disk #vm_name-disk2 == additional disk added via create_disk if run a second time etc the new disk has labels added for vm and pool The GCE lifecycle is to create a new disk (lives independently of the instance) then to attach it to the existing instance. inputs
[String] pool_name : Name of the pool
[String] vm_name : Name of the existing VM
[String] disk_size : The new disk size in GB
returns
[boolean] true : once the operations are finished
231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 |
# File 'lib/vmpooler/providers/gce.rb', line 231 def create_disk(pool_name, vm_name, disk_size) debug_logger('create_disk') pool = pool_config(pool_name) raise("Pool #{pool_name} does not exist for the provider #{name}") if pool.nil? begin vm_object = connection.get_instance(project, zone(pool_name), vm_name) rescue ::Google::Apis::ClientError => e raise e unless e.status_code == 404 # if it does not exist raise("VM #{vm_name} in pool #{pool_name} does not exist for the provider #{name}") end # this number should start at 1 when there is only the boot disk, # eg the new disk will be named spicy-proton-disk1 number_disk = vm_object.disks.length disk_name = "#{vm_name}-disk#{number_disk}" disk = Google::Apis::ComputeV1::Disk.new( name: disk_name, size_gb: disk_size, labels: { 'pool' => pool_name, 'vm' => vm_name } ) debug_logger("trigger insert_disk #{disk_name} for #{vm_name}") result = connection.insert_disk(project, zone(pool_name), disk) wait_for_operation(project, pool_name, result) debug_logger("trigger get_disk #{disk_name} for #{vm_name}") new_disk = connection.get_disk(project, zone(pool_name), disk_name) attached_disk = Google::Apis::ComputeV1::AttachedDisk.new( auto_delete: true, boot: false, source: new_disk.self_link ) debug_logger("trigger attach_disk #{disk_name} for #{vm_name}") result = connection.attach_disk(project, zone(pool_name), vm_object.name, attached_disk) wait_for_operation(project, pool_name, result) true end |
#create_snapshot(pool_name, vm_name, new_snapshot_name) ⇒ Object
create_snapshot creates new snapshots with the unique name new_snapshot_name-#Vmpooler::PoolManager::Provider::Gce.diskdisk.name for one vm, and one create_snapshot() there could be multiple snapshots created, one for each drive. since the snapshot resource needs a unique name in the gce project, we create a unique name by concatenating new_snapshot_name-#Vmpooler::PoolManager::Provider::Gce.diskdisk.name the disk name is based on vm_name which makes it unique. The snapshot is added labels snapshot_name, vm, pool, diskname and boot inputs
[String] pool_name : Name of the pool
[String] vm_name : Name of the existing VM
[String] new_snapshot_name : a unique name for this snapshot, which would be used to refer to it when reverting
returns
[boolean] true : once the operations are finished
raises
RuntimeError if the vm_name cannot be found
RuntimeError if the snapshot_name already exists for this VM
286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 |
# File 'lib/vmpooler/providers/gce.rb', line 286 def create_snapshot(pool_name, vm_name, new_snapshot_name) debug_logger('create_snapshot') begin vm_object = connection.get_instance(project, zone(pool_name), vm_name) rescue ::Google::Apis::ClientError => e raise e unless e.status_code == 404 # if it does not exist raise("VM #{vm_name} in pool #{pool_name} does not exist for the provider #{name}") end old_snap = find_snapshot(vm_name, new_snapshot_name) raise("Snapshot #{new_snapshot_name} for VM #{vm_name} in pool #{pool_name} already exists for the provider #{name}") unless old_snap.nil? result_list = [] vm_object.disks.each do |attached_disk| disk_name = disk_name_from_source(attached_disk) snapshot_obj = ::Google::Apis::ComputeV1::Snapshot.new( name: "#{new_snapshot_name}-#{disk_name}", labels: { 'snapshot_name' => new_snapshot_name, 'vm' => vm_name, 'pool' => pool_name, 'diskname' => disk_name, 'boot' => attached_disk.boot.to_s } ) debug_logger("trigger async create_disk_snapshot #{vm_name}: #{new_snapshot_name}-#{disk_name}") result = connection.create_disk_snapshot(project, zone(pool_name), disk_name, snapshot_obj) # do them all async, keep a list, check later result_list << result end # now check they are done result_list.each do |result| wait_for_operation(project, pool_name, result) end true end |
#create_vm(pool_name, new_vmname) ⇒ Object
create_vm creates a new VM with a default network from the config, a initial disk named #new_vmname-disk0 that uses the ‘template’ as its source image and labels added for vm and pool and an instance configuration for machine_type from the config and labels vm and pool having a label “pool” that match a pool config name. inputs
[String] pool : Name of the pool
[String] new_vmname : Name to give the new VM
returns
[Hashtable] of the VM as per get_vm(pool_name, vm_name)
166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 |
# File 'lib/vmpooler/providers/gce.rb', line 166 def create_vm(pool_name, new_vmname) debug_logger('create_vm') pool = pool_config(pool_name) raise("Pool #{pool_name} does not exist for the provider #{name}") if pool.nil? # harcoded network info network_interfaces = Google::Apis::ComputeV1::NetworkInterface.new( network: network_name ) network_interfaces.subnetwork = subnetwork_name(pool_name) if subnetwork_name(pool_name) init_params = Google::Apis::ComputeV1::AttachedDiskInitializeParams.new( source_image: pool['template'], # The source image to create this disk. labels: { 'vm' => new_vmname, 'pool' => pool_name }, disk_name: "#{new_vmname}-disk0" ) # rubocop:disable Style/IfUnlessModifier if pool['disk_type'] && !pool['disk_type'].empty? init_params.disk_type = "https://www.googleapis.com/compute/v1/projects/#{project}/zones/#{zone(pool_name)}/diskTypes/#{pool['disk_type']}" end # rubocop:enable Style/IfUnlessModifier disk = Google::Apis::ComputeV1::AttachedDisk.new( auto_delete: true, boot: true, initialize_params: init_params ) append_domain = domain(pool_name) fqdn = "#{new_vmname}.#{append_domain}" if append_domain # Assume all pool config is valid i.e. not missing client = ::Google::Apis::ComputeV1::Instance.new( name: new_vmname, hostname: fqdn, machine_type: pool['machine_type'], disks: [disk], network_interfaces: [network_interfaces], labels: { 'vm' => new_vmname, 'pool' => pool_name }, tags: Google::Apis::ComputeV1::Tags.new(items: [project]) ) debug_logger('trigger insert_instance') result = connection.insert_instance(project, zone(pool_name), client) wait_for_operation(project, pool_name, result) get_vm(pool_name, new_vmname) end |
#debug_logger(message, send_to_upstream: false) ⇒ Object
used in local dev environment, set DEBUG_FLAG=true this way the upstream vmpooler manager does not get polluted with logs
707 708 709 710 711 |
# File 'lib/vmpooler/providers/gce.rb', line 707 def debug_logger(, send_to_upstream: false) # the default logger is simple and does not enforce debug levels (the first argument) puts if ENV['DEBUG_FLAG'] logger.log('[g]', ) if send_to_upstream end |
#destroy_vm(pool_name, vm_name) ⇒ Object
destroy_vm deletes an existing VM instance and any disks and snapshots via the labels in gce instances, disks and snapshots are resources that can exist independent of each other inputs
[String] pool_name : Name of the pool
[String] vm_name : Name of the existing VM
returns
[boolean] true : once the operations are finished
412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 |
# File 'lib/vmpooler/providers/gce.rb', line 412 def destroy_vm(pool_name, vm_name) debug_logger('destroy_vm') deleted = false begin connection.get_instance(project, zone(pool_name), vm_name) rescue ::Google::Apis::ClientError => e raise e unless e.status_code == 404 # If a VM doesn't exist then it is effectively deleted deleted = true debug_logger("instance #{vm_name} already deleted") end unless deleted debug_logger("trigger delete_instance #{vm_name}") result = connection.delete_instance(project, zone(pool_name), vm_name) wait_for_operation(project, pool_name, result, 10) end # list and delete any leftover disk, for instance if they were detached from the instance filter = "(labels.vm = #{vm_name})" disk_list = connection.list_disks(project, zone(pool_name), filter: filter) result_list = [] disk_list.items&.each do |disk| debug_logger("trigger delete_disk #{disk.name}") result = connection.delete_disk(project, zone(pool_name), disk.name) # do them all async, keep a list, check later result_list << result end # now check they are done result_list.each do |r| wait_for_operation(project, pool_name, r) end # list and delete leftover snapshots, this could happen if snapshots were taken, # as they are not removed when the original disk is deleted or the instance is detroyed snapshot_list = find_all_snapshots(vm_name) result_list = [] snapshot_list&.each do |snapshot| debug_logger("trigger delete_snapshot #{snapshot.name}") result = connection.delete_snapshot(project, snapshot.name) # do them all async, keep a list, check later result_list << result end # now check they are done result_list.each do |r| wait_for_operation(project, pool_name, r) end true end |
#disk_name_from_source(attached_disk) ⇒ Object
701 702 703 |
# File 'lib/vmpooler/providers/gce.rb', line 701 def disk_name_from_source(attached_disk) attached_disk.source.split('/')[-1] # disk name is after the last / of the full source URL end |
#domain(pool_name) ⇒ Object
85 86 87 88 |
# File 'lib/vmpooler/providers/gce.rb', line 85 def domain(pool_name) dns_plugin_name = pool_config(pool_name)['dns_plugin'] dns_config(dns_plugin_name) end |
#ensured_gce_connection(connection_pool_object) ⇒ Object
644 645 646 647 |
# File 'lib/vmpooler/providers/gce.rb', line 644 def ensured_gce_connection(connection_pool_object) connection_pool_object[:connection] = connect_to_gce unless connection_pool_object[:connection] connection_pool_object[:connection] end |
#find_all_snapshots(vm_name) ⇒ Object
find all snapshots ever created for one vm, regardless of snapshot name, for example when deleting it all
695 696 697 698 699 |
# File 'lib/vmpooler/providers/gce.rb', line 695 def find_all_snapshots(vm_name) filter = "(labels.vm = #{vm_name})" snapshot_list = connection.list_snapshots(project, filter: filter) snapshot_list.items # array of snapshot objects end |
#find_snapshot(vm_name, snapshotname) ⇒ Object
this is used because for one vm, with the same snapshot name there could be multiple snapshots, one for each disk
687 688 689 690 691 |
# File 'lib/vmpooler/providers/gce.rb', line 687 def find_snapshot(vm_name, snapshotname) filter = "(labels.vm = #{vm_name}) AND (labels.snapshot_name = #{snapshotname})" snapshot_list = connection.list_snapshots(project, filter: filter) snapshot_list.items # array of snapshot objects end |
#generate_vm_hash(vm_object, pool_name) ⇒ Object
Return a hash of VM data Provides vmname, hostname, template, poolname, boottime, status, zone, machine_type, labels, label_fingerprint, ip information
625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 |
# File 'lib/vmpooler/providers/gce.rb', line 625 def generate_vm_hash(vm_object, pool_name) pool_configuration = pool_config(pool_name) return nil if pool_configuration.nil? { 'name' => vm_object.name, 'hostname' => vm_object.hostname, 'template' => pool_configuration&.key?('template') ? pool_configuration['template'] : nil, # was expecting to get it from API, not from config, but this is what vSphere does too! 'poolname' => vm_object.labels&.key?('pool') ? vm_object.labels['pool'] : nil, 'boottime' => vm_object., 'status' => vm_object.status, # One of the following values: PROVISIONING, STAGING, RUNNING, STOPPING, SUSPENDING, SUSPENDED, REPAIRING, and TERMINATED 'zone' => vm_object.zone, 'machine_type' => vm_object.machine_type, 'labels' => vm_object.labels, 'label_fingerprint' => vm_object.label_fingerprint, 'ip' => vm_object.network_interfaces ? vm_object.network_interfaces.first.network_ip : nil } end |
#get_current_user(vm_name) ⇒ Object
575 576 577 578 579 580 581 582 583 584 585 586 587 |
# File 'lib/vmpooler/providers/gce.rb', line 575 def get_current_user(vm_name) @redis.with_metrics do |redis| user = redis.hget("vmpooler__vm__#{vm_name}", 'token:user') return '' if user.nil? # cleanup so it's a valid label value # can't have upercase user = user.downcase # replace invalid chars with dash user = user.gsub(/[^0-9a-z_-]/, '-') return user end end |
#get_vm(pool_name, vm_name) ⇒ Object
inputs
[String] pool_name : Name of the pool
[String] vm_name : Name of the VM to find
returns
nil if VM doesn't exist
[Hastable] of the VM
[String] name : The name of the resource, provided by the client when initially creating the resource
[String] hostname : Specifies the hostname of the instance. The specified hostname must be RFC1035 compliant. If hostname is not specified,
the default hostname is [ INSTANCE_NAME].c.[PROJECT_ID].internal when using the global DNS, and
[ INSTANCE_NAME].[ZONE].c.[PROJECT_ID].internal when using zonal DNS
[String] template : This is the name of template
[String] poolname : Name of the pool the VM as per labels
[Time] boottime : Time when the VM was created/booted
[String] status : One of the following values: PROVISIONING, STAGING, RUNNING, STOPPING, SUSPENDING, SUSPENDED, REPAIRING, and TERMINATED
[String] zone : URL of the zone where the instance resides.
[String] machine_type : Full or partial URL of the machine type resource to use for this instance, in the format: zones/zone/machineTypes/machine-type.
136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 |
# File 'lib/vmpooler/providers/gce.rb', line 136 def get_vm(pool_name, vm_name) debug_logger('get_vm') vm_hash = nil begin vm_object = connection.get_instance(project, zone(pool_name), vm_name) rescue ::Google::Apis::ClientError => e raise e unless e.status_code == 404 # swallow the ClientError error 404 and return nil when the VM was not found return nil end return vm_hash if vm_object.nil? vm_hash = generate_vm_hash(vm_object, pool_name) debug_logger("vm_hash #{vm_hash}") vm_hash end |
#get_vm_ip_address(vm_name, pool_name) ⇒ Object
211 212 213 214 |
# File 'lib/vmpooler/providers/gce.rb', line 211 def get_vm_ip_address(vm_name, pool_name) vm_object = get_vm(pool_name, vm_name) vm_object['ip'] end |
#machine_type(pool_name) ⇒ Object
80 81 82 83 |
# File 'lib/vmpooler/providers/gce.rb', line 80 def machine_type(pool_name) return pool_config(pool_name)['machine_type'] if pool_config(pool_name)['machine_type'] return provider_config['machine_type'] if provider_config['machine_type'] end |
#name ⇒ Object
name of the provider class
51 52 53 |
# File 'lib/vmpooler/providers/gce.rb', line 51 def name 'gce' end |
#network_name ⇒ Object
66 67 68 |
# File 'lib/vmpooler/providers/gce.rb', line 66 def network_name provider_config['network_name'] end |
#open_socket(host, domain = nil, timeout = 5, port = 22, &_block) ⇒ Object
This should supercede the open_socket method in the Pool Manager
674 675 676 677 678 679 680 681 682 683 |
# File 'lib/vmpooler/providers/gce.rb', line 674 def open_socket(host, domain = nil, timeout = 5, port = 22, &_block) target_host = host target_host = "#{host}.#{domain}" if domain sock = TCPSocket.new(target_host, port, connect_timeout: timeout) begin yield sock if block_given? ensure sock.close end end |
#project ⇒ Object
main configuration options
62 63 64 |
# File 'lib/vmpooler/providers/gce.rb', line 62 def project provider_config['project'] end |
#purge_unconfigured_resources(allowlist) ⇒ Object
Scans zones that are configured for list of resources (VM, disks, snapshots) that do not have the label.pool set to one of the configured pools. If it is also not in the allowlist, the resource is destroyed
479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 |
# File 'lib/vmpooler/providers/gce.rb', line 479 def purge_unconfigured_resources(allowlist) debug_logger('purge_unconfigured_resources') pools_array = provided_pools filter = {} # we have to group things by zone, because the API search feature is done against a zone and not global # so we will do the searches in each configured zone pools_array.each do |pool| filter[zone(pool)] = [] if filter[zone(pool)].nil? filter[zone(pool)] << "(labels.pool != #{pool})" end filter.each_key do |zone| # this filter should return any item that have a labels.pool that is not in the config OR # do not have a pool label at all filter_string = "#{filter[zone].join(' AND ')} OR -labels.pool:*" # VMs instance_list = connection.list_instances(project, zone, filter: filter_string) result_list = [] instance_list.items&.each do |vm| next if should_be_ignored(vm, allowlist) debug_logger("trigger async delete_instance #{vm.name}") result = connection.delete_instance(project, zone, vm.name) result_list << result end # now check they are done result_list.each do |result| wait_for_zone_operation(project, zone, result) end # Disks disks_list = connection.list_disks(project, zone, filter: filter_string) disks_list.items&.each do |disk| next if should_be_ignored(disk, allowlist) debug_logger("trigger async no wait delete_disk #{disk.name}") connection.delete_disk(project, zone, disk.name) end # Snapshots snapshot_list = connection.list_snapshots(project, filter: filter_string) next if snapshot_list.items.nil? snapshot_list.items.each do |sn| next if should_be_ignored(sn, allowlist) debug_logger("trigger async no wait delete_snapshot #{sn.name}") connection.delete_snapshot(project, sn.name) end end end |
#revert_snapshot(pool_name, vm_name, snapshot_name) ⇒ Object
revert_snapshot reverts an existing VM’s disks to an existing snapshot_name reverting in gce entails
-
shutting down the VM,
-
detaching and deleting the drives,
-
creating new disks with the same name from the snapshot for each disk
-
attach disks and start instance
for one vm, there might be multiple snapshots in time. We select the ones referred to by the snapshot_name, but that may be multiple snapshots, one for each disks The new disk is added labels vm and pool inputs
[String] pool_name : Name of the pool
[String] vm_name : Name of the existing VM
[String] snapshot_name : Name of an existing snapshot
returns
[boolean] true : once the operations are finished
raises
RuntimeError if the vm_name cannot be found
RuntimeError if the snapshot_name already exists for this VM
343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 |
# File 'lib/vmpooler/providers/gce.rb', line 343 def revert_snapshot(pool_name, vm_name, snapshot_name) debug_logger('revert_snapshot') begin vm_object = connection.get_instance(project, zone(pool_name), vm_name) rescue ::Google::Apis::ClientError => e raise e unless e.status_code == 404 # if it does not exist raise("VM #{vm_name} in pool #{pool_name} does not exist for the provider #{name}") end snapshot_object = find_snapshot(vm_name, snapshot_name) raise("Snapshot #{snapshot_name} for VM #{vm_name} in pool #{pool_name} does not exist for the provider #{name}") if snapshot_object.nil? # Shutdown instance debug_logger("trigger stop_instance #{vm_name}") result = connection.stop_instance(project, zone(pool_name), vm_name) wait_for_operation(project, pool_name, result) # Delete existing disks vm_object.disks&.each do |attached_disk| debug_logger("trigger detach_disk #{vm_name}: #{attached_disk.device_name}") result = connection.detach_disk(project, zone(pool_name), vm_name, attached_disk.device_name) wait_for_operation(project, pool_name, result) current_disk_name = disk_name_from_source(attached_disk) debug_logger("trigger delete_disk #{vm_name}: #{current_disk_name}") result = connection.delete_disk(project, zone(pool_name), current_disk_name) wait_for_operation(project, pool_name, result) end # this block is sensitive to disruptions, for example if vmpooler is stopped while this is running snapshot_object.each do |snapshot| current_disk_name = snapshot.labels['diskname'] bootable = (snapshot.labels['boot'] == 'true') disk = Google::Apis::ComputeV1::Disk.new( name: current_disk_name, labels: { 'pool' => pool_name, 'vm' => vm_name }, source_snapshot: snapshot.self_link ) # create disk in GCE as a separate resource debug_logger("trigger insert_disk #{vm_name}: #{current_disk_name} based on #{snapshot.self_link}") result = connection.insert_disk(project, zone(pool_name), disk) wait_for_operation(project, pool_name, result) # read the new disk info new_disk_info = connection.get_disk(project, zone(pool_name), current_disk_name) new_attached_disk = Google::Apis::ComputeV1::AttachedDisk.new( auto_delete: true, boot: bootable, source: new_disk_info.self_link ) # attach the new disk to existing instance debug_logger("trigger attach_disk #{vm_name}: #{current_disk_name}") result = connection.attach_disk(project, zone(pool_name), vm_name, new_attached_disk) wait_for_operation(project, pool_name, result) end debug_logger("trigger start_instance #{vm_name}") result = connection.start_instance(project, zone(pool_name), vm_name) wait_for_operation(project, pool_name, result) true end |
#should_be_ignored(item, allowlist) ⇒ Object
END BASE METHODS
562 563 564 565 566 567 568 569 570 571 572 573 |
# File 'lib/vmpooler/providers/gce.rb', line 562 def should_be_ignored(item, allowlist) return false if allowlist.nil? allowlist.map!(&:downcase) # remove uppercase from configured values because its not valid as resource label array_flattened_labels = [] item.labels&.each do |k, v| array_flattened_labels << "#{k}=#{v}" end (!item.labels.nil? && allowlist&.include?(item.labels['pool'])) || # the allow list specifies the value within the pool label (allowlist&.include?('') && !item.labels&.keys&.include?('pool')) || # the allow list specifies "" string and the pool label is not set !(allowlist & array_flattened_labels).empty? # the allow list specify a fully qualified label eg user=Bob and the item has it end |
#subnetwork_name(pool_name) ⇒ Object
70 71 72 |
# File 'lib/vmpooler/providers/gce.rb', line 70 def subnetwork_name(pool_name) return pool_config(pool_name)['subnetwork_name'] if pool_config(pool_name)['subnetwork_name'] end |
#tag_vm_user(pool, vm_name) ⇒ Object
tag_vm_user This method is called once we know who is using the VM (it is running). This method enables seeing who is using what in the provider pools.
inputs
[String] pool_name : Name of the pool
[String] vm_name : Name of the VM to check if ready
returns
[Boolean] : true if successful, false if an error occurred and it should retry
539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 |
# File 'lib/vmpooler/providers/gce.rb', line 539 def tag_vm_user(pool, vm_name) user = get_current_user(vm_name) vm_hash = get_vm(pool, vm_name) return false if vm_hash.nil? new_labels = vm_hash['labels'] # bailing in this case since labels should exist, and continuing would mean losing them return false if new_labels.nil? # add new label called token-user, with value as user new_labels['token-user'] = user begin instances_set_labels_request_object = Google::Apis::ComputeV1::InstancesSetLabelsRequest.new(label_fingerprint: vm_hash['label_fingerprint'], labels: new_labels) result = connection.set_instance_labels(project, zone(pool), vm_name, instances_set_labels_request_object) wait_for_zone_operation(project, zone(pool), result) rescue StandardError => _e return false end true end |
#vm_ready?(pool_name, vm_name, redis) ⇒ Boolean
463 464 465 466 467 468 469 470 471 472 473 474 475 |
# File 'lib/vmpooler/providers/gce.rb', line 463 def vm_ready?(pool_name, vm_name, redis) debug_logger('vm_ready?') begin # TODO: we could use a healthcheck resource attached to instance domain = domain(pool_name) open_socket(vm_name, domain) rescue StandardError => e redis.hset("vmpooler__vm__#{vm_name}", 'open_socket_error', e.to_s) return false end redis.hdel("vmpooler__vm__#{vm_name}", 'open_socket_error') true end |
#vms_in_pool(pool_name) ⇒ Object
vms_in_pool lists all the VM names in a pool, which is based on the VMs having a label “pool” that match a pool config name. inputs
[String] pool_name : Name of the pool
returns
empty array [] if no VMs found in the pool
[Array]
[Hashtable]
[String] name : the name of the VM instance (unique for whole project)
101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 |
# File 'lib/vmpooler/providers/gce.rb', line 101 def vms_in_pool(pool_name) debug_logger('vms_in_pool') vms = [] pool = pool_config(pool_name) raise("Pool #{pool_name} does not exist for the provider #{name}") if pool.nil? zone = zone(pool_name) filter = "(labels.pool = #{pool_name})" instance_list = connection.list_instances(project, zone, filter: filter) return vms if instance_list.items.nil? instance_list.items.each do |vm| vms << { 'name' => vm.name } end debug_logger(vms) vms end |
#wait_for_operation(project, pool_name, result, retries = 5) ⇒ Object
619 620 621 |
# File 'lib/vmpooler/providers/gce.rb', line 619 def wait_for_operation(project, pool_name, result, retries = 5) wait_for_zone_operation(project, zone(pool_name), result, retries) end |
#wait_for_zone_operation(project, zone, result, retries = 5) ⇒ Object
Compute resource wait for operation to be DONE (synchronous operation)
590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 |
# File 'lib/vmpooler/providers/gce.rb', line 590 def wait_for_zone_operation(project, zone, result, retries = 5) while result.status != 'DONE' result = connection.wait_zone_operation(project, zone, result.name) debug_logger(" -> wait_for_zone_operation status #{result.status} (#{result.name})") end if result.error # unsure what kind of error can be stored here = '' # array of errors, combine them all result.error.errors.each do |error| = "#{} #{error.code}:#{error.}" end raise "Operation: #{result.description} failed with error: #{}" end result rescue Google::Apis::TransmissionError => e # Error returned once timeout reached, each retry typically about 1 minute. if retries.positive? retries -= 1 retry end raise rescue Google::Apis::ClientError => e raise e unless e.status_code == 404 # if the operation is not found, and we are 'waiting' on it, it might be because it # is already finished puts "waited on #{result.name} but was not found, so skipping" end |
#zone(pool_name) ⇒ Object
main configuration options, overridable for each pool
75 76 77 78 |
# File 'lib/vmpooler/providers/gce.rb', line 75 def zone(pool_name) return pool_config(pool_name)['zone'] if pool_config(pool_name)['zone'] return provider_config['zone'] if provider_config['zone'] end |