Puppet Class: profile::hadoop::common
- Defined in:
- modules/profile/manifests/hadoop/common.pp
Overview
SPDX-License-Identifier: Apache-2.0
Class profile::hadoop::common
Configures Hadoop common configuration, the baseline for all the other services/daemons/clients. This includes the Hadoop client packages as well. The main goal of this profile is to keep all the Hadoop cluster daemons/clients in sync with one single configuration.
This profile uses some defaults that are good for a generic use case, like testing in labs, but probably not for production.
Parameters
[*zookeeper_clusters*]
List of available/configured Zookeeper clusters and their properties.
[*hadoop_clusters*]
List of available/configured Hadoop clusters and their properties.
[*hadoop_clusters_secrets*]
Hash of available/configured Hadoop clusters and their secret properties,
like passwords, etc..
The following values will be checked in the hash table only if any Yarn/HDFS/MapRed
TLS config is enabled (see in the code for the exact values).
- 'ssl_keystore_keypassword' -> Related parameter for ssl-(server|client).xml
- 'ssl_keystore_password' -> Related parameter for ssl-(server|client).xml
- 'ssl_trustore_password' -> Related parameter for ssl-(server|client).xml
Default: {}
[*hadoop_cluster_name*]
The Hadoop cluster name to pick up config properties from.
Default: 'bigtop'
[*config_override*]
Hash of Hadoop properties that override the ones defined in the
hadoop_clusters's variable configuration.
Default: {}
[*ensure_ssl_config*]
Extra argument to force the profile to not deploy TLS keys if any of the
Yarn/HDFS/MapReduce TLS config has been added. This is useful in places where
we need the TLS config to be picked (for example to use the encrypted shuffle
in map-reduce jobs) but not TLS keys are available for the host.
Default: false
Hadoop properties
These properties can be added to either hadoop_clusters or config_override's
hashes, and they configure specific Hadoop functionality.
[*zookeeper_cluster_name*]
The zookeeper cluster name to use.
[*resourcemanager_hosts*]
List of hostnames acting as Yarn Resource Managers for the cluster.
[*cluster_name*]
Name of the Hadoop cluster.
[*namenode_hosts*]
List of hostnames acting as HDFS Namenodes for the cluster.
[*journalnode_hosts*]
List of hostnames acting as HDFS Journalnodes for the cluster.
[*datanode_volumes_failed_tolerated*]
Number of disk/volume failures tolerated by the datanode before
shutting down.
Default: undef
[*hdfs_trash_checkpoint_interval*]
Number of minutes to wait before creating a trash checkpoint directory
in each home directory.
Default: undef
[*hdfs_trash_interval*]
Number of minutes to wait before considering a trash checkpoint stale/old
and hence eligible for deletion. This parameter enables the HDFS trash
functionality even without setting hdfs_trash_checkpoint_interval, but
keep in mind that its default value for hadoop will be 0 (every time the
checkpointer runs it creates a new checkpoint out of current and removes
checkpoints created more than hdfs_trash_interval minutes ago).
Default: undef
[*mapreduce_reduce_shuffle_parallelcopies*]
Map-reduce specific setting.
Default: undef
[*mapreduce_task_io_sort_mb*]
Map-reduce specific setting.
Default: undef
[*mapreduce_task_io_sort_factor*]
Map-reduce specific setting.
Default: undef
[*mapreduce_map_memory_mb*]
Map container reserved memory.
Default: undef
[*mapreduce_map_java_opts*]
Map container JVM ops settings.
Default: undef
[*mapreduce_reduce_memory_mb*]
Reduce container reserved memory.
Default: undef
[*mapreduce_reduce_java_opts*]
Reduce container JVM ops settings.
Default: undef
[*yarn_heapsize*]
Yarn Node and Resource Manager max heap size.
Default: undef
[*yarn_nodemanager_opts*]
Yarn Node Manager JVM opts.
Default: undef
[*yarn_resourcemanager_opts*]
Yarn Resource Manager JVM opts.
Default: undef
[*hadoop_heapsize*]
HDFS daemons maximum heapsize.
Default: undef
[*hadoop_datanode_opts*]
HDFS datanode JVM opts.
Default: undef
[*hadoop_journalnode_opts*]
HDFS journalnode JVM opts.
Default: undef
[*hadoop_namenode_opts*]
JVM opts to pass to the HDFS Namenode daemon.
If you change these values please check profile::hadoop::*::namenode_heapsize
since some alarms need to be tuned in the master/standby config too.
Default: undef
[*yarn_app_mapreduce_am_resource_mb*]
Yarn Application Master container size (Mb).
Default: undef
[*yarn_app_mapreduce_am_command_opts*]
Yarn Application Master JVM opts.
Default: undef
[*mapreduce_history_java_opts*]
Map-reduce History server JVM opts.
Default: undef
[*yarn_scheduler_minimum_allocation_vcores*]
Yarn scheduler specific setting.
Default: undef
[*yarn_scheduler_maximum_allocation_vcores*]
Yarn scheduler specific setting.
Default: undef
[*yarn_nodemanager_os_reserved_memory_mb*]
Map-reduce specific setting. If set, yarn_nodemanager_resource_memory_mb will
be set as total_memory_on_host - yarn_nodemanager_os_reserved_memory_mb.
Default: undef
[*yarn_scheduler_minimum_allocation_mb*]
Yarn scheduler specific setting.
Default: undef
[*yarn_scheduler_maximum_allocation_mb*]
Yarn scheduler specific setting. If not set, but reserved_memory_mb and total_memory_mb are,
This will be set to total_memory_mb - reserved_memory_mb.
Default: undef
[*java_home*]
Sets the JAVA_HOME env. variable in hadoop-env.sh
[*net_topology*]
A mapping of FQDN hostname to 'rack'. This will be used by net-topology.py.erb
to render a script that will be used for Hadoop node rack awareness.
[*datanode_mounts_prefix*]
Gets the list of partitions mounted on the host that match a given prefix
to form the list of mountpoints that Yarn and HDFS will rely on.
IMPORTANT: make sure that the partitions are mounted on the OS before using this
parameter.
Default: '/var/lib/hadoop/data'
191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 |
# File 'modules/profile/manifests/hadoop/common.pp', line 191
class profile::hadoop::common (
Hash[String, Any] $zookeeper_clusters = lookup('zookeeper_clusters'),
Hash[String, Any] $hadoop_clusters = lookup('hadoop_clusters'),
String $cluster_name = lookup('profile::hadoop::common::hadoop_cluster_name'),
Hash[String, Any] $hadoop_clusters_secrets = lookup('hadoop_clusters_secrets', { 'default_value' => {} }),
Hash[String, Any] $config_override = lookup('profile::hadoop::common::config_override', { 'default_value' => {} }),
Boolean $ensure_ssl_config = lookup('profile::hadoop::common::ensure_ssl_config', { 'default_value' => false }),
String $datanode_mounts_prefix = lookup('profile::hadoop::common::datanode_mounts_prefix', { 'default_value' => '/var/lib/hadoop/data'}),
Optional[Integer] $min_datanode_mounts = lookup('profile::hadoop::common::min_datanode_mounts', { 'default_value' => undef }),
) {
# Properties that are not meant to have undef as default value (a hash key
# without a correspondent value returns undef) should be listed in here.
$hadoop_default_config = {
'hadoop_var_directory' => '/var/lib/hadoop',
'java_home' => '/usr/lib/jvm/java-8-openjdk-amd64/jre',
'dfs_block_size' => 268435456, # 256M
}
# The final Hadoop configuration is obtained merging three hashes:
# 1) Hadoop properties with a default value different than undef
# 2) Hadoop properies meant to be shared among all Hadoop daemons/services
# 3) Hadoop properties that might get overridden by specific Hadoop role/profiles.
$hadoop_config = $hadoop_default_config + $hadoop_clusters[$cluster_name] + $config_override
# This is a hash of secrets broken down by cluster name. Useful since 1) these info, like passwords,
# cannot be retrieved from the above $hadoop_config 2) All the roles that share this profile can
# get a single source of configuration, that avoids copy/paste config around.
$hadoop_secrets_config = $hadoop_clusters_secrets[$cluster_name]
$zookeeper_cluster_name = $hadoop_config['zookeeper_cluster_name']
$yarn_resourcemanager_zk_state_store_parent_path = $hadoop_config['yarn_resourcemanager_zk_state_store_parent_path']
$yarn_resourcemanager_fs_state_store_uri = $hadoop_config['yarn_resourcemanager_fs_state_store_uri']
$resourcemanager_hosts = $hadoop_config['resourcemanager_hosts']
$namenode_hosts = $hadoop_config['namenode_hosts']
$journalnode_hosts = $hadoop_config['journalnode_hosts']
$hadoop_var_directory = $hadoop_config['hadoop_var_directory']
$datanode_volumes_failed_tolerated = $hadoop_config['datanode_volumes_failed_tolerated']
$dfs_namenode_handler_count = $hadoop_config['dfs_namenode_handler_count']
$dfs_namenode_service_handler_count = $hadoop_config['dfs_namenode_service_handler_count']
$dfs_namenode_service_port = $hadoop_config['dfs_namenode_service_port']
$dfs_block_size = $hadoop_config['dfs_block_size']
$yarn_heapsize = $hadoop_config['yarn_heapsize']
$yarn_nodemanager_opts = $hadoop_config['yarn_nodemanager_opts']
$yarn_resourcemanager_opts = $hadoop_config['yarn_resourcemanager_opts']
$hadoop_heapsize = $hadoop_config['hadoop_heapsize']
$hadoop_datanode_opts = $hadoop_config['hadoop_datanode_opts']
$hadoop_journalnode_opts = $hadoop_config['hadoop_journalnode_opts']
$hadoop_namenode_opts = $hadoop_config['hadoop_namenode_opts']
$mapreduce_history_java_opts = $hadoop_config['mapreduce_history_java_opts']
$yarn_fair_scheduler_template = $hadoop_config['yarn_fair_scheduler_template']
$yarn_node_labels_enabled = $hadoop_config['yarn_node_labels_enabled']
$yarn_nodemanager_resource_memory_mb = $hadoop_config['yarn_nodemanager_os_reserved_memory_mb'] ? {
undef => undef,
default => floor($facts['memorysize_mb']) - $hadoop_config['yarn_nodemanager_os_reserved_memory_mb'],
}
$yarn_scheduler_maximum_allocation_mb = $hadoop_config['yarn_scheduler_maximum_allocation_mb']
$java_home = $hadoop_config['java_home']
$mapreduce_reduce_shuffle_parallelcopies = $hadoop_config['mapreduce_reduce_shuffle_parallelcopies'] ? {
undef => 10,
default => $hadoop_config['mapreduce_reduce_shuffle_parallelcopies'],
}
$mapreduce_task_io_sort_mb = $hadoop_config['mapreduce_task_io_sort_mb'] ? {
undef => 200,
default => $hadoop_config['mapreduce_task_io_sort_mb'],
}
$mapreduce_task_io_sort_factor = $hadoop_config['mapreduce_task_io_sort_factor'] ? {
undef => 10,
default => $hadoop_config['mapreduce_task_io_sort_factor'],
}
# Adding sane defaults to these options in case not explicitly set via hiera.
# More info: T218758
$yarn_resourcemanager_fs_state_store_retry_policy = $hadoop_config['yarn_resourcemanager_fs_state_store_retry_policy'] ? {
undef => '2000,10',
default => $hadoop_config['yarn_resourcemanager_fs_state_store_retry_policy'],
}
$yarn_resourcemanager_max_completed_applications = $hadoop_config['yarn_resourcemanager_max_completed_applications'] ? {
undef => '5000',
default => $hadoop_config['yarn_resourcemanager_max_completed_applications'],
}
$core_site_extra_properties = $hadoop_config['core_site_extra_properties'] ? {
undef => {},
default => $hadoop_config['core_site_extra_properties'],
}
$yarn_site_extra_properties = $hadoop_config['yarn_site_extra_properties'] ? {
undef => {},
default => $hadoop_config['yarn_site_extra_properties'],
}
$hdfs_site_extra_properties = $hadoop_config['hdfs_site_extra_properties'] ? {
undef => {},
default => $hadoop_config['hdfs_site_extra_properties'],
}
$mapred_site_extra_properties = $hadoop_config['mapred_site_extra_properties'] ? {
undef => {},
default => $hadoop_config['mapred_site_extra_properties'],
}
$yarn_nm_container_executor_config = $hadoop_config['yarn_nodemanager_container_executor_config'] ? {
undef => {},
default => $hadoop_config['yarn_nm_container_executor_config'],
}
$yarn_use_spark_shuffle = $hadoop_config['yarn_use_spark_shuffle'] ? {
undef => true,
default => $hadoop_config['yarn_use_spark_shuffle'],
}
$yarn_use_multi_spark_shufflers = $hadoop_config['yarn_use_multi_spark_shufflers'] ? {
undef => false,
default => $hadoop_config['yarn_use_multi_spark_shufflers'],
}
$yarn_multi_spark_shuffler_versions = $hadoop_config['yarn_multi_spark_shuffler_versions'] ? {
undef => {},
default => $hadoop_config['yarn_multi_spark_shuffler_versions'],
}
# The HDFS Trash is configured in this way:
# 1) Once every day a checkpoint is made (that contains all the trash for a day).
# 2) After a month a checkpoint is deleted.
$hdfs_trash_checkpoint_interval = $hadoop_config['hdfs_trash_checkpoint_interval'] ? {
undef => 1440,
default => $hadoop_config['hdfs_trash_checkpoint_interval'],
}
$hdfs_trash_interval = $hadoop_config['hdfs_trash_interval'] ? {
undef => 43200,
default => $hadoop_config['hdfs_trash_interval'],
}
# These Map/Reduce and YARN ApplicationMaster master settings are
# settable per job.
# Choosing 2G for default application container size.
# Map container size and JVM max heap size (-XmX)
$mapreduce_map_memory_mb = $hadoop_config['mapreduce_map_memory_mb'] ? {
undef => 2048,
default => $hadoop_config['mapreduce_map_memory_mb'],
}
$mapreduce_map_java_opts = $hadoop_config['mapreduce_map_java_opts'] ? {
undef => '-Xmx1638m', # 0.8 * 2G
default => $hadoop_config['mapreduce_map_java_opts'],
}
# Reduce container size and JVM max heap size (-Xmx)
$mapreduce_reduce_memory_mb = $hadoop_config['mapreduce_reduce_memory_mb'] ? {
undef => '4096', # 2 * 2G
default => $hadoop_config['mapreduce_reduce_memory_mb'],
}
$mapreduce_reduce_java_opts = $hadoop_config['mapreduce_reduce_java_opts'] ? {
undef => '-Xmx3276m', # 0.8 * 2 * 2G
default => $hadoop_config['mapreduce_reduce_java_opts'],
}
# Yarn ApplicationMaster container size and max heap size (-Xmx)
$yarn_app_mapreduce_am_resource_mb = $hadoop_config['yarn_app_mapreduce_am_resource_mb'] ? {
undef => 4096, # 2 * 2G
default => $hadoop_config['yarn_app_mapreduce_am_resource_mb'],
}
$yarn_app_mapreduce_am_command_opts = $hadoop_config['yarn_app_mapreduce_am_command_opts'] ? {
undef => '-Xmx3276m', # 0.8 * 2 * 2G
default => $hadoop_config['yarn_app_mapreduce_am_command_opts'],
}
# https://issues.apache.org/jira/browse/YARN-5774
# Flink 1.1+ also needs this value to be >= 0
$yarn_scheduler_minimum_allocation_mb = $hadoop_config['yarn_scheduler_minimum_allocation_mb'] ? {
undef => 1,
default => $hadoop_config['yarn_scheduler_minimum_allocation_mb'],
}
# https://issues.apache.org/jira/browse/YARN-5774
$yarn_scheduler_minimum_allocation_vcores = $hadoop_config['yarn_scheduler_minimum_allocation_vcores'] ? {
undef => 1,
default => $hadoop_config['yarn_scheduler_minimum_allocation_vcores'],
}
$yarn_scheduler_maximum_allocation_vcores = $hadoop_config['yarn_scheduler_maximum_allocation_vcores'] ? {
undef => 32,
default => $hadoop_config['yarn_scheduler_maximum_allocation_vcores'],
}
# Raised for T206943
$yarn_resourcemanager_zk_timeout_ms = $hadoop_config['yarn_resourcemanager_zk_timeout_ms'] ? {
undef => 20000,
default => $hadoop_config['yarn_resourcemanager_zk_timeout_ms'],
}
$enable_log4j_extras = $hadoop_config['enable_log4j_extras'] ? {
undef => true,
default => $hadoop_config['enable_log4j_extras'],
}
$yarn_spark_history_server_address = $hadoop_config['yarn_spark_history_server_address'] ? {
undef => undef,
default => $hadoop_config['yarn_spark_history_server_address']
}
# The datanode mountpoints are retrieved from facter, among the list of mounted
# partitions on the host. Once a partition is not available anymore (disk broken for example),
# it is sufficient to run puppet to update the configs (and restart daemons if needed).
$all_partitions = $facts['partitions'].map |$device, $partition_metadata| { $partition_metadata['mount'] }
$datanode_mounts = $all_partitions.filter |$partitions| { $datanode_mounts_prefix in $partitions }
# Fail-safe for Hadoop workers only meant to avoid running a datanode with a low number of partition by mistake.
# The minimum number of datanode partitions is set via hiera following:
# Number of datanode partitions - disk failures tolerated
if $min_datanode_mounts and length($datanode_mounts) < $min_datanode_mounts {
fail("Number of datanode mountpoints (${datanode_mounts.length}) below threshold: ${min_datanode_mounts}, please check.")
}
# Include Wikimedia's thirdparty/bigtop apt component
# as an apt source on all Hadoop hosts.
require ::profile::bigtop::apt
# Need Java before Hadoop is installed.
Class['profile::java'] -> Class['profile::hadoop::common']
$hadoop_name_directory = "${hadoop_var_directory}/name"
$hadoop_data_directory = "${hadoop_var_directory}/data"
$hadoop_journal_directory = "${hadoop_var_directory}/journal"
$zookeeper_hosts = keys($zookeeper_clusters[$zookeeper_cluster_name]['hosts'])
# If specified, this will run the wrapper script needed to generate net_topology
$net_topology = $hadoop_config['net_topology']
$net_topology_script_content = $net_topology ? {
undef => undef,
default => '/usr/local/bin/generate_net_topology.sh',
}
$core_site_extra_properties_default = {
'hadoop.ssl.enabled.protocols' => 'TLSv1.2',
'hadoop.rpc.protection' => 'privacy',
'hadoop.security.authentication' => 'kerberos',
# https://www.cloudera.com/documentation/enterprise/5-16-x/topics/cdh_sg_hiveserver2_security.html#concept_vxf_pgx_nm
'hadoop.proxyuser.hive.hosts' => '*',
'hadoop.proxyuser.hive.groups' => '*',
'hadoop.proxyuser.presto.hosts' => '*',
'hadoop.proxyuser.presto.groups' => '*',
'hadoop.proxyuser.superset.hosts' => '*',
'hadoop.proxyuser.superset.groups' =>'*',
'fs.permissions.umask-mode' => '027',
}
$yarn_site_extra_properties_default = {
# After upgrading to CDH 5.4.0, we are encountering this bug:
# https://issues.apache.org/jira/browse/MAPREDUCE-5799
# This should work around the problem.
'yarn.app.mapreduce.am.env' => 'LD_LIBRARY_PATH=/usr/lib/hadoop/lib/native',
# The default of 90.0 for this was marking older dells as unhealthy when they still
# had 2TB of space left. 99% will mark them at unhealthy with they still have
# > 200G free.
'yarn.nodemanager.disk-health-checker.max-disk-utilization-per-disk-percentage' => '99.0',
'yarn.resourcemanager.principal' => 'yarn/_HOST@WIKIMEDIA',
'yarn.nodemanager.principal' => 'yarn/_HOST@WIKIMEDIA',
'yarn.resourcemanager.keytab' => '/etc/security/keytabs/hadoop/yarn.keytab',
'yarn.nodemanager.keytab' => '/etc/security/keytabs/hadoop/yarn.keytab',
'yarn.nodemanager.linux-container-executor.group' => 'yarn',
'yarn.nodemanager.container-executor.class' => 'org.apache.hadoop.yarn.server.nodemanager.LinuxContainerExecutor',
'spark.authenticate' => true,
'spark.network.crypto.enabled' => true,
# We tried to bump the yarn.nodemanager.vmem-pmem-ratio to 10.1 in T278441,
# but in several use cases Spark containers were killed due to excessive vmem usage
# (but not the same for pmem). Bumping the ratio even more is not productive,
# it is more efficient to just disable the vmem check.
'yarn.nodemanager.vmem-check-enabled' => false,
}
$yarn_nodemanager_container_executor_config_default = {
'yarn.nodemanager.linux-container-executor.group' => 'yarn',
'min.user.id' => '100',
# hdfs is allowed to make distCP working (mapred user that can be started by hdfs
# to be able to avoid permission issues while copying).
'allowed.system.users' => 'hive,hdfs',
'banned.users' => 'root,yarn,mapred,bin,nobody',
}
$hdfs_site_extra_properties_default = {
'dfs.cluster.administrators' => 'hdfs analytics-admins,ops',
'dfs.block.access.token.enable' => 'true',
'dfs.namenode.keytab.file' => '/etc/security/keytabs/hadoop/hdfs.keytab',
'dfs.secondary.namenode.keytab.file' => '/etc/security/keytabs/hadoop/hdfs.keytab',
'dfs.namenode.kerberos.principal' => 'hdfs/_HOST@WIKIMEDIA',
'dfs.secondary.namenode.kerberos.principal' => 'hdfs/_HOST@WIKIMEDIA',
'dfs.journalnode.keytab.file' => '/etc/security/keytabs/hadoop/hdfs.keytab',
'dfs.journalnode.kerberos.principal' => 'hdfs/_HOST@WIKIMEDIA',
'dfs.journalnode.kerberos.internal.spnego.principal' => 'HTTP/_HOST@WIKIMEDIA',
'dfs.web.authentication.kerberos.keytab' => '/etc/security/keytabs/hadoop/HTTP.keytab',
'dfs.web.authentication.kerberos.principal' => 'HTTP/_HOST@WIKIMEDIA',
'dfs.datanode.keytab.file' => '/etc/security/keytabs/hadoop/hdfs.keytab',
'dfs.datanode.kerberos.principal' => 'hdfs/_HOST@WIKIMEDIA',
'dfs.encrypt.data.transfer' => 'true',
'dfs.data.transfer.protection' => 'privacy',
# This is required to allow the datanode to start:
# https://hadoop.apache.org/docs/r2.6.0/hadoop-project-dist/hadoop-common/SecureMode.html#Secure_DataNode
'dfs.http.policy' => 'HTTPS_ONLY',
'dfs.encrypt.data.transfer.cipher.suites' => 'AES/CTR/NoPadding',
'dfs.encrypt.data.transfer.cipher.key.bitlength' => '128',
}
$mapred_site_extra_properties_default = {
'mapreduce.ssl.enabled' => true,
'mapreduce.shuffle.ssl.enabled' => true,
'mapreduce.jobhistory.principal' => 'mapred/_HOST@WIKIMEDIA',
'mapreduce.jobhistory.keytab' => '/etc/security/keytabs/hadoop/mapred.keytab',
}
class { 'bigtop::hadoop':
# Default to using running resourcemanager on the same hosts
# as the namenodes.
resourcemanager_hosts => $resourcemanager_hosts,
zookeeper_hosts => $zookeeper_hosts,
yarn_resourcemanager_zk_timeout_ms => $yarn_resourcemanager_zk_timeout_ms,
yarn_resourcemanager_zk_state_store_parent_path => $yarn_resourcemanager_zk_state_store_parent_path,
yarn_resourcemanager_fs_state_store_uri => $yarn_resourcemanager_fs_state_store_uri,
yarn_resourcemanager_fs_state_store_retry_policy => $yarn_resourcemanager_fs_state_store_retry_policy,
yarn_resourcemanager_max_completed_applications => $yarn_resourcemanager_max_completed_applications,
dfs_name_dir => [$hadoop_name_directory],
dfs_journalnode_edits_dir => $hadoop_journal_directory,
dfs_datanode_failed_volumes_tolerated => $datanode_volumes_failed_tolerated,
fs_trash_checkpoint_interval => $hdfs_trash_checkpoint_interval,
fs_trash_interval => $hdfs_trash_interval,
dfs_namenode_handler_count => $dfs_namenode_handler_count,
dfs_namenode_service_handler_count => $dfs_namenode_service_handler_count,
dfs_namenode_service_port => $dfs_namenode_service_port,
cluster_name => $cluster_name,
namenode_hosts => $namenode_hosts,
journalnode_hosts => $journalnode_hosts,
datanode_mounts => $datanode_mounts,
yarn_heapsize => $yarn_heapsize,
hadoop_heapsize => $hadoop_heapsize,
yarn_nodemanager_opts => $yarn_nodemanager_opts,
yarn_resourcemanager_opts => $yarn_resourcemanager_opts,
hadoop_namenode_opts => $hadoop_namenode_opts,
hadoop_datanode_opts => $hadoop_datanode_opts,
hadoop_journalnode_opts => $hadoop_journalnode_opts,
mapreduce_history_java_opts => $mapreduce_history_java_opts,
yarn_app_mapreduce_am_resource_mb => $yarn_app_mapreduce_am_resource_mb,
yarn_app_mapreduce_am_command_opts => $yarn_app_mapreduce_am_command_opts,
yarn_nodemanager_resource_memory_mb => $yarn_nodemanager_resource_memory_mb,
yarn_scheduler_minimum_allocation_mb => $yarn_scheduler_minimum_allocation_mb,
yarn_scheduler_maximum_allocation_mb => $yarn_scheduler_maximum_allocation_mb,
yarn_scheduler_minimum_allocation_vcores => $yarn_scheduler_minimum_allocation_vcores,
yarn_scheduler_maximum_allocation_vcores => $yarn_scheduler_maximum_allocation_vcores,
yarn_use_spark_shuffle => $yarn_use_spark_shuffle,
yarn_use_multi_spark_shufflers => $yarn_use_multi_spark_shufflers,
yarn_multi_spark_shuffler_versions => $yarn_multi_spark_shuffler_versions,
yarn_node_labels_enabled => $yarn_node_labels_enabled,
dfs_block_size => 268435456, # 256 MB
io_file_buffer_size => 131072,
# Turn on Snappy compression by default for maps and final outputs
mapreduce_intermediate_compression_codec => 'org.apache.hadoop.io.compress.SnappyCodec',
mapreduce_output_compression => true,
mapreduce_output_compression_codec => 'org.apache.hadoop.io.compress.SnappyCodec',
mapreduce_output_compression_type => 'BLOCK',
mapreduce_job_reuse_jvm_num_tasks => 1,
mapreduce_reduce_shuffle_parallelcopies => $mapreduce_reduce_shuffle_parallelcopies,
mapreduce_task_io_sort_mb => $mapreduce_task_io_sort_mb,
mapreduce_task_io_sort_factor => $mapreduce_task_io_sort_factor,
mapreduce_map_memory_mb => $mapreduce_map_memory_mb,
mapreduce_map_java_opts => $mapreduce_map_java_opts,
mapreduce_reduce_memory_mb => $mapreduce_reduce_memory_mb,
mapreduce_reduce_java_opts => $mapreduce_reduce_java_opts,
net_topology_script_content => $net_topology_script_content,
net_topology => $net_topology,
# This needs to be set in order to use Impala
dfs_datanode_hdfs_blocks_metadata_enabled => true,
# Whether or not to use fair-scheduler.xml.erb to define FairScheduler queues.
fair_scheduler_template => $yarn_fair_scheduler_template,
# Yarn App Master possible port ranges
yarn_app_mapreduce_am_job_client_port_range => '55000-55199',
core_site_extra_properties => $core_site_extra_properties_default + $core_site_extra_properties,
yarn_site_extra_properties => $yarn_site_extra_properties_default + $yarn_site_extra_properties,
hdfs_site_extra_properties => $hdfs_site_extra_properties_default + $hdfs_site_extra_properties,
mapred_site_extra_properties => $mapred_site_extra_properties_default + $mapred_site_extra_properties,
yarn_nodemanager_container_executor_config => $yarn_nodemanager_container_executor_config_default + $yarn_nm_container_executor_config,
java_home => $java_home,
enable_log4j_extras => $enable_log4j_extras,
}
# The following code deploys TLS certificates to the Hadoop cluster hosts.
# Very important note too keep in mind:
# "Ensure that common name (CN) matches exactly with the fully qualified domain name (FQDN) of the server.
# The client compares the CN with the DNS domain name to ensure
# that it is indeed connecting to the desired server, not the malicious one."
# Source: https://it.hortonworks.com/blog/deploying-https-hdfs/
# When 'ensure_ssl_config' is set to true, the following assumptions are made for
# the puppet private repository:
# 1) TLS certificates are not needed since we use the host's puppet ones.
# 2) There is puppet code that copies the host's puppet TLS cert into a pks12 keystore (with custom
# password, see below) to a known location.
# 3) keystores/trustores are all encrypted with the passwords stated in the $hadoop_clusters_secrets
# hash (stored in the private repo as well).
# Also please remember that the configuration below takes care of deploying the trustores/keystores and
# the related ssl-(client|server).xml configs, but it does not enable any TLS setting for Yarn/HDFS.
# In order to do it, specific settings to the main hadoop_clusters hiera config need to be made.
#
if $ensure_ssl_config {
$hadoop_ssl_config_name = "hadoop_${cluster_name}"
$hostname_suffix = $::realm ? {
'labs' => '.eqiad.wmflabs',
default => "${::site}.wmnet",
}
$keystore_type = 'pkcs12'
$keystore_password = $hadoop_secrets_config['ssl_keystore_password']
# The keystore password is needed for the Journalnode to start,
# since not adding it or using an empty value lead to null pointer
# exceptions. Even if the key in the keystore is not pretected by a password,
# the setting needs to be present anyway. Upstream tutorials suggest to
# put this value equal to the value of the keystore password.
$keystore_keypassword = $hadoop_secrets_config['ssl_keystore_keypassword']
$keystore_path = "${bigtop::hadoop::config_directory}/ssl/server.p12"
# TODO: consider using profile::pki::get_cert
puppet::expose_agent_certs{$bigtop::hadoop::config_directory:
user => 'root',
group => 'hadoop',
provide_p12 => true,
provide_pem => false,
p12_password => $keystore_password,
}
$ssl_server_config = {
'ssl.server.keystore.type' => $keystore_type,
'ssl.server.keystore.keypassword' => $keystore_keypassword,
'ssl.server.keystore.password' => $keystore_password,
'ssl.server.keystore.location' => $keystore_path,
}
# By default we ensure that the puppet CA is trusted in the default
# JVM's truststore. No need for ssl-client.xml config in this case.
class { 'bigtop::hadoop::ssl_config':
config_directory => $::bigtop::hadoop::config_directory,
ssl_server_config => $ssl_server_config,
}
}
# Starting with Bullseye the systemd unit for systemd-logind uses ProtectSystem=strict,
# which doesn't work with HDFS, so exclude /mnt from the list of inaccessible paths for
# the systemd-logind service
if debian::codename::ge('bullseye') {
systemd::unit { 'systemd-logind.service':
content => "[Service]\nInaccessiblePaths=-/mnt\n",
restart => false,
override => true,
}
}
if $::realm == 'labs' {
# Hadoop directories in labs should be created by puppet.
# This conditional could be added to each worker,master,standby
# classes, but since it doesn't hurt to have these directories
# in labs, and since I don't want to add the $::realm conditionals
# in each class, I do it here.
file { [
$hadoop_var_directory,
$hadoop_data_directory,
]:
ensure => 'directory',
before => Class['bigtop::hadoop'],
}
}
contain 'bigtop::hadoop'
}
|