Revision 8b6fa853b54f91cf316d0078ab8d40948c6f2b55 authored by elukey on 07 June 2017, 12:24:30 UTC, committed by elukey on 07 June 2017, 12:24:30 UTC
The zookeeper client role needs to pick up a zookeeper_version value to avoid conflicts/overrides with the Cloudera repos. This commit is only temporary to allow proper testing on conf2002. Change-Id: Ibb0f246c08d3f95e5ad2540d8fdd37e6a2178e72
1 parent d0adc4c
site.pp
# vim: set tabstop=4 shiftwidth=4 softtabstop=4 expandtab textwidth=80 smarttab
# site.pp
import 'realm.pp' # These ones first
# Base nodes
# Default variables. this way, they work with an ENC (as in labs) as well.
if $cluster == undef {
$cluster = 'misc'
}
# Node definitions (alphabetic order)
node /^(acamar|achernar)\.wikimedia\.org$/ {
role(dnsrecursor, ntp)
interface::add_ip6_mapped { 'main': }
}
# url-downloaders
node /^(alsafi|aluminium)\.wikimedia\.org$/ {
role(url_downloader)
interface::add_ip6_mapped { 'main': }
}
# analytics1001 is the Hadoop master node:
# - primary active NameNode
# - YARN ResourceManager
node 'analytics1001.eqiad.wmnet' {
role(analytics_cluster::hadoop::master,
analytics_cluster::users,
# Need druid user and HDFS directories
analytics_cluster::druid::hadoop)
include ::standard
include ::base::firewall
}
# analytics1002 is the Hadoop standby NameNode and ResourceManager.
node 'analytics1002.eqiad.wmnet' {
role(analytics_cluster::hadoop::standby,
analytics_cluster::users,
# analytics1002 is usually inactive, and it has a
# decent amount of disk space. We use it to
# store backups of the analytics_cluster::database::meta
# (MySQL analytics-meta) instance. If you move this,
# make sure /srv/backup/mysql/analytics-meta has
# enough space to store backups.
analytics_cluster::database::meta::backup_dest,
# Need druid user and HDFS directories
analytics_cluster::druid::hadoop)
include ::standard
include ::base::firewall
}
# This node hosts Oozie and Hive servers,
# as well as a MySQL instance that stores
# meta data associated with those and other
# Analytics Cluster services.
#
# This node is also is a launch pad for various cron based Hadoop jobs.
# Many ingestion jobs need a starting point. Oozie is a great
# Hadoop job scheduler, but it is not better than cron
# for some jobs that need to be launched at regular time
# intervals. Cron is used for those. These crons
# do not use local resources, instead, they launch
# Hadoop jobs that run throughout the cluster.
#
node 'analytics1003.eqiad.wmnet' {
role(analytics_cluster::client,
analytics_cluster::database::meta,
# Back up analytics-meta MySQL instance
# to analytics1002. $dest is configured in
# hieradata/role/eqiad/analytics_cluster/database/meta/backup.yaml
analytics_cluster::database::meta::backup,
analytics_cluster::hive::metastore::database,
analytics_cluster::oozie::server::database,
analytics_cluster::hive::metastore,
analytics_cluster::hive::server,
analytics_cluster::oozie::server,
# Include a weekly cron job to run hdfs balancer.
analytics_cluster::hadoop::balancer,
# We need hive-site.xml in HDFS. This can be included
# on any node with a Hive client, but we really only
# want to include it in one place. analytics1003
# is a little special and standalone, so we do it here.
analytics_cluster::hive::site_hdfs,
# Camus crons import data into
# from Kafka into HDFS.
analytics_cluster::refinery::job::camus,
# Various crons that launch Hadoop jobs.
analytics_cluster::refinery,
analytics_cluster::refinery::job::data_drop,
analytics_cluster::refinery::job::project_namespace_map,
analytics_cluster::refinery::job::sqoop_mediawiki)
include ::standard
include ::base::firewall
}
# analytics1028-analytics1068 are Hadoop worker nodes.
#
# NOTE: If you add, remove or move Hadoop nodes, you should edit
# modules/role/templates/analytics_cluster/hadoop/net-topology.py.erb
# to make sure the hostname -> /datacenter/rack/row id is correct.
# This is used for Hadoop network topology awareness.
node /analytics10(2[89]|3[0-9]|4[0-9]|5[0-9]|6[0-8]).eqiad.wmnet/ {
role(analytics_cluster::hadoop::worker)
include ::base::firewall
include ::standard
}
# Analytics Query Service
node /aqs100[456789]\.eqiad\.wmnet/ {
role(aqs)
}
node 'auth1001.eqiad.wmnet' {
role('yubiauth_server')
}
node 'auth2001.codfw.wmnet' {
role('yubiauth_server')
}
node 'baham.wikimedia.org' {
role(authdns::server)
interface::add_ip6_mapped { 'main': }
}
# Bastion in Virginia
node 'bast1001.wikimedia.org' {
role(bastionhost::general)
interface::add_ip6_mapped { 'main': }
}
# Bastion in Texas
node 'bast2001.wikimedia.org' {
role(bastionhost::general)
interface::add_ip6_mapped { 'main': }
}
# Bastion in the Netherlands (replaced bast3001)
node 'bast3002.wikimedia.org' {
role(bastionhost::general,
installserver::tftp,
prometheus::ops)
interface::add_ip6_mapped { 'main': }
class { '::ganglia::monitor::aggregator': sites => 'esams', }
}
# Bastion in California
node 'bast4001.wikimedia.org' {
role(bastionhost::general,
ipmi::mgmt,
installserver::tftp,
prometheus::ops)
interface::add_ip6_mapped { 'main': }
class { '::ganglia::monitor::aggregator':
sites => 'ulsfo',
}
}
node 'bohrium.eqiad.wmnet' {
role(piwik::server)
}
# VM with webserver for misc. static sites
node 'bromine.eqiad.wmnet' {
role(webserver_misc_static)
}
# Californium hosts openstack-dashboard AKA horizon
# and Tool Labs admin console AKA Striker
# It's proxied by the misc-web varnishes
node 'californium.wikimedia.org' {
role(horizon, striker::web, labs::instance_info_dumper)
include ::standard
include ::base::firewall
include ::openstack::horizon::puppetpanel
include ::ldap::role::client::labs
}
# cerium, praseodymium and xenon are Cassandra test hosts
node /^(cerium|praseodymium|xenon)\.eqiad\.wmnet$/ {
role(restbase::test_cluster)
}
# DNS recursor
node 'chromium.wikimedia.org' {
role(dnsrecursor, ntp)
interface::add_ip6_mapped { 'main': }
}
# All gerrit servers (swap master status in hiera)
node 'cobalt.wikimedia.org', 'gerrit2001.wikimedia.org' {
role(gerrit::server)
interface::add_ip6_mapped { 'main': }
}
# conf100x are zookeeper and etcd discovery service nodes in eqiad
node /^conf100[123]\.eqiad\.wmnet$/ {
role(configcluster)
}
# conf200x are etcd/zookeeper service nodes in codfw
node /^conf200[123]\.codfw\.wmnet$/ {
role(configcluster)
}
# CI master / CI standby (switch in Hiera)
node /^(contint1001|contint2001)\.wikimedia\.org$/ {
role(ci::master,
ci::slave,
ci::website,
zuul::merger,
zuul::server)
interface::add_ip6_mapped { 'main': }
include ::contint::firewall
}
# Debian package/docker images building host in production
node 'copper.eqiad.wmnet' {
role(builder)
}
# cp1008: prod-like SSL test host
node 'cp1008.wikimedia.org' {
role(cache::text, authdns::testns)
include ::tlsproxy::prometheus
interface::add_ip6_mapped { 'main': }
}
node /^cp10(45|5[18]|61)\.eqiad\.wmnet$/ {
interface::add_ip6_mapped { 'main': }
role(cache::misc, ipsec)
}
node 'cp1046.eqiad.wmnet', 'cp1047.eqiad.wmnet', 'cp1059.eqiad.wmnet', 'cp1060.eqiad.wmnet' {
interface::add_ip6_mapped { 'main': }
role(cache::maps, ipsec)
}
node /^cp10(4[89]|50|6[234]|7[1-4]|99)\.eqiad\.wmnet$/ {
interface::add_ip6_mapped { 'main': }
role(cache::upload, ipsec)
}
node /^cp10(5[2-5]|6[5-8])\.eqiad\.wmnet$/ {
interface::add_ip6_mapped { 'main': }
role(cache::text, ipsec)
}
node /^cp20(0[147]|1[0369]|23)\.codfw\.wmnet$/ {
interface::add_ip6_mapped { 'main': }
role(cache::text, ipsec)
}
node /^cp20(0[258]|1[147]|2[0246])\.codfw\.wmnet$/ {
interface::add_ip6_mapped { 'main': }
role(cache::upload, ipsec)
}
node /^cp20(0[39]|15|21)\.codfw\.wmnet$/ {
interface::add_ip6_mapped { 'main': }
role(cache::maps, ipsec)
}
node /^cp20(06|1[28]|25)\.codfw\.wmnet$/ {
interface::add_ip6_mapped { 'main': }
role(cache::misc, ipsec)
}
node /^cp300[3-6]\.esams\.wmnet$/ {
interface::add_ip6_mapped { 'main': }
role(cache::maps, ipsec)
}
node /^cp30(0[789]|10)\.esams\.wmnet$/ {
interface::add_ip6_mapped { 'main': }
role(cache::misc, ipsec)
}
node 'cp3022.esams.wmnet' {
include ::standard
}
node /^cp30[34][0123]\.esams\.wmnet$/ {
interface::add_ip6_mapped { 'main': }
role(cache::text, ipsec)
}
node /^cp30[34][4-9]\.esams\.wmnet$/ {
interface::add_ip6_mapped { 'main': }
role(cache::upload, ipsec)
}
#
# ulsfo varnishes
#
node /^cp400[1-4]\.ulsfo\.wmnet$/ {
interface::add_ip6_mapped { 'main': }
role(cache::misc, ipsec)
}
node /^cp40(0[5-7]|1[3-5])\.ulsfo\.wmnet$/ {
interface::add_ip6_mapped { 'main': }
role(cache::upload, ipsec)
}
node /^cp40(0[89]|1[0678])\.ulsfo\.wmnet$/ {
interface::add_ip6_mapped { 'main': }
role(cache::text, ipsec)
}
node /^cp40(1[129]|20)\.ulsfo\.wmnet$/ {
interface::add_ip6_mapped { 'main': }
role(cache::maps, ipsec)
}
# temporary entry for testing new cache node hardware setup...
node 'cp4021.ulsfo.wmnet' {
interface::add_ip6_mapped { 'main': }
role(cache::upload, ipsec)
}
node 'darmstadtium.eqiad.wmnet' {
role(docker::registry)
}
node 'dataset1001.wikimedia.org' {
role(dataset::primary, dumps::server)
interface::add_ip6_mapped { 'main': }
}
# MariaDB 10
# s1 (enwiki) core production dbs on eqiad
# eqiad master
node 'db1052.eqiad.wmnet' {
class { '::role::mariadb::core':
shard => 's1',
master => true,
binlog_format => 'STATEMENT',
}
}
node /^db10(51|55|66|67|72|73|80|83|89)\.eqiad\.wmnet/ {
class { '::role::mariadb::core':
shard => 's1',
}
}
# row-based replication to sanitarium2 (T150960)
node 'db1065.eqiad.wmnet' {
class { '::role::mariadb::core':
shard => 's1',
binlog_format => 'ROW',
}
}
# s1 (enwiki) core production dbs on codfw
# codfw master
node 'db2016.codfw.wmnet' {
class { '::role::mariadb::core':
shard => 's1',
master => true,
binlog_format => 'STATEMENT',
}
}
node /^db20(34|42|48|55|62|69|70|71)\.codfw\.wmnet/ {
class { '::role::mariadb::core':
shard => 's1',
binlog_format => 'ROW',
}
}
# s2 (large wikis) core production dbs on eqiad
# eqiad master
node 'db1054.eqiad.wmnet' {
class { '::role::mariadb::core':
shard => 's2',
master => true,
binlog_format => 'STATEMENT',
}
}
node /^db10(18|21|36|60|74|76|90)\.eqiad\.wmnet/ {
class { '::role::mariadb::core':
shard => 's2',
}
}
# To be decommissioned in T162699
node 'db1024.eqiad.wmnet' {
role(spare::system)
}
# s2 (large wikis) core production dbs on codfw
# codfw master
node 'db2017.codfw.wmnet' {
class { '::role::mariadb::core':
shard => 's2',
master => true,
binlog_format => 'STATEMENT',
}
}
node /^db20(35|41|49|56|63|64)\.codfw\.wmnet/ {
class { '::role::mariadb::core':
shard => 's2',
binlog_format => 'ROW',
}
}
# s3 (default) core production dbs on eqiad
# Lots of tables!
# eqiad master
node 'db1075.eqiad.wmnet' {
class { '::role::mariadb::core':
shard => 's3',
master => true,
binlog_format => 'STATEMENT',
}
}
node /^db10(15|35|38|77|78)\.eqiad\.wmnet/ {
class { '::role::mariadb::core':
shard => 's3',
}
}
# Testing row-based replication to sanitarium2 (T150802)
node 'db1044.eqiad.wmnet' {
class { '::role::mariadb::core':
shard => 's3',
binlog_format => 'ROW',
}
}
# s3 (default) core production dbs on codfw
# codfw master
node 'db2018.codfw.wmnet' {
class { '::role::mariadb::core':
shard => 's3',
master => true,
binlog_format => 'STATEMENT',
}
}
node /^db20(36|43|50|57)\.codfw\.wmnet/ {
class { '::role::mariadb::core':
shard => 's3',
binlog_format => 'ROW',
}
}
# s4 (commons) core production dbs on eqiad
# eqiad master
node 'db1068.eqiad.wmnet' {
class { '::role::mariadb::core':
shard => 's4',
master => true,
binlog_format => 'STATEMENT',
}
}
node /^db10(53|56|59|81|84|91|97)\.eqiad\.wmnet/ {
class { '::role::mariadb::core':
shard => 's4',
}
}
# row-based replication to sanitarium2 (T150960)
node 'db1064.eqiad.wmnet' {
class { 'role::mariadb::core':
shard => 's4',
binlog_format => 'ROW',
}
}
# s4 (commons) core production dbs on codfw
# codfw master
node 'db2019.codfw.wmnet' {
class { '::role::mariadb::core':
shard => 's4',
master => true,
binlog_format => 'STATEMENT',
}
}
node /^db20(37|44|51|58|65)\.codfw\.wmnet/ {
class { '::role::mariadb::core':
shard => 's4',
binlog_format => 'ROW',
}
}
# s5 (wikidata/dewiki) core production dbs on eqiad
# eqiad master
node 'db1063.eqiad.wmnet' {
class { '::role::mariadb::core':
shard => 's5',
master => true,
binlog_format => 'STATEMENT',
}
}
# row-based replication to sanitarium2 (T153743)
node 'db1070.eqiad.wmnet' {
class { 'role::mariadb::core':
shard => 's5',
binlog_format => 'ROW',
}
}
node /^db10(26|45|49|71|82|87|92)\.eqiad\.wmnet/ {
class { '::role::mariadb::core':
shard => 's5',
}
}
# s5 (wikidata/dewiki) core production dbs on codfw
# codfw master
node 'db2023.codfw.wmnet' {
class { '::role::mariadb::core':
shard => 's5',
master => true,
binlog_format => 'STATEMENT',
}
}
node /^db20(38|45|52|59|66)\.codfw\.wmnet/ {
class { '::role::mariadb::core':
shard => 's5',
binlog_format => 'ROW',
}
}
# s6 core production dbs on eqiad
# eqiad master
node 'db1061.eqiad.wmnet' {
class { '::role::mariadb::core':
shard => 's6',
master => true,
binlog_format => 'STATEMENT',
}
}
node /^db10(30|37|50|85|88|93)\.eqiad\.wmnet/ {
class { '::role::mariadb::core':
shard => 's6',
}
}
# To be decommissioned in T166486
node 'db1023.eqiad.wmnet' {
role(spare::system)
}
# s6 core production dbs on codfw
# codfw master
node 'db2028.codfw.wmnet' {
class { '::role::mariadb::core':
shard => 's6',
master => true,
binlog_format => 'STATEMENT',
}
}
node /^db20(39|46|53|60|67)\.codfw\.wmnet/ {
class { '::role::mariadb::core':
shard => 's6',
binlog_format => 'ROW',
}
}
# To be decommissioned in T163778
node 'db1022.eqiad.wmnet' {
role(spare::system)
}
# s7 (centralauth, meta et al.) core production dbs on eqiad
# eqiad master
node 'db1062.eqiad.wmnet' {
class { '::role::mariadb::core':
shard => 's7',
master => true,
binlog_format => 'STATEMENT',
}
}
node /^db10(28|33|34|39|41|79|86|94)\.eqiad\.wmnet/ {
class { '::role::mariadb::core':
shard => 's7',
}
}
# s7 (centralauth, meta et al.) core production dbs on codfw
# codfw master
node 'db2029.codfw.wmnet' {
class { '::role::mariadb::core':
shard => 's7',
master => true,
binlog_format => 'STATEMENT',
}
}
node /^db20(40|47|54|61|68)\.codfw\.wmnet/ {
class { '::role::mariadb::core':
shard => 's7',
binlog_format => 'ROW',
}
}
## x1 shard
# eqiad
node 'db1031.eqiad.wmnet' {
class { '::role::mariadb::core':
shard => 'x1',
master => true,
binlog_format => 'ROW',
}
}
node 'db1029.eqiad.wmnet' {
class { '::role::mariadb::core':
shard => 'x1',
binlog_format => 'ROW',
}
}
# codfw
node 'db2033.codfw.wmnet' {
class { '::role::mariadb::core':
shard => 'x1',
master => true,
binlog_format => 'ROW',
}
}
## m1 shard
node 'db1016.eqiad.wmnet' {
class { '::role::mariadb::misc':
shard => 'm1',
master => true,
}
include ::base::firewall
}
node 'db1001.eqiad.wmnet' {
class { '::role::mariadb::misc':
shard => 'm1',
}
include ::base::firewall
}
node 'db2010.codfw.wmnet' {
class { '::role::mariadb::misc':
shard => 'm1',
}
include ::base::firewall
}
## m2 shard
node 'db1020.eqiad.wmnet' {
class { '::role::mariadb::misc':
shard => 'm2',
master => true,
}
}
node 'db2011.codfw.wmnet' {
class { '::role::mariadb::misc':
shard => 'm2',
}
}
## m3 shard
node 'db1043.eqiad.wmnet' {
class { '::role::mariadb::misc::phabricator':
shard => 'm3',
master => true,
}
include ::base::firewall
}
node 'db1048.eqiad.wmnet' {
class { '::role::mariadb::misc::phabricator':
shard => 'm3',
}
include ::base::firewall
}
node 'db2012.codfw.wmnet' {
class { '::role::mariadb::misc::phabricator':
shard => 'm3',
}
include ::base::firewall
}
# m4 shard
node 'db1046.eqiad.wmnet' {
class { '::role::mariadb::misc::eventlogging':
shard => 'm4',
master => true,
}
include ::base::firewall
}
node 'db1047.eqiad.wmnet' {
# this slave has an m4 custom replication protocol
# this slave additionally replicates s1 and s2
role(mariadb::analytics, mariadb::analytics::custom_repl_slave)
class { '::role::mariadb::misc::eventlogging':
shard => 'm4',
master => false,
}
include ::base::firewall
}
# m5 shard
node 'db1009.eqiad.wmnet' {
class { '::role::mariadb::misc':
shard => 'm5',
master => true,
}
}
node 'db2030.codfw.wmnet' {
class { '::role::mariadb::misc':
shard => 'm5',
}
include ::base::firewall
}
# sanitarium
node 'db1069.eqiad.wmnet' {
role(mariadb::sanitarium)
include ::base::firewall
}
node 'db1095.eqiad.wmnet' {
role(mariadb::sanitarium2)
}
# tendril db
node 'db1011.eqiad.wmnet' {
role(mariadb::tendril)
include ::base::firewall
}
node 'dbstore1001.eqiad.wmnet' {
include ::role::mariadb::backup
# 24h delay on all repl streams
class { '::role::mariadb::dbstore':
lag_warn => 90000,
lag_crit => 180000,
# Delayed slaves legitimately and cleanly (errno = 0) stop the SQL thread, so
# don't spam Icinga with warnings. This will not block properly critical alerts.
warn_stopped => false,
}
include ::base::firewall
}
node 'dbstore1002.eqiad.wmnet' {
# this slave has an m4 custom replication protocol
role(mariadb::dbstore, mariadb::analytics::custom_repl_slave)
include ::base::firewall
}
node 'dbstore2001.codfw.wmnet' {
# 24h delay on all repl streams
class { 'role::mariadb::dbstore2':
lag_warn => 90000,
lag_crit => 180000,
# Delayed slaves legitimately and cleanly (errno = 0) stop the SQL thread, so
# don't spam Icinga with warnings. This will not block properly critical alerts.
warn_stopped => false,
}
include ::base::firewall
}
node 'dbstore2002.codfw.wmnet' {
role(mariadb::dbstore)
include ::base::firewall
}
# Proxies for misc databases
node /^dbproxy100(1|6)\.eqiad\.wmnet$/ {
class { '::role::mariadb::proxy::master':
shard => 'm1',
primary_name => 'db1016',
primary_addr => '10.64.0.20',
secondary_name => 'db1001',
secondary_addr => '10.64.0.5',
}
}
node /^dbproxy100(2|7)\.eqiad\.wmnet$/ {
class { '::role::mariadb::proxy::master':
shard => 'm2',
primary_name => 'db1020',
primary_addr => '10.64.16.9',
secondary_name => 'db2011',
secondary_addr => '10.192.0.14',
}
}
node /^dbproxy100(3|8)\.eqiad\.wmnet$/ {
class { '::role::mariadb::proxy::master':
shard => 'm3',
primary_name => 'db1043',
primary_addr => '10.64.16.32',
secondary_name => 'db1048',
secondary_addr => '10.64.16.37',
}
}
node /^dbproxy100(4|9)\.eqiad\.wmnet$/ {
class { '::role::mariadb::proxy::master':
shard => 'm4',
primary_name => 'db1046',
primary_addr => '10.64.16.35',
secondary_name => 'db1047',
secondary_addr => '10.64.16.36',
}
}
node 'dbproxy1005.eqiad.wmnet' {
class { '::role::mariadb::proxy::master':
shard => 'm5',
primary_name => 'db1009',
primary_addr => '10.64.0.13',
secondary_name => 'db2030',
secondary_addr => '10.192.16.18',
}
}
# labsdb proxies (controling replica service dbs)
node 'dbproxy1010.eqiad.wmnet' {
class { '::role::mariadb::proxy::master':
shard => 'labsdb',
primary_name => 'labsdb1009',
primary_addr => '10.64.4.14',
secondary_name => 'labsdb1010',
secondary_addr => '10.64.37.23',
}
}
node 'dbproxy1011.eqiad.wmnet' {
class { '::role::mariadb::proxy::master':
shard => 'labsdb',
primary_name => 'labsdb1011',
primary_addr => '10.64.37.24',
secondary_name => 'labsdb1010',
secondary_addr => '10.64.37.23',
}
}
node /^dbmonitor[12]001\.wikimedia\.org$/ {
role(tendril)
}
# Analytics Druid servers.
# https://wikitech.wikimedia.org/wiki/Analytics/Data_Lake#Druid
node /^druid100[123].eqiad.wmnet$/ {
role(analytics_cluster::druid::worker,
analytics_cluster::hadoop::client,
# zookeeper_cluster_name is set in hiera
# in hieradata/hosts/druid100*.yaml. This
# is a separate druid zookeeper cluster.
zookeeper::server)
include ::base::firewall
include ::standard
}
node 'eeden.wikimedia.org' {
role(authdns::server)
interface::add_ip6_mapped { 'main': }
}
# icinga based monitoring hosts (einsteinium = eqiad, tegmen = codfw)
node 'einsteinium.wikimedia.org', 'tegmen.wikimedia.org' {
role(icinga, tendril, tcpircbot, certspotter)
interface::add_ip6_mapped { 'main': }
}
node /^elastic101[7-9]\.eqiad\.wmnet/ {
role(elasticsearch::cirrus)
}
node /^elastic10[2-4][0-9]\.eqiad\.wmnet/ {
role(elasticsearch::cirrus)
}
node /^elastic105[0-2]\.eqiad\.wmnet/ {
role(elasticsearch::cirrus)
}
node /^elastic20[0-3][0-9]\.codfw\.wmnet/ {
role(elasticsearch::cirrus)
}
# External Storage, Shard 1 (es1) databases
## eqiad servers
node /^es101[268]\.eqiad\.wmnet/ {
class { '::role::mariadb::core':
shard => 'es1',
}
}
## codfw servers
node /^es201[123]\.codfw\.wmnet/ {
class { '::role::mariadb::core':
shard => 'es1',
binlog_format => 'ROW',
}
}
# External Storage, Shard 2 (es2) databases
## eqiad servers
node 'es1011.eqiad.wmnet' {
class { '::role::mariadb::core':
shard => 'es2',
master => true,
binlog_format => 'ROW',
}
}
node /^es101[35]\.eqiad\.wmnet/ {
class { '::role::mariadb::core':
shard => 'es2',
binlog_format => 'ROW',
}
}
## codfw servers
node 'es2016.codfw.wmnet' {
class { '::role::mariadb::core':
shard => 'es2',
master => true,
binlog_format => 'ROW',
}
}
node /^es201[45]\.codfw\.wmnet/ {
class { '::role::mariadb::core':
shard => 'es2',
binlog_format => 'ROW',
}
}
# External Storage, Shard 3 (es3) databases
## eqiad servers
node 'es1014.eqiad.wmnet' {
class { '::role::mariadb::core':
shard => 'es3',
master => true,
binlog_format => 'ROW',
}
}
node /^es101[79]\.eqiad\.wmnet/ {
class { '::role::mariadb::core':
shard => 'es3',
binlog_format => 'ROW',
}
}
## codfw servers
node 'es2018.codfw.wmnet' {
class { '::role::mariadb::core':
shard => 'es3',
master => true,
binlog_format => 'ROW',
}
}
node /^es201[79]\.codfw\.wmnet/ {
class { '::role::mariadb::core':
shard => 'es3',
binlog_format => 'ROW',
}
}
# Disaster recovery hosts for external storage
# These nodes are temporarilly up until we get proper
# storage on the backup system
node 'es2001.codfw.wmnet' {
role(mariadb::otrsbackups)
include ::standard
include ::base::firewall
# temporary measure until mysql is uninstalled
include ::mariadb::mysqld_safe
}
node /^es200[234]\.codfw\.wmnet/ {
include ::standard
include ::base::firewall
# temporary measure until mysql is uninstalled
include ::mariadb::mysqld_safe
}
# Etcd cluster for kubernetes
# TODO: Rename the eqiad etcds to the codfw etcds naming scheme
node /^(kub)?etcd[12]00[123]\.(eqiad|codfw)\.wmnet$/ {
role(etcd::kubernetes)
}
# kubernetes masters
node /^(acrab|acrux|argon|chlorine)\.(eqiad|codfw)\.wmnet$/ {
role(kubernetes::master)
}
# Etcd cluster for "virtual" networking
node /^etcd100[456]\.eqiad\.wmnet$/ {
role(etcd::networking)
}
# Etherpad (virtual machine)
node 'etherpad1001.eqiad.wmnet' {
role(etherpad)
}
# Receives log data from Kafka processes it, and broadcasts
# to Kafka Schema based topics.
node 'eventlog1001.eqiad.wmnet' {
role(eventlogging::analytics::zeromq,
eventlogging::analytics::processor,
eventlogging::analytics::mysql,
eventlogging::analytics::files,
logging::mediawiki::errors)
include ::standard
include ::base::firewall
interface::add_ip6_mapped { 'main': }
}
# EventLogging Analytics does not (yet?) run in codfw.
node 'eventlog2001.codfw.wmnet' {
include ::standard
include ::base::firewall
}
# virtual machine for mailman list server
node 'fermium.wikimedia.org' {
role(lists)
interface::add_ip6_mapped { 'main': }
}
# ZIM dumps (https://en.wikipedia.org/wiki/ZIM_%28file_format%29)
node 'francium.eqiad.wmnet' {
role(dumps::zim)
}
# Virtualization hosts
node /^ganeti[12]00[0-9]\.(codfw|eqiad)\.wmnet$/ {
role(ganeti)
}
# Hosts visualization / monitoring of EventLogging event streams
# and MediaWiki errors.
node 'hafnium.eqiad.wmnet' {
role(webperf)
}
# debug_proxy hosts; Varnish backend for X-Wikimedia-Debug reqs
node /^(hassaleh|hassium)\.(codfw|eqiad)\.wmnet$/ {
role(debug_proxy)
}
node 'helium.eqiad.wmnet' {
role(backup)
interface::add_ip6_mapped { 'main': }
}
# Bacula storage
node 'heze.codfw.wmnet' {
role(backup::offsite)
}
# DNS recursor
node 'hydrogen.wikimedia.org' {
role(dnsrecursor, ntp)
interface::add_ip6_mapped { 'main': }
}
# irc.wikimedia.org
node 'kraz.wikimedia.org' {
role(mw_rc_irc)
interface::add_ip6_mapped { 'main': }
}
# labservices1001 hosts openstack-designate, the labs DNS service.
node 'labservices1001.wikimedia.org' {
role(labs::dns, labs::openstack::designate::server, labs::dnsrecursor,
labs::dns_floating_ip_updater)
include ::standard
include ::base::firewall
include ::ldap::role::client::labs
}
node 'labservices1002.wikimedia.org' {
role(labs::dns, labs::openstack::designate::server, labs::dnsrecursor)
include ::standard
include ::base::firewall
include ::ldap::role::client::labs
}
node 'labtestneutron2001.codfw.wmnet' {
include ::standard
}
node /^labtestvirt200[1-3]\.codfw\.wmnet$/ {
role(labs::openstack::nova::compute)
include ::standard
}
node 'labtestnet2001.codfw.wmnet' {
role(labs::openstack::nova::api, labs::openstack::nova::network)
include ::standard
}
node 'labtestcontrol2001.wikimedia.org' {
include ::standard
include ::base::firewall
role(labs::openstack::nova::controller, labs::puppetmaster)
# Labtest is weird; the mysql server is on labtestcontrol2001. So
# we need some special fw rules to allow that
$designate = ipresolve(hiera('labs_designate_hostname'),4)
$horizon = ipresolve(hiera('labs_horizon_host'),4)
$wikitech = ipresolve(hiera('labs_osm_host'),4)
$fwrules = {
mysql_designate => {
rule => "saddr ${designate} proto tcp dport (3306) ACCEPT;",
},
mysql_horizon => {
rule => "saddr ${horizon} proto tcp dport (3306) ACCEPT;",
},
mysql_wikitech => {
rule => "saddr ${wikitech} proto tcp dport (3306) ACCEPT;",
},
labspuppetbackend_horizon => {
rule => "saddr ${horizon} proto tcp dport (8100) ACCEPT;",
},
}
create_resources (ferm::rule, $fwrules)
}
node 'labtestservices2001.wikimedia.org' {
role(labs::dns, labs::openstack::designate::server, labs::dnsrecursor, openldap::labtest,
labs::dns_floating_ip_updater)
include ::standard
include ::base::firewall
}
# Primary graphite machines
node 'graphite1001.eqiad.wmnet' {
role(graphite::production, statsd, performance::site, graphite::alerts,
restbase::alerts, graphite::alerts::reqstats, elasticsearch::alerts)
}
# graphite test machine, currently with SSD caching + spinning disks
node 'graphite1002.eqiad.wmnet' {
role(test::system)
}
# graphite additional machine, for additional space
node 'graphite1003.eqiad.wmnet' {
role(graphite::production, statsd)
}
# Primary graphite machines
node 'graphite2001.codfw.wmnet' {
role(graphite::production, statsd, performance::site)
}
# graphite additional machine, for additional space
node 'graphite2002.codfw.wmnet' {
role(graphite::production, statsd)
}
# replaced carbon and install1001/install2001 (T132757, T84380, T156440)
node /^install[12]002\.wikimedia\.org$/ {
role(installserver::tftp,
installserver::dhcp,
installserver::http,
installserver::proxy,
installserver::preseed,
aptrepo::wikimedia)
interface::add_ip6_mapped { 'main': }
class { '::ganglia::monitor::aggregator':
sites => $::site,
}
}
# Phabricator
node /^(iridium\.eqiad|phab1001\.eqiad|phab2001\.codfw)\.wmnet$/ {
role(phabricator_server)
include ::ganglia
interface::add_ip6_mapped { 'main': }
}
node 'iron.wikimedia.org' {
system::role { 'misc':
description => 'Experimental Yubico two factor authentication bastion',
}
interface::add_ip6_mapped { 'main': }
role(bastionhost::twofa, access_new_install)
}
# Analytics Kafka Brokers
node /kafka10(12|13|14|18|20|22)\.eqiad\.wmnet/ {
# Kafka brokers are routed via IPv6 so that
# other DCs can address without public IPv4
# addresses.
interface::add_ip6_mapped { 'main': }
role(kafka::analytics::broker,
# Mirror all other Kafka cluster data into the analytics Kafka cluster.
kafka::analytics::mirror,
ipsec)
include ::standard
include ::base::firewall
}
# Kafka Brokers - main-eqiad and main-codfw Kafka clusters.
# For now, eventlogging-service-eventbus is also colocated
# on these brokers.
node /kafka[12]00[123]\.(eqiad|codfw)\.wmnet/ {
role(kafka::main::broker,
# Mirror eqiad.* topics from main-eqiad into main-codfw,
# or mirror codfw.* topics from main-codfw into main-eqiad.
kafka::main::mirror,
eventbus::eventbus)
include ::standard
}
# virtual machine for misc. PHP apps
node 'krypton.eqiad.wmnet' {
# kafka::analytics::burrow is a Kafka consumer lag monitor.
# Running this here because krypton is a 'misc' Jessie
# <s>monitoring host</s> (not really, it's just misc apps)
role(wikimania_scholarships, iegreview::app, grafana::production,
kafka::analytics::burrow, racktables::server)
include ::standard
}
node /kubernetes[12]00[1-4]\.(codfw|eqiad)\.wmnet/ {
role(kubernetes::worker)
include ::standard
interface::add_ip6_mapped { 'main': }
}
node 'labcontrol1001.wikimedia.org' {
role(labs::openstack::nova::controller,
labs::puppetmaster,
salt::masters::labs,
deployment::salt_masters)
include ::base::firewall
include ::standard
include ::ldap::role::client::labs
}
# labcontrol1002 is a hot spare for 1001. Switching it on
# involves uncommenting the dns role, below, and also
# changing the keystone catalog to point to labcontrol1002:
# basically repeated use of 'keystone endpoint-list,'
# 'keystone endpoint-create' and 'keystone endpoint-delete.'
node 'labcontrol1002.wikimedia.org' {
role(labs::openstack::nova::controller,
labs::puppetmaster,
salt::masters::labs,
deployment::salt_masters)
include ::base::firewall
include ::standard
include ::ldap::role::client::labs
}
# This is the testlab server that implements both:
# - silver (wikitech.wikimedia.org), and
# - californium (horizon.wikimedia.org)
node 'labtestweb2001.wikimedia.org' {
role(labs::openstack::nova::manager, mariadb::wikitech, horizon)
include ::base::firewall
include ::standard
include ::openstack::horizon::puppetpanel
include ::ldap::role::client::labs
interface::add_ip6_mapped { 'main': }
}
# Labs Graphite and StatsD host
node 'labmon1001.eqiad.wmnet' {
role(labs::graphite, grafana::labs)
include ::standard
include ::base::firewall
}
node 'labnet1001.eqiad.wmnet' {
role(labs::openstack::nova::api,
labs::openstack::nova::network,
labs::openstack::nova::fullstack)
include ::standard
}
node 'labnet1002.eqiad.wmnet' {
role(labs::openstack::nova::api)
include ::standard
}
node 'labnodepool1001.eqiad.wmnet' {
$nagios_contact_group = 'admins,contint'
role(labs::openstack::nodepool)
include ::standard
include ::base::firewall
}
## labsdb dbs
node /labsdb100[13]\.eqiad\.wmnet/ {
# this role is depecated and its nodes scheduled for decom
role(mariadb::labs_deprecated)
}
node /labsdb10(09|10|11)\.eqiad\.wmnet/ {
role(labs::db::replica)
}
node 'labsdb1004.eqiad.wmnet' {
role(postgres::master, labs::db::slave)
}
node 'labsdb1005.eqiad.wmnet' {
role(labs::db::master)
}
node 'labsdb1006.eqiad.wmnet' {
role(osm::slave)
}
node 'labsdb1007.eqiad.wmnet' {
role(osm::master)
}
node /labstore100[12]\.eqiad\.wmnet/ {
# soon to be recommissioned in T158196
include ::standard
}
node 'labstore1003.eqiad.wmnet' {
role(labs::nfs::misc)
include ::standard
}
node /labstore100[45]\.eqiad\.wmnet/ {
role(labs::nfs::secondary)
include ::standard
}
node /labstore200[1-2]\.codfw\.wmnet/ {
include ::standard
}
node 'labstore2003.codfw.wmnet' {
role(labs::nfs::secondary_backup::tools)
include ::standard
}
node 'labstore2004.codfw.wmnet' {
role(labs::nfs::secondary_backup::misc)
include ::standard
}
node 'lithium.eqiad.wmnet' {
role(syslog::centralserver)
}
node /^logstash100[1-2]\.eqiad\.wmnet$/ {
role(logstash::collector, kibana, logstash::apifeatureusage)
include ::lvs::realserver
}
node /^logstash1003\.eqiad\.wmnet$/ {
role(logstash::collector, kibana, logstash::apifeatureusage, logstash::eventlogging)
include ::lvs::realserver
}
node /^logstash100[4-6]\.eqiad\.wmnet$/ {
role(logstash::elasticsearch)
}
node /lvs100[1-6]\.wikimedia\.org/ {
# lvs100[25] are LVS balancers for the eqiad recursive DNS IP,
# so they need to use the recursive DNS backends directly
# (chromium and hydrogen) with fallback to codfw
# (doing this for all lvs for now, see T103921)
$nameservers_override = [ '208.80.154.157', '208.80.154.50', '208.80.153.254' ]
role(lvs::balancer)
interface::add_ip6_mapped { 'main': }
include ::lvs::configuration
$ips = $lvs::configuration::subnet_ips
# Set up tagged interfaces to all subnets with real servers in them
# lint:ignore:case_without_default
case $::hostname {
# lint:endignore
/^lvs100[1-3]$/: {
# Row A subnets on eth0
interface::tagged { 'eth0.1017':
base_interface => 'eth0',
vlan_id => '1017',
address => $ips['private1-a-eqiad'][$::hostname],
netmask => '255.255.252.0',
}
# Row B subnets on eth1
interface::tagged { 'eth1.1002':
base_interface => 'eth1',
vlan_id => '1002',
address => $ips['public1-b-eqiad'][$::hostname],
netmask => '255.255.255.192',
}
interface::tagged { 'eth1.1018':
base_interface => 'eth1',
vlan_id => '1018',
address => $ips['private1-b-eqiad'][$::hostname],
netmask => '255.255.252.0',
}
}
/^lvs100[4-6]$/: {
# Row B subnets on eth0
interface::tagged { 'eth0.1018':
base_interface => 'eth0',
vlan_id => '1018',
address => $ips['private1-b-eqiad'][$::hostname],
netmask => '255.255.252.0',
}
# Row A subnets on eth1
interface::tagged { 'eth1.1001':
base_interface => 'eth1',
vlan_id => '1001',
address => $ips['public1-a-eqiad'][$::hostname],
netmask => '255.255.255.192',
}
interface::tagged { 'eth1.1017':
base_interface => 'eth1',
vlan_id => '1017',
address => $ips['private1-a-eqiad'][$::hostname],
netmask => '255.255.252.0',
}
}
}
# Row C subnets on eth2
interface::tagged { 'eth2.1003':
base_interface => 'eth2',
vlan_id => '1003',
address => $ips['public1-c-eqiad'][$::hostname],
netmask => '255.255.255.192',
}
interface::tagged { 'eth2.1019':
base_interface => 'eth2',
vlan_id => '1019',
address => $ips['private1-c-eqiad'][$::hostname],
netmask => '255.255.252.0',
}
# Row D subnets on eth3
interface::tagged { 'eth3.1004':
base_interface => 'eth3',
vlan_id => '1004',
address => $ips['public1-d-eqiad'][$::hostname],
netmask => '255.255.255.224',
}
interface::tagged { 'eth3.1020':
base_interface => 'eth3',
vlan_id => '1020',
address => $ips['private1-d-eqiad'][$::hostname],
netmask => '255.255.252.0',
}
lvs::interface_tweaks {
'eth0':;
'eth1':;
'eth2':;
'eth3':;
}
}
node /^lvs10(0[789]|10)\.eqiad\.wmnet$/ {
# lvs1008,10 are LVS balancers for the eqiad recursive DNS IP,
# so they need to use the recursive DNS backends directly
# (chromium and hydrogen) with fallback to codfw
# (doing this for all lvs for now, see T103921)
$nameservers_override = [ '208.80.154.157', '208.80.154.50', '208.80.153.254' ]
role(lvs::balancer)
interface::add_ip6_mapped { 'main': }
include ::lvs::configuration
$ips = $lvs::configuration::subnet_ips
# Set up tagged interfaces to all subnets with real servers in them
# lint:ignore:case_without_default
case $::hostname {
# lint:endignore
/^lvs100[789]$/: {
# Row A subnets on eth0
interface::tagged { 'eth0.1001':
base_interface => 'eth0',
vlan_id => '1001',
address => $ips['public1-a-eqiad'][$::hostname],
netmask => '255.255.255.192',
}
# Row C subnets on eth1
interface::tagged { 'eth1.1003':
base_interface => 'eth1',
vlan_id => '1003',
address => $ips['public1-c-eqiad'][$::hostname],
netmask => '255.255.255.192',
}
interface::tagged { 'eth1.1019':
base_interface => 'eth1',
vlan_id => '1019',
address => $ips['private1-c-eqiad'][$::hostname],
netmask => '255.255.252.0',
}
}
/^lvs101[012]$/: {
# Row C subnets on eth0
interface::tagged { 'eth0.1003':
base_interface => 'eth0',
vlan_id => '1003',
address => $ips['public1-c-eqiad'][$::hostname],
netmask => '255.255.255.192',
}
# Row A subnets on eth1
interface::tagged { 'eth1.1001':
base_interface => 'eth1',
vlan_id => '1001',
address => $ips['public1-a-eqiad'][$::hostname],
netmask => '255.255.255.192',
}
interface::tagged { 'eth1.1017':
base_interface => 'eth1',
vlan_id => '1017',
address => $ips['private1-a-eqiad'][$::hostname],
netmask => '255.255.252.0',
}
}
}
# Row B subnets on eth2
interface::tagged { 'eth2.1002':
base_interface => 'eth2',
vlan_id => '1002',
address => $ips['public1-b-eqiad'][$::hostname],
netmask => '255.255.255.192',
}
interface::tagged { 'eth2.1018':
base_interface => 'eth2',
vlan_id => '1018',
address => $ips['private1-b-eqiad'][$::hostname],
netmask => '255.255.252.0',
}
# Row D subnets on eth3
interface::tagged { 'eth3.1004':
base_interface => 'eth3',
vlan_id => '1004',
address => $ips['public1-d-eqiad'][$::hostname],
netmask => '255.255.255.224',
}
interface::tagged { 'eth3.1020':
base_interface => 'eth3',
vlan_id => '1020',
address => $ips['private1-d-eqiad'][$::hostname],
netmask => '255.255.252.0',
}
lvs::interface_tweaks {
'eth0': bnx2x => true, txqlen => 10000;
'eth1': bnx2x => true, txqlen => 10000;
'eth2': bnx2x => true, txqlen => 10000;
'eth3': bnx2x => true, txqlen => 10000;
}
}
node /^lvs101[12]\.eqiad\.wmnet$/ {
role(spare::system)
}
# codfw lvs
node /lvs200[1-6]\.codfw\.wmnet/ {
# lvs200[25] are LVS balancers for the codfw recursive DNS IP,
# so they need to use the recursive DNS backends directly
# (acamar and achernar) with fallback to eqiad
# (doing this for all lvs for now, see T103921)
$nameservers_override = [ '208.80.153.12', '208.80.153.42', '208.80.154.254' ]
role(lvs::balancer)
interface::add_ip6_mapped { 'main': }
include ::lvs::configuration
$ips = $lvs::configuration::subnet_ips
# Set up tagged interfaces to all subnets with real servers in them
# lint:ignore:case_without_default
case $::hostname {
# lint:endignore
/^lvs200[1-3]$/: {
# Row A subnets on eth0
interface::tagged { 'eth0.2001':
base_interface => 'eth0',
vlan_id => '2001',
address => $ips['public1-a-codfw'][$::hostname],
netmask => '255.255.255.224',
}
# Row B subnets on eth1
interface::tagged { 'eth1.2002':
base_interface => 'eth1',
vlan_id => '2002',
address => $ips['public1-b-codfw'][$::hostname],
netmask => '255.255.255.224',
}
interface::tagged { 'eth1.2018':
base_interface => 'eth1',
vlan_id => '2018',
address => $ips['private1-b-codfw'][$::hostname],
netmask => '255.255.252.0',
}
}
/^lvs200[4-6]$/: {
# Row B subnets on eth0
interface::tagged { 'eth0.2002':
base_interface => 'eth0',
vlan_id => '2002',
address => $ips['public1-b-codfw'][$::hostname],
netmask => '255.255.255.224',
}
# Row A subnets on eth1
interface::tagged { 'eth1.2001':
base_interface => 'eth1',
vlan_id => '2001',
address => $ips['public1-a-codfw'][$::hostname],
netmask => '255.255.255.224',
}
interface::tagged { 'eth1.2017':
base_interface => 'eth1',
vlan_id => '2017',
address => $ips['private1-a-codfw'][$::hostname],
netmask => '255.255.252.0',
}
}
}
# Row C subnets on eth2
interface::tagged { 'eth2.2003':
base_interface => 'eth2',
vlan_id => '2003',
address => $ips['public1-c-codfw'][$::hostname],
netmask => '255.255.255.224',
}
interface::tagged { 'eth2.2019':
base_interface => 'eth2',
vlan_id => '2019',
address => $ips['private1-c-codfw'][$::hostname],
netmask => '255.255.252.0',
}
# Row D subnets on eth3
interface::tagged { 'eth3.2004':
base_interface => 'eth3',
vlan_id => '2004',
address => $ips['public1-d-codfw'][$::hostname],
netmask => '255.255.255.224',
}
interface::tagged { 'eth3.2020':
base_interface => 'eth3',
vlan_id => '2020',
address => $ips['private1-d-codfw'][$::hostname],
netmask => '255.255.252.0',
}
lvs::interface_tweaks {
'eth0': bnx2x => true, txqlen => 10000;
'eth1': bnx2x => true, txqlen => 10000;
'eth2': bnx2x => true, txqlen => 10000;
'eth3': bnx2x => true, txqlen => 10000;
}
}
# ESAMS lvs servers
node /^lvs300[1-4]\.esams\.wmnet$/ {
# lvs300[24] are LVS balancers for the esams recursive DNS IP,
# so they need to use the recursive DNS backends directly
# (nescio and maerlant) with fallback to eqiad
# (doing this for all lvs for now, see T103921)
$nameservers_override = [ '91.198.174.106', '91.198.174.122', '208.80.154.254' ]
role(lvs::balancer)
interface::add_ip6_mapped { 'main': }
include ::lvs::configuration
$ips = $lvs::configuration::subnet_ips
interface::tagged { 'eth0.100':
base_interface => 'eth0',
vlan_id => '100',
address => $ips['public1-esams'][$::hostname],
netmask => '255.255.255.128',
}
# txqueuelen 20K for 10Gbps LVS in esams:
# Higher traffic than ulsfo. There is no perfect value based
# on hardware alone, but this seems to get rid of common
# spiky drops currently in esams. The real answer is
# probably a red or codel variant within each multiqueue
# class, but we need a much newer kernel + driver to
# be able to do that (both to get good schedulers
# and driver updates for XPS).
lvs::interface_tweaks {
'eth0': bnx2x => true, txqlen => 20000;
}
}
# ULSFO lvs servers
node /^lvs400[1-4]\.ulsfo\.wmnet$/ {
# ns override for all lvs for now, see T103921
$nameservers_override = [ '208.80.153.12', '208.80.153.42', '208.80.154.254' ]
role(lvs::balancer)
interface::add_ip6_mapped { 'main': }
lvs::interface_tweaks {
'eth0': bnx2x => true, txqlen => 10000;
}
}
node 'maerlant.wikimedia.org' {
role(dnsrecursor, ntp)
interface::add_ip6_mapped { 'main': }
}
node 'maps-test2001.codfw.wmnet' {
role(maps::server, maps::master)
}
node /^maps-test200[2-4]\.codfw\.wmnet/ {
role(maps::server, maps::slave)
}
node 'maps1001.eqiad.wmnet' {
role(maps::server, maps::master)
}
node /^maps100[2-4]\.eqiad\.wmnet/ {
role(maps::server, maps::slave)
}
node 'maps2001.codfw.wmnet' {
role(maps::server, maps::master)
}
node /^maps200[2-4]\.codfw\.wmnet/ {
role(maps::server, maps::slave)
}
node /^mc10(0[1-9]|1[0-8])\.eqiad\.wmnet/ {
role(spare::system)
}
node /^mc10(19|2[0-9]|3[0-6])\.eqiad\.wmnet/ {
role(mediawiki::memcached)
}
node /^mc20(19|2[0-9]|3[0-6])\.codfw\.wmnet/ {
role(mediawiki::memcached)
}
# archiva.wikimedia.org
node 'meitnerium.wikimedia.org' {
role(archiva)
include ::standard
}
# OTRS - ticket.wikimedia.org
node 'mendelevium.eqiad.wmnet' {
role(otrs::webserver)
}
# misc. test server, keep (T156208)
node 'multatuli.wikimedia.org' {
role(test::system)
interface::add_ip6_mapped { 'main': }
}
# also see dataset1001
node 'ms1001.wikimedia.org' {
role(dataset::secondary, dumps::server)
interface::add_ip6_mapped { 'main': }
}
node 'ms1002.eqiad.wmnet' {
include ::standard
}
node /^ms-fe1005\.eqiad\.wmnet$/ {
role(swift::proxy, swift::stats_reporter)
include ::lvs::realserver
}
node /^ms-fe100[6-8]\.eqiad\.wmnet$/ {
role(swift::proxy)
include ::lvs::realserver
}
node /^ms-be10(0[0-9]|1[0-5])\.eqiad\.wmnet$/ {
role(swift::storage)
}
# HP machines have different disk ordering T90922
node /^ms-be10(1[6-9]|2[0-9]|3[0-9])\.eqiad\.wmnet$/ {
role(swift::storage)
}
node /^ms-fe300[1-2]\.esams\.wmnet$/ {
role(swift::proxy)
}
node /^ms-be300[1-4]\.esams\.wmnet$/ {
role(swift::storage)
}
node /^ms-fe2005\.codfw\.wmnet$/ {
role(swift::proxy, swift::stats_reporter)
include ::lvs::realserver
}
node /^ms-fe200[6-8]\.codfw\.wmnet$/ {
role(swift::proxy)
include ::lvs::realserver
}
node /^ms-be201[3-5]\.codfw\.wmnet$/ {
role(swift::storage)
}
node /^ms-be20(0[1-9]|1[0-2])\.codfw\.wmnet$/ {
role(spare::system)
}
# HP machines have different disk ordering T90922
node /^ms-be20(1[6-9]|2[0-9]|3[0-9])\.codfw\.wmnet$/ {
role(swift::storage)
}
# mwdebug servers are for mediawiki testing/debugging
# They replace mw1017 and mw1099
node /^mwdebug100[12]\.eqiad\.wmnet$/ {
role(mediawiki::canary_appserver)
include ::base::firewall
}
# mw1161-1167 are job runners
node /^mw116[1-7]\.eqiad\.wmnet$/ {
role(mediawiki::jobrunner)
include ::base::firewall
}
# mw1170-1188 are apaches
node /^mw11(7[0-9]|8[0-8])\.eqiad\.wmnet$/ {
role(mediawiki::appserver)
include ::base::firewall
}
# mw1189-1208 are api apaches
node /^mw1(189|19[0-9]|20[0-8])\.eqiad\.wmnet$/ {
role(mediawiki::appserver::api)
include ::base::firewall
}
# mw1209-1216, 1218-1220 are apaches
node /^mw12(09|1[012345689]|20)\.eqiad\.wmnet$/ {
role(mediawiki::appserver)
include ::base::firewall
}
#mw1221-mw1235 are api apaches
node /^mw12(2[1-9]|3[0-5])\.eqiad\.wmnet$/ {
role(mediawiki::appserver::api)
include ::base::firewall
}
#mw1236-mw1258 are apaches
node /^mw12(3[6-9]|4[0-9]|5[0-8])\.eqiad\.wmnet$/ {
role(mediawiki::appserver)
include ::base::firewall
}
#mw1259-60 are videoscalers
#mw1168-69 were previously jobrunners (T153488)
node /^mw1(16[89]|259|260)\.eqiad\.wmnet/ {
role(mediawiki::videoscaler)
}
# ROW A eqiad appservers
#mw1261 - mw1275
node /^mw126[1-5]\.eqiad\.wmnet$/ {
role(mediawiki::canary_appserver)
include ::base::firewall
}
node /^mw12(6[6-9]|7[0-5])\.eqiad\.wmnet$/ {
role(mediawiki::appserver)
include ::base::firewall
}
# ROW A eqiad api appserver
# mw1276 - mw1290
node /^mw127[6-9]\.eqiad\.wmnet$/ {
role(mediawiki::appserver::canary_api)
include ::base::firewall
}
node /^mw12(8[0-9]|90)\.eqiad\.wmnet$/ {
role(mediawiki::appserver::api)
include ::base::firewall
}
# ROW A eqiad imagescalers
node /^mw129[3-8]\.eqiad\.wmnet$/ {
role(mediawiki::imagescaler)
}
# ROW A eqiad jobrunners
node /^mw1(299|30[0-6])\.eqiad\.wmnet$/ {
role(mediawiki::jobrunner)
include ::base::firewall
}
# ROW A codfw appservers: mw2017, mw2075-mw2079, and mw2215-2250
# mw2017/mw2099 are codfw test appservers
node /^mw20(17|99)\.codfw\.wmnet$/ {
role(mediawiki::canary_appserver)
include ::base::firewall
}
#mw2097, mw2100-mw2117 are appservers
node /^mw2(097|10[0-9]|11[0-7])\.codfw\.wmnet$/ {
role(mediawiki::appserver)
include ::base::firewall
}
#mw2120-2147 are api appservers
node /^mw21([2-3][0-9]|4[0-7])\.codfw\.wmnet$/ {
role(mediawiki::appserver::api)
include ::base::firewall
}
# ROW B codfw appservers
node /^mw211[89]\.codfw\.wmnet$/ {
role(mediawiki::videoscaler)
}
# ROW C codfw appservers: mw2148-mw2234
#mw2148-mw2151 are imagescalers
node /^mw21(4[89]|5[01])\.codfw\.wmnet$/ {
role(mediawiki::imagescaler)
}
#mw2152 is a videoscaler
node 'mw2152.codfw.wmnet' {
role(mediawiki::videoscaler)
}
#mw2153-62 are jobrunners
node /^mw21(5[3-9]|6[0-2])\.codfw\.wmnet$/ {
role(mediawiki::jobrunner)
include ::base::firewall
}
#mw2163-mw2199 are appservers
node /^mw21(6[3-9]|[6-9][0-9])\.codfw\.wmnet$/ {
role(mediawiki::appserver)
include ::base::firewall
}
#mw2200-2214 are api appservers
node /^mw22(0[0-9]|1[0-4])\.codfw\.wmnet$/ {
role(mediawiki::appserver::api)
include ::base::firewall
}
# New Appservers, in row A3/A4
#mw2215-2223 are api appservers
node /^mw22(1[5-9]|2[0123])\.codfw\.wmnet$/ {
role(mediawiki::appserver::api)
include ::base::firewall
}
# mw2224-42 are appservers
node /^mw22(2[4-9]|3[0-9]|4[0-2])\.codfw\.wmnet$/ {
role(mediawiki::appserver)
include ::base::firewall
}
#mw2244-mw2245 are imagescalers
node /^mw224[45]\.codfw\.wmnet$/ {
role(mediawiki::imagescaler)
}
# mw2246 is a videoscaler
node 'mw2246.codfw.wmnet' {
role(mediawiki::videoscaler)
}
# mw2247-2250 are jobrunners
node /^mw22(4[3789]|50)\.codfw\.wmnet$/ {
role(mediawiki::jobrunner)
include ::base::firewall
}
#mw2251-2253 are api-appservers
node /^mw225[1-3]\.codfw\.wmnet$/ {
role(mediawiki::appserver::api)
include ::base::firewall
}
#mw2254-2260 are appservers
node /^mw22(5[4-9]|60)\.codfw\.wmnet$/ {
role(mediawiki::appserver)
include ::base::firewall
}
# mw logging host codfw
node 'mwlog2001.codfw.wmnet' {
role(xenon)
include ::base::firewall
include ::standard
class { 'role::logging::mediawiki::udp2log':
monitor => false,
}
}
# mw logging host eqiad
node 'mwlog1001.eqiad.wmnet' {
role(xenon)
include ::base::firewall
include ::standard
class { 'role::logging::mediawiki::udp2log':
monitor => false,
}
}
node 'mx1001.wikimedia.org' {
role(mail::mx)
include ::standard
interface::add_ip6_mapped { 'main': }
interface::alias { 'wiki-mail-eqiad.wikimedia.org':
ipv4 => '208.80.154.91',
ipv6 => '2620:0:861:3:208:80:154:91',
}
}
node 'mx2001.wikimedia.org' {
role(mail::mx)
include ::standard
interface::add_ip6_mapped { 'main': }
interface::alias { 'wiki-mail-codfw.wikimedia.org':
ipv4 => '208.80.153.46',
ipv6 => '2620:0:860:2:208:80:153:46',
}
}
# Experimental Jupyter notebook servers
node 'notebook1001.eqiad.wmnet' {
# Equivalent of stat1003
role(paws_internal::jupyterhub, paws_internal::mysql_access)
include ::standard
}
node 'notebook1002.eqiad.wmnet' {
# Equivalent to stat1002
role(paws_internal::jupyterhub, analytics_cluster::client)
include ::standard
}
# cluster management (salt master, cumin master) + other management tools
node 'neodymium.eqiad.wmnet' {
role(deployment::salt_masters, debdeploy::master, ipmi::mgmt,
access_new_install, mgmt::drac_ilo, cluster::management)
interface::add_ip6_mapped { 'main': }
}
node 'nescio.wikimedia.org' {
role(dnsrecursor, ntp)
interface::add_ip6_mapped { 'main': }
}
# network monitoring tool server
node 'netmon1001.wikimedia.org' {
role(rancid::server, librenms, servermon::wmf, smokeping,
network::monitor)
include ::standard
include ::passwords::network
include ::base::firewall
interface::add_ip6_mapped { 'main': }
class { '::ganglia::monitor::aggregator':
sites => ['eqiad', 'codfw'],
}
}
# network monitoring tool server - replacement server (T125020)
node 'netmon1002.wikimedia.org' {
# role(rancid::server, librenms, servermon::wmf, torrus, smokeping)
include ::standard
include ::passwords::network
include ::base::firewall
interface::add_ip6_mapped { 'main': }
}
node /^(nihal\.codfw|nitrogen\.eqiad)\.wmnet$/ {
role(puppetmaster::puppetdb)
}
# Offline Content Generator
node /^ocg100[123]\.eqiad\.wmnet$/ {
role(ocg)
}
node /^oresrdb100[12]\.eqiad\.wmnet$/ {
role(ores::redis)
include ::standard
}
node /^oresrdb200[12]\.codfw\.wmnet$/ {
role(ores::redis)
include ::standard
}
# VisualEditor performance testing rig
node 'osmium.eqiad.wmnet' {
role(ve)
include ::standard
include ::base::firewall
}
# oxygen runs a kafkatee instance that consumes webrequest from Kafka
# and writes to a couple of files for quick and easy ops debugging.,
node 'oxygen.eqiad.wmnet'
{
role(logging::kafkatee::webrequest::ops)
include ::base::firewall
include ::standard
}
# parser cache databases
# eqiad
node 'pc1004.eqiad.wmnet' {
class { '::role::mariadb::parsercache':
shard => 'pc1',
}
include ::base::firewall
}
node 'pc1005.eqiad.wmnet' {
class { '::role::mariadb::parsercache':
shard => 'pc2',
}
include ::base::firewall
}
node 'pc1006.eqiad.wmnet' {
class { '::role::mariadb::parsercache':
shard => 'pc3',
}
include ::base::firewall
}
# codfw
node 'pc2004.codfw.wmnet' {
class { '::role::mariadb::parsercache':
shard => 'pc1',
}
include ::base::firewall
}
node 'pc2005.codfw.wmnet' {
class { '::role::mariadb::parsercache':
shard => 'pc2',
}
include ::base::firewall
}
node 'pc2006.codfw.wmnet' {
class { '::role::mariadb::parsercache':
shard => 'pc3',
}
include ::base::firewall
}
# virtual machines hosting https://wikitech.wikimedia.org/wiki/Planet.wikimedia.org
node /^planet[12]001\.(eqiad|codfw)\.wmnet$/ {
role(planet_server)
interface::add_ip6_mapped { 'main': }
}
# LDAP servers relied on by OIT for mail
node /(dubnium|pollux)\.wikimedia\.org/ {
role(openldap::corp)
include ::standard
}
node 'poolcounter1001.eqiad.wmnet' {
role(poolcounter::server)
}
node 'poolcounter1002.eqiad.wmnet' {
role(poolcounter::server)
}
node /^prometheus200[34]\.codfw\.wmnet$/ {
role(prometheus::ops, prometheus::global)
include ::base::firewall
include ::standard
include ::lvs::realserver
interface::add_ip6_mapped { 'main': }
}
node /^prometheus100[34]\.eqiad\.wmnet$/ {
role(prometheus::ops, prometheus::global)
include ::lvs::realserver
interface::add_ip6_mapped { 'main': }
}
node /^puppetmaster[12]001\.(codfw|eqiad)\.wmnet$/ {
role(
ipmi::mgmt, access_new_install,
puppetmaster::frontend,
)
include ::standard
interface::add_ip6_mapped { 'main': }
}
node /^puppetmaster[12]002\.(codfw|eqiad)\.wmnet$/ {
role(puppetmaster::backend)
interface::add_ip6_mapped { 'main': }
}
# pybal-test200X VMs are used for pybal testing/development
node /^pybal-test200[123]\.codfw\.wmnet$/ {
role(pybaltest)
include ::standard
}
# Tor relay
node 'radium.wikimedia.org' {
role(tor_relay)
interface::add_ip6_mapped { 'main': }
}
node 'radon.wikimedia.org' {
role(authdns::server)
interface::add_ip6_mapped { 'main': }
}
# Live Recent Changes WebSocket stream
node 'rcs1001.eqiad.wmnet', 'rcs1002.eqiad.wmnet' {
interface::add_ip6_mapped { 'main': }
role(rcstream)
include ::base::firewall
}
node /^rdb100[1357]\.eqiad\.wmnet$/ {
role(jobqueue_redis::master)
}
node /^rdb100[24689]\.eqiad\.wmnet/ {
role(jobqueue_redis::slave)
}
node /^rdb200[135]\.codfw\.wmnet$/ {
role(jobqueue_redis::master)
}
node /^rdb200[246]\.codfw\.wmnet/ {
role(jobqueue_redis::slave)
}
node /^relforge100[1-2]\.eqiad\.wmnet/ {
role(elasticsearch::relforge)
}
# restbase eqiad cluster
node /^restbase10[01][0-9]\.eqiad\.wmnet$/ {
role(restbase::production)
}
# restbase codfw cluster
node /^restbase20[01][0-9]\.codfw\.wmnet$/ {
role(restbase::production)
}
# cassandra multi-dc temporary test T111382
node /^restbase-test200[1-3]\.codfw\.wmnet$/ {
role(restbase::test_cluster)
}
# cassandra/restbase dev cluster
node /^restbase-dev100[1-3]\.eqiad\.wmnet$/ {
role(restbase::dev_cluster)
}
# network insights (netflow/pmacct, etc.)
node 'rhenium.wikimedia.org' {
role(pmacct)
}
# Failoid service (Ganeti VM)
node 'roentgenium.eqiad.wmnet' {
role(failoid)
}
# people.wikimedia.org, for all shell users
node 'rutherfordium.eqiad.wmnet' {
role(microsites::peopleweb)
}
# ruthenium is a parsoid regression test server
# https://www.mediawiki.org/wiki/Parsoid/Round-trip_testing
# Right now, both rt-server and rt-clients run on the same node
# But, we are likely going to split them into different boxes soon.
node 'ruthenium.eqiad.wmnet' {
role(test::system,
parsoid::testing, parsoid::rt_server, parsoid::rt_client,
parsoid::vd_server, parsoid::vd_client, parsoid::diffserver)
}
# cluster management (salt master, cumin master)
node 'sarin.codfw.wmnet' {
role(cluster::management)
interface::add_ip6_mapped { 'main': }
}
# Services 'A'
node /^sca[12]00[1234]\.(eqiad|codfw)\.wmnet$/ {
role(sca)
interface::add_ip6_mapped { 'main': }
}
# Services 'B'
node /^scb[12]00[123456]\.(eqiad|codfw)\.wmnet$/ {
role(scb)
interface::add_ip6_mapped { 'main': }
}
# Codfw, eqiad ldap servers, aka ldap-$::site
node /^(seaborgium|serpens)\.wikimedia\.org$/ {
role(openldap::labs)
include ::standard
include ::base::firewall
}
# Silver is the new home of the wikitech web server.
node 'silver.wikimedia.org' {
role(labs::openstack::nova::manager, mariadb::wikitech)
include ::base::firewall
include ::standard
interface::add_ip6_mapped { 'main': }
}
node 'sodium.wikimedia.org' {
role(mirrors)
include ::standard
interface::add_ip6_mapped { 'main': }
}
node /^rhodium.eqiad.wmnet/ {
role(puppetmaster::backend)
interface::add_ip6_mapped { 'main': }
}
node 'thorium.eqiad.wmnet' {
# thorium is mainly used to host Analytics websites like:
# - https://stats.wikimedia.org (Wikistats)
# - https://datasets.wikimedia.org
# - https://metrics.wikimedia.org (https://metrics.wmflabs.org/ (Wikimetrics))
# - https://pivot.wikimedia.org (Imply's Pivot UI for Druid data)
# - https://hue.wikimedia.org (Hadoop User Experience GUI)
#
# For a complete and up to date list please check the
# related role/module.
#
# This node is not intended for data processing.
role(statistics::web,
analytics_cluster::druid::pivot,
analytics_cluster::hue)
include ::standard
include ::base::firewall
}
# Failoid service (Ganeti VM)
node 'tureis.codfw.wmnet' {
role(failoid)
}
node 'stat1002.eqiad.wmnet' {
# stat1002 is intended to be the private
# webrequest access log storage host.
# Users should not use it for app development.
# Data processing on this machine is fine.
# Include classes needed for storing and crunching
# private data on stat1002.
role(statistics::private,
# stat1002 is also a Hadoop client, and should
# have any special analytics system users on it
# for interacting with HDFS.
analytics_cluster::users,
# Include Hadoop and other analytics cluster
# clients so that analysts can access Hadoop
# from here.
analytics_cluster::client,
# Include analytics/refinery deployment target.
analytics_cluster::refinery,
# Include analytics/refinery checks that send email about
# webrequest partitions faultyness.
analytics_cluster::refinery::job::data_check,
# Include analytics/refinery/source guard checks
# Disabled due to T166937
# analytics_cluster::refinery::job::guard,
# Set up a read only rsync module to allow access
# to public data generated by the Analytics Cluster.
analytics_cluster::rsyncd,
# Deploy wikimedia/discovery/analytics repository
# to this node.
elasticsearch::analytics)
include ::standard
# Include the MySQL research password at
# /etc/mysql/conf.d/analytics-research-client.cnf
# and only readable by users in the
# analytics-privatedata-users group.
statistics::mysql_credentials { 'analytics-research':
group => 'analytics-privatedata-users',
}
# The eventlogging code is useful for scripting
# EventLogging consumers. Install this on
# stat1002, but don't run any daemons.
include ::eventlogging
}
# stat1003 is a general purpose number cruncher for
# researchers and analysts. It is primarily used
# to connect to MySQL research databases and save
# query results for further processing on this node.
node 'stat1003.eqiad.wmnet' {
role(statistics::cruncher)
}
node 'stat1004.eqiad.wmnet' {
# stat1004 contains all the tools and libraries to access
# the Analytics Cluster services.
role(analytics_cluster::client, analytics_cluster::refinery)
include ::standard
}
node /^snapshot1001\.eqiad\.wmnet/ {
role(snapshot::testbed)
include ::standard
}
node /^snapshot100[5-7]\.eqiad\.wmnet/ {
# NOTE: New snapshot hosts must also be manually added
# to hiera/common.yaml:dataset_clients_snapshots,
# hieradata/hosts/ with a file named after the host,
# and modules/scap/files/dsh/group/mediawiki-installation
role(snapshot::dumper, snapshot::monitor, snapshot::cronrunner)
}
# codfw poolcounters
node /(subra|suhail)\.codfw\.wmnet/ {
role(poolcounter::server)
}
# mediawiki maintenance servers (https://wikitech.wikimedia.org/wiki/Terbium)
node 'terbium.eqiad.wmnet', 'wasat.codfw.wmnet' {
role(mediawiki_maintenance)
interface::add_ip6_mapped { 'main': }
}
# Thumbor servers for MediaWiki image scaling
node /^thumbor100[12].eqiad.wmnet/ {
role(thumbor::mediawiki)
}
# deployment servers
node 'tin.eqiad.wmnet', 'naos.codfw.wmnet' {
role(deployment_server)
interface::add_ip6_mapped { 'main': }
}
# test system for performance team (T117888)
node 'tungsten.eqiad.wmnet' {
role(test::system, xhgui::app)
}
# replaced magnesium (RT) (T119112 T123713)
node 'ununpentium.wikimedia.org' {
role(requesttracker_server)
interface::add_ip6_mapped { 'main': }
}
# Ganglia Web UI
node 'uranium.wikimedia.org' {
role(ganglia::web)
interface::add_ip6_mapped { 'main': }
}
node /^labvirt100[0-9].eqiad.wmnet/ {
openstack::nova::partition{ '/dev/sdb': }
role(labs::openstack::nova::compute)
include ::standard
}
node /^labvirt101[0-4].eqiad.wmnet/ {
role(labs::openstack::nova::compute)
include ::standard
}
# Wikidata query service
node /^wdqs100[1-3]\.eqiad\.wmnet$/ {
role(wdqs)
include ::lvs::realserver
}
node /^wdqs200[1-3]\.codfw\.wmnet$/ {
role(wdqs)
include ::lvs::realserver
}
node 'wezen.codfw.wmnet' {
role(syslog::centralserver)
}
# https://www.mediawiki.org/wiki/Parsoid
node /^wtp10(0[1-9]|1[0-9]|2[0-4])\.eqiad\.wmnet$/ {
role(parsoid)
}
node /^wtp20(0[1-9]|1[0-9]|2[0-4])\.codfw\.wmnet$/ {
role(parsoid)
}
# T138650 - tools for the security team
node 'zosma.codfw.wmnet' {
role(security::tools)
interface::add_ip6_mapped { 'main': }
}
node default {
if $::realm == 'production' {
include ::standard
} else {
# Require instead of include so we get NFS and other
# base things setup properly
require ::role::labs::instance
}
}
![swh spinner](/static/img/swh-spinner.gif)
Computing file changes ...