Puppet Class: profile::presto::server

Defined in:
modules/profile/manifests/presto/server.pp

Overview

SPDX-License-Identifier: Apache-2.0

Class profile::presto::server

Sets up a presto server in a presto cluster. By default this node will be set up as a worker node. To enable the coordinator or discovery, provide appropriate settings in $config_properties.

See also: https://prestodb.io/docs/current/installation/deployment.html

Parameters

[cluster_name] Name of the Presto cluster. This will be used as the default node.environment.

[discovery_uri] URI to the Presto discovery server.

[enabled] Allows selectively disabling nodes within the cluster. Default: true

[node_properties] Specific node.properties settings. This profile attempts to use sane defaults. Only set this if you need to override them. Note that node.id will be set automatically by the presto::server module class based on the current node's $::fqdn.

[config_properties] Specific config.properties settings. This profile attempts to use sane defaults. Only set this if you need to override them.

[resource_groups_enabled] Enable the file-based resource groups manager on this cluster. Default: false

[resource_config] Hash of overrides for the resource groups limits/weights rendered into resource-groups.json. Merged over $default_resource_config. Only used when $resource_groups_enabled is true.

[spill_enabled] Create the on-disk spill directory and add the configuration file property.

[log_properties] Specific log.properties settings.

[heap_max] -Xmx argument. Default: 2G

[generate_certificate] Use the pki::get_cert function to generate a certificate for TLS authentication, using the discovery CA, as opposed to using the Puppet CA certificates. If this option is enabled, a certificate will be generated using the host's $::fqdn. Default: false

[ssl_certnames] Optional array of SubjectAltNames that should be added to a certificate generated by the pki::get_cert function. This is only used if generate_certificates is true. Default: []

[presto_clusters_secrets] Hash of available/configured Presto clusters and their secret properties, like passwords, etc.. The following values will be checked in the hash table only if TLS/Kerberos configs are enabled (see in the code for the exact values). - 'ssl_keystore_password' - 'ssl_trustore_password' Default: {}

Parameters:

  • cluster_name (String) (defaults to: lookup('profile::presto::cluster_name'))
  • discovery_uri (String) (defaults to: lookup('profile::presto::discovery_uri'))
  • enabled (Boolean) (defaults to: lookup('profile::presto::enabled', { 'default_value' => true }))
  • node_properties (Hash) (defaults to: lookup('profile::presto::server::node_properties', { 'default_value' => {} }))
  • config_properties (Hash) (defaults to: lookup('profile::presto::server::config_properties', { 'default_value' => {} }))
  • resource_groups_enabled (Boolean) (defaults to: lookup('profile::presto::server::resource_groups_enabled', { 'default_value' => false }))
  • resource_config (Hash) (defaults to: lookup('profile::presto::server::resource_config', { 'default_value' => {} }))
  • spill_enabled (Boolean) (defaults to: lookup('profile::presto::server::spill_enabled', { 'default_value' => false }))
  • catalogs (Hash) (defaults to: lookup('profile::presto::server::catalogs', { 'default_value' => {} }))
  • log_properties (Hash) (defaults to: lookup('profile::presto::server::log_properties', { 'default_value' => {} }))
  • heap_max (String) (defaults to: lookup('profile::presto::server::heap_max', { 'default_value' => '2G' }))
  • ssl_certnames (Array[String]) (defaults to: lookup('profile::presto::server::ssl_certnames', { 'default_value' => [] }))
  • generate_certificate (Boolean) (defaults to: lookup('profile::presto::server::generate_certificate', { 'default_value' => false }))
  • use_kerberos (Boolean) (defaults to: lookup('profile::presto::use_kerberos', { 'default_value' => true }))
  • monitoring_enabled (Boolean) (defaults to: lookup('profile::presto::monitoring_enabled', { 'default_value' => false }))
  • firewall_access (Array[String]) (defaults to: lookup('profile::presto::server::firewall_access'))
  • presto_clusters_secrets (Optional[Hash[String, Hash[String, String]]]) (defaults to: lookup('presto_clusters_secrets', { 'default_value' => {} }))


68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
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
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
# File 'modules/profile/manifests/presto/server.pp', line 68

class profile::presto::server (
    String        $cluster_name              = lookup('profile::presto::cluster_name'),
    String        $discovery_uri             = lookup('profile::presto::discovery_uri'),
    Boolean       $enabled                   = lookup('profile::presto::enabled', { 'default_value' => true }),
    Hash          $node_properties           = lookup('profile::presto::server::node_properties', { 'default_value' => {} }),
    Hash          $config_properties         = lookup('profile::presto::server::config_properties', { 'default_value' => {} }),
    Boolean       $resource_groups_enabled   = lookup('profile::presto::server::resource_groups_enabled', { 'default_value' => false }),
    Hash          $resource_config           = lookup('profile::presto::server::resource_config', { 'default_value' => {} }),
    Boolean       $spill_enabled             = lookup('profile::presto::server::spill_enabled', { 'default_value' => false }),
    Hash          $catalogs                  = lookup('profile::presto::server::catalogs', { 'default_value' => {} }),
    Hash          $log_properties            = lookup('profile::presto::server::log_properties', { 'default_value' => {} }),
    String        $heap_max                  = lookup('profile::presto::server::heap_max', { 'default_value' => '2G' }),
    Array[String] $ssl_certnames             = lookup('profile::presto::server::ssl_certnames', { 'default_value' => [] }),
    Boolean       $generate_certificate      = lookup('profile::presto::server::generate_certificate', { 'default_value' => false }),
    Boolean       $use_kerberos              = lookup('profile::presto::use_kerberos', { 'default_value' => true }),
    Boolean       $monitoring_enabled        = lookup('profile::presto::monitoring_enabled', { 'default_value' => false }),
    Array[String] $firewall_access           = lookup('profile::presto::server::firewall_access'),
    Optional[Hash[String, Hash[String, String]]] $presto_clusters_secrets = lookup('presto_clusters_secrets', { 'default_value' => {} }),
) {
    # node.environment must not contain any -
    $sanitize_cluster_name = regsubst($cluster_name, '-', '', 'G')
    $default_node_properties = {
        'node.environment'              => $sanitize_cluster_name,
        'node.data-dir'                => '/srv/presto',
        'node.internal-address-source' => 'FQDN',
    }

    $default_config_properties = {
        'jmx.rmiregistry.port'               => 8279,
        'discovery.uri'                      => $discovery_uri,

        # flat will try to schedule splits on the host where the data is located by reserving
        # 50% of the work queue for local splits. It is recommended to use flat for clusters
        # where distributed storage runs on the same nodes as Presto workers.
        # You should change this if your Presto cluster is not colocated with storage.
        'node-scheduler.network-topology'    => 'flat',

        # Enables accessing the Presto UI over http. Access it as follows:
        # ssh -NL 8280:an-test-coord1001.eqiad.wmnet:8280 an-test-coord1001.eqiad.wmnet
        # (for the test cluster, for the production cluster use an-coord1003).
        # Then visit http://localhost:8280 (which should redirect to http://localhost:8280/ui/)
        'http-server.http.port'              => 8280,
        'http-server.http.enabled'           => true,
    }

    if $use_kerberos {
        $hostname_suffix = $::realm ? {
            'labs'  => '.eqiad.wmflabs',
            default => "${::site}.wmnet",
        }

        $keystore_password = $presto_clusters_secrets[$cluster_name]['ssl_keystore_password']
        $ssl_keystore_path = '/etc/presto/ssl/server.p12'
        $ssl_truststore_path = '/etc/ssl/certs/java/cacerts'
        $ssl_truststore_password = 'changeit'

        if $generate_certificate {
            # Generate a certificate for each presto server. These will be used to secure traffic between
            # 1) the clients and the co-ordinator and 2) between each server in the cluster
            # The co-ordinator will use several SubJectAltNames in order to support DNS discovery
            $ssl_cert = profile::pki::get_cert('discovery2026', $facts['networking']['fqdn'], {
                'hosts'  => $ssl_certnames,
                'outdir' => '/etc/presto/ssl',
                'owner'  => 'presto',
                notify   => Sslcert::X509_to_pkcs12['presto_keystore'],
                require  => Package['presto-server'],
                }
            )
            sslcert::x509_to_pkcs12 { 'presto_keystore' :
                owner       => 'presto',
                group       => 'presto',
                public_key  => $ssl_cert['chained'],
                private_key => $ssl_cert['key'],
                outfile     => $ssl_keystore_path,
                password    => $keystore_password,
                require     => Package['presto-server'],
                notify      => Service['presto-server'],
            }
        } else {
            # TODO: consider using profile::pki::get_cert
            puppet::expose_agent_certs { '/etc/presto':
                user         => 'root',
                group        => 'presto',
                provide_p12  => true,
                provide_pem  => false,
                p12_password => $keystore_password,
                require      => Package['presto-server'],
            }
        }

        file { '/usr/local/bin/presto':
            owner   => 'root',
            group   => 'root',
            mode    => '0555',
            content => template('profile/presto/presto_client_ssl_kerberos.erb'),
            require => Package['presto-server'],
        }
    }

    if $presto_clusters_secrets[$cluster_name] {
        $ssl_keystore_password = $presto_clusters_secrets[$cluster_name]['ssl_keystore_password']
        $default_ssl_properties = {
            'http-server.https.keystore.path' => $ssl_keystore_path,
            'http-server.https.keystore.key' => $ssl_keystore_password,
            'internal-communication.https.required' => true,
            'internal-communication.https.keystore.path' => $ssl_keystore_path,
            'internal-communication.https.keystore.key' => $ssl_keystore_password,
            'internal-communication.https.trust-store-path' => '/etc/ssl/certs/java/cacerts',
            'internal-communication.https.trust-store-password' => 'changeit',
            'http-server.https.port' => 8281,
            'http-server.https.enabled' => true,
        }
    } else {
        $default_ssl_properties = {}
    }

    if $use_kerberos {
        $default_kerberos_properties = {
            'internal-communication.kerberos.enabled' => true,
            'http-server.authentication.type' => 'KERBEROS',
            'http.authentication.krb5.config' => '/etc/krb5.conf',
            'http.server.authentication.krb5.keytab' => '/etc/security/keytabs/presto/presto.keytab',
            'http.server.authentication.krb5.service-name' => 'presto',
        }
    } else {
        $default_kerberos_properties = {}
    }

    # Sensible defaults for the resource groups manager. Only consumed when
    # $resource_groups_enabled is true; hieradata overrides individual values.
    $default_resource_config = {
        # global
        'glob_mem_limit'    => '100%',
        'glob_concur_limit' => 40,
        'glob_max_queued'   => 200,
        # high priority
        'hp_mem_limit'      => '35%',
        'hp_concur_limit'   => 12,
        'hp_max_queued'     => 40,
        'hp_sched_weight'   => 6,
        # standard
        'std_mem_limit'     => '45%',
        'std_concur_limit'  => 20,
        'std_max_queued'    => 100,
        'std_sched_weight'  => 3,
        # heavy
        'hv_mem_limit'      => '20%',
        'hv_concur_limit'   => 5,
        'hv_max_queued'     => 60,
        'hv_sched_weight'   => 1,
    }

    # Merge in any overrides for properties
    $_node_properties = $default_node_properties + $node_properties
    $_config_properties = $default_config_properties + $config_properties + $default_ssl_properties + $default_kerberos_properties
    $_resource_config = $default_resource_config + $resource_config

    if $monitoring_enabled {
        include ::profile::presto::monitoring::server
        $jmx_agent_port = $::profile::presto::monitoring::server::prometheus_jmx_exporter_server_port
        $jmx_agent_config_file = $::profile::presto::monitoring::server::jmx_exporter_config_file
        $extra_jvm_configs = "-javaagent:/usr/share/java/prometheus/jmx_prometheus_javaagent.jar=[::]:${jmx_agent_port}:${jmx_agent_config_file}"

        nrpe::monitor_service { 'presto-server':
            description    => 'Presto Server',
            nrpe_command   => '/usr/lib/nagios/plugins/check_procs -c 1:1 -C java -a "com.facebook.presto.server.PrestoServer"',
            contact_group  => 'team-data-platform',
            require        => Class['presto::server'],
            notes_url      => 'https://wikitech.wikimedia.org/wiki/Analytics/Systems/Presto/Administration#Presto_server_down',
            migration_task => 'T357099',
        }
    } else {
        $extra_jvm_configs = undef
    }

    class { '::presto::server':
        enabled                 => $enabled,
        node_properties         => $_node_properties,
        config_properties       => $_config_properties,
        resource_groups_enabled => $resource_groups_enabled,
        resource_config         => $_resource_config,
        spill_enabled           => $spill_enabled,
        log_properties          => $log_properties,
        catalogs                => $catalogs,
        heap_max                => $heap_max,
        extra_jvm_configs       => $extra_jvm_configs,
    }

    if $presto_clusters_secrets[$cluster_name] {
        firewall::service { 'presto-https':
            proto    => 'tcp',
            port     => $_config_properties['http-server.https.port'],
            src_sets => $firewall_access,
        }
    } else {
        firewall::service { 'presto-http':
            proto    => 'tcp',
            port     => $_config_properties['http-server.http.port'],
            src_sets => $firewall_access,
        }
    }

    firewall::service { 'presto-jmx-rmiregistry':
        proto    => 'tcp',
        port     => $_config_properties['jmx.rmiregistry.port'],
        src_sets => $firewall_access,
    }
}