Puppet Class: profile::hadoop::backup::namenode

Defined in:
modules/profile/manifests/hadoop/backup/namenode.pp

Overview

SPDX-License-Identifier: Apache-2.0

Class profile::hadoop::backup::namenode

Periodically backup the HDFS FSImage locally, and optionnally on HDFS. Also ensures that bacula backs up Hadoop NameNode fsimages, in the case we need to recover if both Hadoop NameNodes.

Parameters:

  • monitoring_enabled (Boolean) (defaults to: lookup('profile::hadoop::backup::namenode::monitoring_enabled', {default_value => false})) —

    Alerts in case of staling backup.

  • fsimage_retention_days (Integer) (defaults to: lookup('profile::hadoop::backup::namenode::fsimage_retention_days', {default_value => 10})) —

    Number of days during which the backups are kept.

  • hdfs_backup_dir (Optional[String]) (defaults to: lookup('profile::hadoop::backup::namenode::hdfs_backup_dir', {default_value => undef})) —

    HDFS directory to send the backup to. If present, it will add a systemd timer.



11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
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
 
# File 'modules/profile/manifests/hadoop/backup/namenode.pp', line 11

class profile::hadoop::backup::namenode(
    Boolean $monitoring_enabled       = lookup('profile::hadoop::backup::namenode::monitoring_enabled', {default_value => false}),
    Integer $fsimage_retention_days   = lookup('profile::hadoop::backup::namenode::fsimage_retention_days', {default_value => 10}),
    Optional[String] $hdfs_backup_dir = lookup('profile::hadoop::backup::namenode::hdfs_backup_dir', {default_value => undef}),
) {
    require ::profile::hadoop::common

    $backup_dir_group = $::realm ? {
        'production' => 'analytics-admins',
        'labs'       => "project-${::wmcs_project}",
    }

    if !defined(File['/srv/backup']) {
        file { '/srv/backup':
            ensure  => 'directory',
            owner   => 'root',
            group   => $backup_dir_group,
            mode    => '0755',
            require => Group[$backup_dir_group],
        }
    }

    $backup_dir = '/srv/backup/hadoop/namenode'
    file { [
            '/srv/backup/hadoop',
            $backup_dir
        ]:
        ensure  => 'directory',
        owner   => 'hdfs',
        group   => $backup_dir_group,
        mode    => '0750',
        require => File['/srv/backup']
    }

    $backup_file_path_cmd = "${backup_dir}/fsimage_$(/usr/bin/date +%%Y-%%m-%%d)"  # %% is for escaping % in systemd syntax
    kerberos::systemd_timer { 'hadoop-namenode-backup-fetchimage':
        description => 'Downloads the most recent fsimage from the NameNode and saves it in the specified local directory.',
        command     => "/usr/bin/bash -c \"/usr/bin/hdfs dfsadmin -fetchImage ${backup_file_path_cmd} && /usr/bin/gzip ${backup_file_path_cmd}\"",
        interval    => '*-*-* 00:00:00',  # everyday at midnight
        user        => 'hdfs',
    }

    if $hdfs_backup_dir != undef {

        $backup_manager_script = '/etc/hadoop/hdfs_fsimage_backup_manager_script.sh'
        file { $backup_manager_script:
            ensure => present,
            owner  => 'hdfs',
            group  => 'hdfs',
            mode   => '0750',
            source => 'puppet:///modules/profile/hadoop/hdfs_fsimage_backup_manager.sh'
        }

        kerberos::systemd_timer { 'hadoop-namenode-backup-hdfs':
            description => 'Checks that the local backup has been created, and sends it to HDFS.',
            command     => "${backup_manager_script} ${backup_dir} ${hdfs_backup_dir}",
            interval    => 'Mon *-*-* 02:00:00',  #  2am on Monday
            user        => 'hdfs'
        }
    }

    systemd::timer::job { 'hadoop-namenode-backup-prune':
      description     => "Deletes namenode's fsimage backups in ${backup_dir} older than ${fsimage_retention_days} days.",
      command         => "/usr/bin/find ${backup_dir} -mtime +${fsimage_retention_days} -delete",
      interval        => {
          'start'    => 'OnCalendar',
          'interval' => '*-*-* 04:00:00',
      },
      logging_enabled => false,
      user            => 'hdfs',
      send_mail_to    => 'data-engineering-alerts@wikimedia.org',
    }

    if !defined(Sudo::User['nagios_check_newest_file_age']) {
        sudo::user { 'nagios_check_newest_file_age':
            ensure => absent,
        }
    }

    if $monitoring_enabled {
        # Alert if backup gets stale.
        $warning_threshold_hours = 26
        $critical_threshold_hours = 48
        nrpe::monitor_service { 'hadoop-namenode-backup-age':
            description   => 'Age of most recent Hadoop NameNode backup files',
            nrpe_command  => "/usr/local/lib/nagios/plugins/check_newest_file_age -V -C -d ${backup_dir} -w ${$warning_threshold_hours} -c ${critical_threshold_hours}",
            sudo_user     => 'root',
            contact_group => 'team-data-platform',
            notes_url     => 'https://wikitech.wikimedia.org/wiki/Analytics/Systems/Cluster/Hadoop/Alerts#HDFS_Namenode_backup_age',
        }
    }

    # Bacula will also back up this directory.
    # See: bacula::director::fileset { 'hadoop-namenode-backup'
    # in profile::backup::director
    include ::profile::backup::host
    backup::set { 'hadoop-namenode-backup' : }
}