{"id":2938,"date":"2026-01-07T10:57:52","date_gmt":"2026-01-07T10:57:52","guid":{"rendered":"https:\/\/bynatree.com\/?p=2938"},"modified":"2026-01-07T10:57:52","modified_gmt":"2026-01-07T10:57:52","slug":"high-availability-postgresql-16-11-cluster-with-repmgr-5-5-on-centos-stream-10-coughlan","status":"publish","type":"post","link":"https:\/\/divaind.com\/ie1\/2026\/01\/07\/high-availability-postgresql-16-11-cluster-with-repmgr-5-5-on-centos-stream-10-coughlan\/","title":{"rendered":"High Availability PostgreSQL 16.11 Cluster with repmgr 5.5 on CentOS Stream 10 (Coughlan)"},"content":{"rendered":"

Introduction<\/span><\/h2>\n

High availability (HA) is a critical requirement for modern database infrastructures, ensuring minimal downtime and seamless failover in case of node failures. PostgreSQL, while robust as a standalone database, requires additional tooling to manage replication, failover, and cluster orchestration in multi\u2011node environments.<\/span><\/p><\/blockquote>\n

repmgr (Replication Manager) is a widely adopted open\u2011source tool that simplifies the administration of PostgreSQL replication and failover. Version 5.5 of repmgr introduces enhanced support for PostgreSQL 16.x, making it a reliable choice for production\u2011grade HA setups.<\/span><\/p>\n

This guide focuses on deploying PostgreSQL 16.11 with repmgr 5.5 on CentOS Stream 10 (Coughlan), the rolling\u2011release distribution aligned with Red Hat Enterprise Linux 10. The target architecture is a three\u2011node cluster consisting of one primary and two standby nodes, configured for synchronous replication and automatic failover<\/span><\/p>\n

Cluster Topology:<\/b> Three nodes<\/span><\/p>\n

    \n
  • Node 1 (Primary):<\/b> Handles read\/write traffic and replicates changes downstream.<\/span><\/li>\n
  • Node 2 (Standby):<\/b> Hot standby, ready to be promoted in case of primary failure.<\/span><\/li>\n
  • Node 3 (Witness):<\/b> Provides quorum and ensures consistent failover decisions.<\/span><\/li>\n<\/ul>\n

    On a host:<\/span><\/h2>\n
    192.168.134.129\u00a0 node1 \u00a0 pg-primary\n\n192.168.134.130\u00a0 node2 \u00a0 pg-standby\n\n192.168.134.131\u00a0 witness pg-witness<\/pre>\n
    Checking OS Version:<\/span><\/h2>\n
    Command:
    \n<\/b><\/p>\n
    cat \/etc\/os-release<\/pre>\n
    Output:<\/b><\/p>\n\n\n\n
    \n
    NAME=\"CentOS Stream\"\n\nVERSION=\"10 (Coughlan)\"\n\n\u2026\u2026\u2026\u2026\u2026\u2026\u2026\n\n\u2026\u2026\u2026\u2026\u2026\u2026<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
    Installing Prerequisites for the repmgr:<\/span><\/h3>\n
    wget https:\/\/github.com\/EnterpriseDB\/repmgr\/releases\/download\/v5.5.0\/repmgr-5.5.0.tar.gz\nsudo yum install -y gcc make readline-devel zlib-devel \u00a0 \u00a0 postgresql postgresql-server postgresql-contrib\nsudo yum install -y perl-App-cpanminus\nsudo cpanm IPC::Run\nsudo dnf install -y epel-release\nsudo dnf config-manager --set-enabled crb\nsudo dnf install -y perl-IPC-Run\nsudo yum install -y postgresql16-devel\nsudo yum install -y libcurl-devel json-c-devel\nsudo yum install -y flex\nsudo yum install -y lz4-devel libxslt-devel libxml2-devel pam-devel<\/pre>\n
    Untar repmgr package:<\/span><\/h3>\n\n\n\n
    \n
    tar -xvf repmgr-5.5.0.tar.gz\ncd repmgr-5.5.0\nwhich pg_config<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
    Output:<\/span><\/p>\n
    \/usr\/pgsql-16\/bin\/pg_config<\/pre>\n
    Installing\u00a0 repmgr in pg_config path:<\/span><\/h3>\n\n\n\n
    \n
    export PG_CONFIG=\/usr\/pgsql-16\/bin\/pg_config\n.\/configure\nmake\nsudo make install<\/pre>\n
    Added in bash_profile<\/b><\/p>\n
    export PATH=\/usr\/pgsql-16\/bin:$PATH\necho 'export PATH=\/usr\/pgsql-16\/bin:$PATH' >> ~\/.bash_profile\nsource ~\/.bash_profile<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
     <\/p>\n\n\n\n
    \n
    [postgres@test3 data]$ repmgr --version\nrepmgr 5.5.0\n\n[postgres@test3 data]$ repmgrd --version\nrepmgrd 5.5.0<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
    \"\"<\/p>\n
    Prerequisites:<\/span><\/h2>\n
    Enable ssh in all the three machine:<\/span><\/h3>\n\n\n\n
    \n
    ssh-keygen -t ed25519\n\nssh-copy-id postgres@192.168.134.129\nssh-copy-id postgres@192.168.134.130\n\nssh postgres@192.168.134.130 \"hostname\"\nssh postgres@192.168.134.131 \"hostname\"<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
     <\/p>\n
    Added entry for allowing port 5432 in tcp\u00a0<\/span><\/h3>\n
    sudo firewall-cmd --add-port=5432\/tcp --permanent\nsudo firewall-cmd --reload<\/pre>\n
    \"\"<\/span><\/p>\n
    Verify:<\/span><\/p>\n
    \"\"
    \nAdding host in all the three node in \/etc\/hosts :<\/span><\/p>\n
    Vi \/etc\/hosts<\/pre>\n
    Repmgr Setup:<\/span><\/h2>\n
    Creating user in the primary and witness node:<\/span>
    \n<\/span><\/h3>\n\n\n\n
    \n
    CREATE USER repmgr WITH REPLICATION SUPERUSER LOGIN;\u00a0\nALTER USER repmgr SET search_path TO repmgr, public;\u00a0\u00a0\nCREATE DATABASE repmgr OWNER repmgr;<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
    In postgresql.conf add following configuration changes in primary (node1) and (node2) standby :<\/span><\/h4>\n\n\n\n
    \n
    listen_addresses = '*'\u00a0\nport = 5432\u00a0\nmax_wal_senders = 10\u00a0\nmax_replication_slots = 10\u00a0\nwal_level = replica\u00a0\nhot_standby = on\u00a0\narchive_mode = on\u00a0\narchive_command = '\/bin\/true'\u00a0 # or your preferred archiving command\u00a0\nwal_log_hints = on\u00a0\nshared_preload_libraries = 'repmgr'<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
     <\/p>\n
    Added entry for the repmgr in pg_hba.conf<\/span><\/h4>\n
    # Add lines for repmgr\u00a0\nhost\u00a0 \u00a0 repmgr\u00a0 \u00a0 \u00a0 repmgr\u00a0 \u00a0 \u00a0 192.168.134.0\/24\u00a0 \u00a0 trust\nhost\u00a0 \u00a0 replication repmgr\u00a0 \u00a0 \u00a0 192.168.134.0\/24\u00a0 \u00a0 trust<\/pre>\n
    In a primary node create \/etc\/repmgr.conf\u00a0<\/span><\/h3>\n
    sudo vi \/etc\/repmgr.conf<\/pre>\n
    \"\"<\/p>\n
    Node 01 (primary)\u00a0<\/span><\/h4>\n\n\n\n
    \n
    node_id=1\u00a0\nnode_name='node1'\u00a0\nconninfo='host=192.168.134.129 user=repmgr dbname=repmgr connect_timeout=2'\u00a0\ndata_directory='\/var\/lib\/pgsql\/16\/data'\u00a0\npg_bindir='\/usr\/pgsql-16\/bin'\u00a0\nconfig_directory='\/var\/lib\/pgsql\/16\/data\/postgresql.conf'\u00a0\nreplication_user='repmgr'\u00a0\nfailover=automatic\npromote_command='repmgr standby promote -f \/etc\/repmgr.conf --log-to-file'\u00a0\nfollow_command='repmgr standby follow -f \/etc\/repmgr.conf --log-to-file --upstream-node-id=%n'\u00a0\nlog_file='\/var\/log\/repmgr\/repmgr.log'<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
     <\/p>\n
    Node 02 (secondary)\u00a0<\/span><\/h4>\n
    sudo vi \/etc\/repmgr.conf\u00a0<\/span><\/p>\n\n\n\n
    \n
    node_id=2\u00a0\nnode_name='node2'\u00a0\nconninfo='host=192.168.134.130 user=repmgr dbname=repmgr connect_timeout=2'\u00a0\ndata_directory='\/var\/lib\/pgsql\/16\/data'\u00a0\npg_bindir='\/usr\/pgsql-16\/bin'\u00a0\nconfig_directory='\/var\/lib\/pgsql\/16\/data\/postgresql.conf'\u00a0\nreplication_user='repmgr'\u00a0\nfailover=automatic\npromote_command='repmgr standby promote -f \/etc\/repmgr.conf --log-to-file'\u00a0\nfollow_command='repmgr standby follow -f \/etc\/repmgr.conf --log-to-file --upstream-node-id=%n'\u00a0\nlog_file='\/var\/log\/repmgr\/repmgr.log'<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
    Node 03 (Witness)\u00a0<\/span><\/h4>\n
    Sudo vi \/etc\/repmgr.conf\u00a0<\/span><\/p>\n\n\n\n
    \n
    node_id=3\u00a0\nnode_name='witness'\u00a0\nconninfo='host=192.168.134.131 user=repmgr dbname=repmgr connect_timeout=2'\u00a0\ndata_directory='\/var\/lib\/pgsql\/16\/data'\u00a0\npg_bindir='\/usr\/pgsql-16\/bin'\u00a0\nreplication_user='repmgr'\u00a0\nlog_file='\/var\/log\/repmgr\/repmgr.log'<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
    Create Log Directory\u00a0<\/span><\/h4>\n
    In all nodes create log directory\u00a0<\/span><\/p>\n\n\n\n
    \n
    sudo mkdir -p \/var\/log\/repmgr\u00a0\nsudo chown postgres:postgres \/var\/log\/repmgr<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
    Register nodes in a cluster:<\/span><\/h2>\n
    Register Primary Node in a repmgr cluster:<\/span><\/h3>\n
    repmgr -f \/etc\/repmgr.conf primary register<\/pre>\n
    \"\"<\/span><\/h3>\n
    Building standby from the primary:<\/span><\/h3>\n
    repmgr -h 192.168.134.129 -U repmgr -d repmgr -f \/etc\/repmgr.conf standby clone -F<\/pre>\n
    \"\"<\/p>\n
    Register standby node:
    \n\"\"
    \n<\/span><\/p>\n
    Register Witness\u00a0 node in a repmgr cluster make sure datadirectory name is different from primary:<\/span><\/h3>\n
    repmgr -f \/etc\/repmgr.conf witness register -h 192.168.134.129 -p 5432 -U repmgr<\/pre>\n
    \"\"Register repmgrd service in all the three node:<\/span><\/h3>\n
    sudo vi \/etc\/systemd\/system\/repmgrd.service<\/pre>\n\n\n\n
    \n
    [Unit]\u00a0\nDescription=A replication manager, and failover management tool for PostgreSQL\u00a0\nAfter=syslog.target\u00a0\nAfter=network.target\u00a0\nAfter=postgresql.service\u00a0\n\n[Service]\u00a0\nType=forking\u00a0\nUser=postgres\u00a0\nExecStart=\/usr\/pgsql-16\/bin\/repmgrd -f \/etc\/repmgr.conf --pid-file=\/var\/run\/repmgrd.pid --daemonize\u00a0\nExecReload=\/bin\/kill -HUP $MAINPID\u00a0\nPIDFile=\/var\/run\/repmgrd.pid\u00a0\n\n[Install]\u00a0\nWantedBy=multi-user.target<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
    Reload repmgrd and its services:<\/span><\/h4>\n
    sudo systemctl daemon-reload\u00a0\nsudo systemctl enable repmgrd\u00a0\nsudo systemctl start repmgrd<\/pre>\n
    \"\"Verify repmgr running and checking status:<\/span><\/h2>\n
    repmgr -f \/etc\/repmgr.conf\u00a0 cluster show\nrepmgr -f \/etc\/repmgr.conf\u00a0 service status\nrepmgr -f \/etc\/repmgr.conf\u00a0 cluster crosscheck<\/pre>\n
    \"\"<\/p>\n
    Conclusion:<\/h2>\n
    Deploying a PostgreSQL 16.11 high-availability cluster using repmgr 5.5 on CentOS Stream 10 showcases the ability of open-source tools to provide enterprise-level reliability. The setup involves a primary node, a hot standby, and a witness server to enable synchronous replication and automatic failover. The detailed setup process emphasizes the significance of planning and reproducible configuration. The cluster supports seamless failover and minimal downtime, making it suitable for critical workloads. With improved support for PostgreSQL 16.x, repmgr offers an effective solution for managing replication and failover, creating a reliable and scalable HA platform for modern data infrastructures.<\/p><\/blockquote>\n","protected":false},"excerpt":{"rendered":"
    Introduction High availability (HA) is a critical requirement for modern database infrastructures, ensuring minimal downtime and seamless failover in case of node failures. PostgreSQL, while robust as a standalone database, requires additional tooling to manage replication, failover, and cluster orchestration in multi\u2011node environments. repmgr (Replication Manager) is a widely adopted open\u2011source tool that simplifies the…<\/p>\n","protected":false},"author":1,"featured_media":2951,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[28],"tags":[58,94,124,138,143,153,159,191,232,266,305,307],"class_list":["post-2938","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-postgresql","tag-sysadmin","tag-centosstream","tag-databasecluster","tag-dba","tag-devops","tag-failover","tag-highavailability","tag-linuxadmin","tag-opensource","tag-postgresql","tag-replication","tag-repmgr","category-28","description-off"],"aioseo_notices":[],"_links":{"self":[{"href":"https:\/\/divaind.com\/ie1\/wp-json\/wp\/v2\/posts\/2938","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/divaind.com\/ie1\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/divaind.com\/ie1\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/divaind.com\/ie1\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/divaind.com\/ie1\/wp-json\/wp\/v2\/comments?post=2938"}],"version-history":[{"count":0,"href":"https:\/\/divaind.com\/ie1\/wp-json\/wp\/v2\/posts\/2938\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/divaind.com\/ie1\/wp-json\/wp\/v2\/media\/2951"}],"wp:attachment":[{"href":"https:\/\/divaind.com\/ie1\/wp-json\/wp\/v2\/media?parent=2938"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/divaind.com\/ie1\/wp-json\/wp\/v2\/categories?post=2938"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/divaind.com\/ie1\/wp-json\/wp\/v2\/tags?post=2938"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}