If you deployed your PGD cluster with Trusted Postgres Architect (TPA), EDB Postgres Distributed for Kubernetes, or Hybrid Manager, that tool manages your Postgres and PGD upgrades together, rather than through the CLI or manual package installation described elsewhere in this section.
Upgrading with Trusted Postgres Architect (TPA)
Run tpaexec upgrade to let TPA automate your Postgres and PGD upgrade, staging a BDR 3/4 cluster to PGD 5
or a PGD 5.9+ cluster to PGD 6 by migrating to the built-in Connection Manager before performing the
software upgrade.
Note
TPA supports upgrading PGD 4 to PGD 5 and PGD 5 to PGD 6, but not a direct PGD 4 to PGD 6 upgrade.
Refer to Upgrading from BDR-Always-ON to PGD-Always-ON and Upgrading from PGD-Always-ON to PGD-X in the TPA documentation for instructions.
Upgrading with EDB Postgres Distributed for Kubernetes
Change the operand image to upgrade Postgres and PGD together. EDB CloudNativePG Global Cluster performs the change as a rolling, node-by-node operation, managed through the Kubernetes operator rather than the CLI or manual package installation.
Refer to Installation and upgrade in the EDB CloudNativePG Global Cluster documentation for instructions.
Upgrading with Hybrid Manager
Use the Hybrid Manager console to upgrade Postgres and PGD together for Distributed High-Availability (DHA) and Advanced High Availability (AHA) clusters. For a combined major-version upgrade, HM uses a physical join between data groups so they can temporarily run different major versions during the upgrade. Upgrade one major version at a time, direct version skips aren't supported.
Refer to Postgres major version upgrades and Postgres major version upgrades for distributed high availability in the Hybrid Manager documentation for instructions.