A.1. Postgres Pro Shardman 14.17.1 #

Release date: 2025-03-17

This release is based on PostgreSQL 14.17 and Shardman 14.15.4 and provides new features, optimizations and bug fixes. Major changes are as follows:

A.1.1. Core and Extensions #

  • The shardman.silk_shmem_size, shardman.silk_netbuf_size shardman.silk_suspend_shmqe_limit, shardman.silk_resume_shmqe_limit, shardman.silk_suspend_netqe_limit, shardman.silk_resume_netqe_limit parameters now cannot change their values and are only kept for compatibility purposes.

  • Added a detailed description for the following configuration parameters: shardman.silk_unassigned_job_queue_size, shardman.silk_max_message, shardman.silkworm_fetch_size, and shardman.silk_hello_timeout.

  • Optimized mechanisms to result in receiving of a consistent syncpoint.

  • Improved error messages for temporary sharded or global tables creation failures.

  • Updated the CREATE USER MAPPING, ALTER USER MAPPING, and DROP USER MAPPING commands that are now prohibited when applied to mappings for foreign servers from the Shardman cluster.

  • Updated the shardman.users and pg_user_mapping catalogs that are now not stored in plain-text.

  • Fixed an issue with the processing of ALTER INDEX commands for the sharded tables.

  • Fixed a bug related to the Silk transport that previously resulted in a recursive error and the postmaster crash.

  • Fixed the BDU:2025-01601 vulnerabilitiy.

A.1.2. Management Utilities #

  • Added a new parameter --lock-lifetime to the probackup backup command to allow setting the maximum time that pg_probackup can hold the lock, in seconds.

  • Updated the shardmanctl forall, shardmanctl load, and shardmanctl history commands so they can run concurrently and do not block other processes.

  • Added a new option -n|--node to the commands shardmanctl shard stop and shardmanctl shard start to specify the node to start or stop.

  • Improved error messaging for the pg_probackup-related tools.

  • Implemented safe restoration of the etcd cluster from the dump for the cold backup by adding a shardmanctl store restore command.

  • Optimized the backup validation process by adding new options --data-validate, -remote-port, --remote-user, and --ssh-key to the shardmanctl probackup restore command.

  • Added a new filter restart_required_params to the shardmanctl status command that checks that all the postgres parameters requiring a postgres instance restart are applied. The successful output shows no pending restart parameters.

  • Implemented the automatic confirmation of the restart for the shardmanctl probackup archive-command add and the probackup backup commands with the -y|--yes option.

  • Updated the shardmanctl history output to show whether the listed commands succeeded or failed.

  • Implemented the automatic confirmation of the restart necessary for the parameters to take effect for the shardmanctl config update and shardmanctl config set command with the -y|--yes option. If this option is not specified, and the parameters update requires a restart, the manual confirmation will be requested. If not confirmed, the cluster will continue to work, yet the new parameter values will only take effect after the restart.

  • Fixed an issue that previously resulted in the pg_hba.conf row duplicates.

  • Fixed a bug that previously resulted in the shardmanctl status command failure.

  • Updated the supported version of pg_probackup to 2.8.8.