Release Announcement EDB Bi-Directional Replication 3.7.9

Highlights

Type

Add bdr.local_group_slot_name() function which returns the group slot name. Useful primarily for monitoring.

Enhancement

Add bdr.workers view which shows additional information about BDR workers. Helps with monitoring of BDR specific activity. Useful especially when joined with bdr.stat_activity.

Enhancement

Allow Parallel Apply on logical standbys for forwarded transactions. Previously, parallel apply could be used only for changes replicated directly from the upstream of the logical standby, but not for any changes coming from another node.

Enhancement

Introduce bdr.batch_inserts configuration variable. This sets after how many INSERTs into the same table in a row (in the same transaction) BDR will switch to a multi insert strategy. This normally improves performance of replication of large data loads, be it via INSERTs or the COPY command. However BDR 3.7.8 would try to use this strategy always which would result in performance degradation in workloads that do many single row inserts only. [Support tickets 71004, 70727]

Enhancement

Fix group slot advance on subscriber-only nodes. This solves excessive WAL log retention on subscriber-only nodes. [Support ticket 71182]

Bug

Use the correct slot name when joining subscriber-only nodes using bdr_init_physical. The bdr_init_physical used to create the wrong slot, which resulted in 2 slots existing on the join source node when subscriber-only node was joined using this method. This would result in excessive WAL retention on the join source node. [Support ticket 71124]

Bug

Fix group monitoring view to allow more than one row per node. Group monitoring views would previously truncate the information from any node reporting more than one row of information. This would result in for example slots missing in bdr.group_replslots_details.

Bug

Fix Raft snapshot creation for auto partitioned tables. Previously the Raft snapshot didn't take into account the state of autopartition tasks on all nodes when writing the information. This could result in some nodes skipping partition creation after a prolonged period of downtime. [Support ticket 71178]

Bug

Fix DDL replication and locking management of several commands. ANALYZE, CHECKPOINT, CLUSTER, PREPARE/COMMIT/ABORT TRANSACTION, MOVE, RELEASE, ROLLBACK were documented as replicated and some of these even tried to take DDL lock which they should not.

Bug

Reduce logging of some unreplicated utility commands. PREPARE and EXECUTE don't need to spam about not being replicated as nobody expects that they would be.

Bug

Fix global locking of ALTER TABLE ... SET. It should not take a global DML lock.

Bug

Fix documentation about how the TRUNCATE command is replicated. While TRUNCATE can acquire global locks, it's not replicated the way other DDL commands are, it's replicated like DML, according to replication set settings.

Bug

Document that CAMO and Eager currently don't work with Single Decoding Worker.

Bug