Release notes for EDB Postgres Distributed version 5.2.0¶
EDB Postgres Distributed version 5.2.0 is a minor version of EDB Postgres Distributed.
Highlights of EDB Postgres Distributed 5.2¶
Parallel apply is now available for PGD’s Commit at Most Once (CAMO) synchronous commit scope and improving replication performance.
Parallel apply for native Postgres asynchronous and synchronous replication has been improved for workloads where the same key is being modified concurrently by multiple transactions to maintain commit sequence and avoid deadlocks
PGD Proxy has added HTTP(S) APIs to allow the health of the proxy to be monitored directly for readiness and liveness. See Proxy health check
Component |
Version |
Type |
Description |
|---|---|---|---|
BDR |
5.2.0 |
Feature |
Parallel apply for synchronous commit scopes with CAMO |
BDR |
5.2.0 |
Feature |
Allow multiple SYNCHRONOUS_COMMIT clauses in a Commit Scope rule |
BDR |
5.2.0 |
Enhancement |
Allow transaction streaming with SYNCHRONOUS_COMMIT and LAG CONTROL Commit Scopes |
BDR |
5.2.0 |
Enhancement |
Improve handling of concurrent workloads with parallel apply |
BDR |
5.2.0 |
Enhancement |
Modify bdr.stat_subscription for new columns |
BDR |
5.2.0 |
Bug fix |
Allow a logical join of node if there are foreign key constraints violations. (RT91745) |
BDR |
5.2.0 |
Bug fix |
Change to group_raft_details view to avoid deadlock possibility |
BDR |
5.2.0 |
Bug fix |
Log the extension upgrade |
BDR |
5.2.0 |
Bug fix |
Add check for conflicting node names |
BDR |
5.2.0 |
Bug fix |
Fix a crash during Raft manual snapshot restore |
BDR |
5.2.0 |
Bug fix |
Don’t try to establish consensus connections to parting or parted nodes |
BDR |
5.2.0 |
Bug fix |
Fix tcp_user_timeout GUC to use the correct unit |
BDR |
5.2.0 |
Bug fix |
Fix the consensus snapshot compatibility with PGD 3.7. (RT93022) |
BDR |
5.2.0 |
Bug fix |
Address a crash when BDR is used together with pgaudit |
BDR |
5.2.0 |
Bug fix |
Skip parting synchronization to witness node |
BDR |
5.2.0 |
Bug fix |
Generate correct keepalive parameters in connection strings |
BDR |
5.2.0 |
Bug fix |
Enable various scenarios of switching nodes between groups and their subgroups. For example, transition node from a group to any of the nested sub-groups |
BDR |
5.2.0 |
Bug fix |
Reduce the amount of WAL produced by consensus on idle server |
BDR |
5.2.0 |
Bug fix |
Fixed deadlock on autopartition catalogs when a concurrent DROP EXTENSION is executing |
BDR |
5.2.0 |
Bug fix |
Fix sporadic failure when dropping extension after node restart |
BDR |
5.2.0 |
Bug fix |
Add workaround for crash due to pgaudit bug 212 |
BDR |
5.2.0 |
Bug fix |
Fix deadlock between consensus and global monitoring queries |
BDR |
5.2.0 |
Bug fix |
Fix query cancellation propagation across bdr.run_on_all_nodes |
BDR |
5.2.0 |
Bug fix |
Disallow invoking bdr.run_on_nodes(), bdr.run_on_group() and bdr.run_on_all_nodes() on parted nodes |
CLI |
5.2.0 |
Enhancement |
Add new GUCs verification in verify-settings command |
CLI |
5.2.0 |
Bug fix |
Truncate the long value of GUC in tabular output of verify-settings |
CLI |
5.2.0 |
Bug fix |
Upgrade database driver library version which fixes connect_timeout issue when sslmode=allow or sslmode=prefer |
Proxy |
5.2.0 |
Feature |
Add HTTP(S) APIs for Proxy health check |
Proxy |
5.2.0 |
Enhancement |
Improve route change events handling mechanism |
Proxy |
5.2.0 |
Enhancement |
Add retry mechanism on consensus query error |
Proxy |
5.2.0 |
Bug fix |
Upgrade database driver library version which fixes connect_timeout issue when sslmode=allow or sslmode=prefer |
