yugabyte/yugabyte-db v2.3.2.0

v2.3.2.0 (Released September 29, 2020)

on GitHub

YSQL

• Enable pg_stat_statements extension by default. #5750
• Clean up unused PgStatement objects on hot path after usage to avoid memory consumption. #5374
• Fix restarting writes (INSERT and UPDATE) on transaction conflicts and avoid surfacing transaction errors to user applications. #4291
• Enable batching of updates for YSQL procedures and functions. #5257

YCQL

• Add additional system queries (SELECT_LOCAL, SELECT_LOCAL_TOKENS, and SELECT_PEERS_TOKENS) for caching. #5753
• Fix crash for YCQL index scan when ORDER BY columns exist in table, but not in the chosen index, and raise an error. #5690

Core database

• Allow setting a custom replacement policy per table through yb_admin. #5368
• Do not clean up transaction that potentially was committed. #5789 #5783 #5755
• Fix CDC shutdown races. Moved thread pool shutdown before other CDC objects cleanup. Added explicit CDC poller clear in CDC consumer shutdown. #5583
• Fix preparer deadlock after switching from leader to follower. #5741
• Fix RaftGroupMetadata construction issues found using PVS Studio. #5578
• Change rlimit flags to have at-least semantics. Only change the process soft limit if it is smaller than the limit specified by the corresponding flag value. #5715
• Add extra logging when initializing syscatalog. #5708
• Set default webserver_interface to rpc_bind_address IP/host component. #5320
• Add environmental variables functionality to yugabyted. New env variables for YSQL (YSQL_USER, YSQL_PASSWORD, and YSQL_DB) and YCQL(YCQL_USER, YCQL_PASSWORD, and YCQL_KEYSPACE). (#5499
• Improve cleanup of aborted transactions. #5669
• Fix dumping an empty batch by updating MaxReplicateOpId. #5679
• Set default staleness for read from followers (using max_stale_read_time_ms) to 60 seconds. Previously, a follower could return requested data no matter how state the data is. #5590
• Calculate the placement UUID from Raft upon tserver registration. #5558
• Enable ignoring deleted tablets. #5122
• Remove unreachable nodes from UI (/tablet-servers page) after nodes are removed from the cluster. #4759

Yugabyte Platform

• Fix issue parsing multiple comma-separated addresses for rpc_bind_addresses. #5634
• Add functionality to delete old backups after a specified time period. In the Create Backup dialog, added Number of Days to Retain Backup field. #4493
• Combine the Platform and community Grafana dashboards. Grafana dashboard shows the node status and other panels correctly. #5707
• Sum table sizes across all nodes of a universe. #5745
• Fixes releasing OnPrem nodes should not delete them from the universe metadata. Fixes releasing a node should only attempt to shut down processes on that node. Nodes can be re-added using Node > Actions > Add flow. #5703 #5791 #5792
• Add additional property to MetricsPanel to handle width of Replication graph and increase height. Aggregates replication metrics values from all tservers into an average and max value graph across all tservers. #5429 #5613 #5632
• Update yb-client JAR version to 0.8.1. #5759
• Stop and restart relevant services during Yugabyte Platform backup. Backups stop and restart services. Restores stop and restart Yugabyte Platform and Prometheus, if necessary. #5685
• Authenticate API tokens even when using OAuth. This allows curl requests. #5587
• Fix Metrics tab and columns misalignment on table page. #5582
• Rename In Use column to Universe Name and show NOT USED if the node is not in use. #5713
• Do not generate node certificates for custom certificates. #5545
• Make instance name field optional for on-premises provider. #5234
• Fix button for Metrics tab to reset active and focus state when key is unpressed. #5620
• Add Docker- and VM-aware options to the backup script and will include Prometheus snapshots. #5120 #5615 #5612 #5614

Note:
Prior to v2.0, YSQL was still in beta. Upon release of v2.0, a backward-incompatible file format change was made for YSQL. For existing clusters running pre-2.0 release with YSQL enabled, you cannot upgrade to v2.0 or later. Instead, export your data from existing clusters and then import the data into a new cluster (v2.0 or later).