What orioledb/orioledb shipped

OrioleDB v1.10.2 is out with the corrected version stamp. This week also fixed row lock conflict handling, UPDATE during concurrent index builds, expression index memory on TOASTed rows, tablespace moves on standbys and NULL-argument crashes. Nightly builds are available.

OrioleDB v1.10.2 is released; it corrects ORIOLEDB_VERSION after v1.10.1 left it unchanged, with no storage or WAL format changes. Work landed this week on row lock conflicts, including stopping the walk at the transaction's own version, and on UPDATE during concurrent index builds. Expression index builds on TOASTed rows no longer leak memory, tablespace moves on physical standbys are fixed, and four structure and index functions no longer crash on NULL arguments. A plain-text ORIOLEDB_VERSION file is now written to the orioledb_data directory so tools can identify a cluster's version without connecting.

This week: fixed a bgwriter crash on range, array and composite index keys, a decoder crash in TOASTed tables, and the OLD tuple for AFTER UPDATE triggers. Also fixed a failover slot sync race and a CREATE INDEX CONCURRENTLY spool buffer issue.

This week's changes: the background writer no longer crashes on range, multirange, array and composite index keys, and the logical decoder no longer crashes on TOASTed tables with short chunks. AFTER UPDATE triggers now get the correct OLD tuple, CREATE INDEX CONCURRENTLY gives each spool stream its own buffers, and failover slot sync takes the standby base backup before creating the slot. Autonomous transaction abort now takes changes back before marking the transaction aborted, and the checkpointer waits instead of dying on transient page pool exhaustion.

Dropping a database no longer leaves an empty orioledb_data directory. Parallel scans no longer hit a stuck spinlock PANIC when DSM creation fails. Rewrite copies from table-rewriting ALTERs are skipped in logical decoding, so subscribers no longer stop on a duplicated key.

This week's OrioleDB changes include removing the leftover empty orioledb_data directory after DROP DATABASE, releasing workerStart before parallel scan DSM creation so a failed creation no longer leaves workers spinning until a PANIC restarts the instance, and skipping rewrite copies in logical decoding so subscribers no longer stop with a duplicated key error. We also fixed waking oxid waiters by granting their lock, kept bitmap scan EXPLAIN counters across rescans, removed a false assertion in undo_it_next_page, and made backpatch detection stop at the first matching declaration to avoid SIGPIPE aborts.

This week: stopped writing replident full WAL when wal_level<logical, ran wal_rollback in a critical section, limited undo link checks to where they are followed, fixed orphaned relfilenodes after ROLLBACK, reduced isol_rr iterations to avoid Valgrind timeouts, prevented duplicat

This week in OrioleDB: we stopped writing replident full WAL when wal_level is below logical, ran wal_rollback in a critical section to prevent a second ROLLBACK after an ERROR, limited page undo link checks to where links are followed to avoid a false PANIC after restart, fixed orphaned relfilenode files after ROLLBACK, reduced isol_rr iterations so the test no longer times out under Valgrind, prevented duplicate downlink insertion on BROKEN_SPLIT, handled torn OTable/OIndex reads during concurrent ALTER TABLE, released undo reservations on early exits in btree_try_merge_and_unlock, uploaded the isolation server log for failed CI runs, and fixed assertions that were always true due to assignment instead of comparison. Stable branches now also run full CI.

This week: logical replication streaming redesigned so a mixed transaction cannot leave a broken slot; enum and composite type cache consistency fixed; enum cache inserts made rollback-safe; ALTER TABLE SET ACCESS METHOD orioledb rejected on non-orioledb tables; index bridge no l

This week's changes to the project focus on correctness in places that are hard to test by hand. Logical replication transaction streaming was redesigned around a decoding startup hook, so a mixed transaction that has only touched heap cannot leave a broken slot without rebuilding the downstream. Enum and composite type cache consistency was fixed across a family of bugs, and enum cache inserts are now rollback-safe. ALTER TABLE SET ACCESS METHOD orioledb is now rejected on non-orioledb tables instead of crashing. Bridged indexes are no longer written twice during ON CONFLICT insert, and index and bitmap scans now read the bridge with the query's snapshot. A rollback under row locks no longer rewrites the lock chain. CI work for dockerTEST and pg_upgrade is now limited to main branch pushe

Beta19 is out. Key fixes: replicas no longer lose rows during crash recovery from table rewrites, CREATE DATABASE with enum indexes no longer crashes, and DROP TYPE handles undo correctly during checkpoint cleanup.

OrioleDB beta19 released with critical data integrity fixes. Replicas can no longer lose rows before a restartpoint during crash recovery from table rewrites. System cache handling improved to prevent use-after-free and ensure DROP TYPE works correctly with checkpoint cleanup. CREATE DATABASE with enum-indexed columns now works without segmentation faults. Several test flakes on loaded CI also resolved.

OrioleDB beta18 ships with DROP TABLESPACE replication and improved logical decoding. This week also fixed COPY corruption in index-organized tables, replication switchover divergence, and several recovery safety issues.

OrioleDB beta18 is out with DROP TABLESPACE replication support and enhanced logical decoding that carries TOASTed attributes inline. This week the team fixed critical issues: COPY corruption in index-organized tables that could leave B-trees inconsistent, replication switchover divergence that blocked recovery, memory safety problems in undo chain handling, and descriptor lifecycle issues causing crash assertions. These fixes target production reliability for replication and bulk load workloads.

This week: fixed multi-insert undo tracking, parallel scan interruption, recovery collation handling, and seven validation gaps in WAL replay. Several addressed crash scenarios and data loss risks.

OrioleDB shipped fixes this week for critical issues in crash recovery and concurrent operations. Highlights include restoring undo tracking accuracy in multi-insert batches during checkpoints, making parallel scans properly interruptible, fixing recovery crashes with collated keys, and adding validation for seven quantities decoded from WAL before they're trusted as sizes or indices. Together these improve the robustness of recovery paths and prevent data loss scenarios.