Skip to content

Logs from etgar#48

Open
krleonid wants to merge 4 commits into
etgar_9bde721from
logs-from-etgar
Open

Logs from etgar#48
krleonid wants to merge 4 commits into
etgar_9bde721from
logs-from-etgar

Conversation

@krleonid
Copy link
Copy Markdown
Owner

No description provided.

…ed > 1s

Logs via DUCKDB_LOG_WARNING with queue_size and dead_nodes count.
Also prints to stderr in DEBUG builds for local debugging.

Requires: SET GLOBAL enable_logging = true; SET GLOBAL logging_level = 'warning';

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
krleonid and others added 3 commits May 13, 2026 18:28
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Defer re-enqueue of alive nodes until after the purge loop completes.
Previously, enqueue_bulk fed alive nodes back into the purge thread's
own sub-queue between iterations, causing try_dequeue_bulk to re-drain
those same alive nodes instead of reaching dead entries in worker
thread sub-queues. Also add stderr print for purge warnings.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add IsDeadNode() to correctly identify dead nodes (expired weak_ptr or
  sequence mismatch) without conflating pinned-but-alive nodes as dead
- Use IsDeadNode() in PurgeIteration instead of TryGetBlockMemory() which
  incorrectly counted pinned nodes as dead
- Remove GetBuffer() guard in BlockMemory destructor so IncrementDeadNodes
  fires for blocks destroyed after eviction (buffer already null)
- Add total_dequeued breakdown (alive/dead) to Purge warning log

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant