Skip to content

handoff: exports created after takeover flush - #132

Closed
jaredLunde wants to merge 2 commits into
mainfrom
jared/successor-flush-gate
Closed

jaredLunde wants to merge 2 commits into
mainfrom
jared/successor-flush-gate

Conversation

@jaredLunde

Copy link
Copy Markdown
Contributor

Fixes exports created in a handoff-successor process never uploading on their own (dirty-age, threshold): GLIDEFS_HANDOFF_SUCCESSOR_PASSIVE was set at successor start and never cleared, so every later WriteCache was born Warming and flush_packs skipped it (checkpoint WAL truncation too).

The flag is now a process-global cleared right after set_all_caches_phase(Idle). Test tests/successor_flush.rs (own process; fails without the clear): Warming cache does not flush; existing cache flushes after takeover; post-takeover cache flushes by dirty age and by the 1,024-block threshold.

🤖 Generated with Claude Code

https://claude.ai/code/session_01WpMBiw38GkJFe49jfZowsH

jaredLunde and others added 2 commits November 7, 2026 06:46
…e flag was never cleared)

run_server_as_successor set GLIDEFS_HANDOFF_SUCCESSOR_PASSIVE for the life of the
process, so every WriteCache built after takeover was born Warming and
flush_packs skipped it forever: no dirty-age, threshold or any scheduler upload,
and no WAL truncation. The flag is now a process-global cleared when takeover
completes; no env var.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WpMBiw38GkJFe49jfZowsH
@jaredLunde

Copy link
Copy Markdown
Contributor Author

Superseded by the rebased PR on branch jared/successor-flush-gate-v2.

@jaredLunde jaredLunde closed this Nov 7, 2026
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