Repository navigation
Epic: Data safety #189
Description
Outcome
Prevent storage corruption, false write success, incomplete subscription state and diagnostic deadlock. R1 delivers S01/S02/S05/S06. Issue #93 stays open for its later storage-disappearance/health probe; atomic durability is tracked in the Atomic programs epic.
Child issues
- Exclusive lock on data_dir at startup #142
- SDK: propagate server and statement failures through public write APIs #164
- Engine: fact inserts return success but store nothing when the data directory vanishes (WAL write failure swallowed) #93
- Subscriptions: reject capped results before replacing complete state #165
- Engine: remove recursive storage reads in diagnostics and proofs #166
Completion
Keep this tracking issue open until its selected release scope passes. Child issues in a later milestone remain visibly deferred; a partial slice does not complete a broader issue. Every change runs the performance pipeline in #163 before a pull request.
Activity
jsam commentedon Oct 4, 2026
Moved to the Engine production readiness milestone, the current focus (tracker #199). Data safety is part of production readiness (EN-1 #93, EN-9 #262).
Roadmap per the neurosymbolic decision layer plan: engine production readiness first, then gateway neural integration with Jev, CLM and Laya (#266).
jsam commentedon Oct 6, 2026
Completed by https://github.com/inputlayer/inputlayer-symbolic/pull/5 (merged): operator runbooks (docs/content/docs/guides/operations.mdx) for scheduled online backups with verify, a restore drill run for real, Prometheus alert rules for the engine, replication and backup metrics (promtool-tested and fired live), upgrade/rollback (tested against the 2026-10-03 image), and manual warm-standby failover/switchover (zero acknowledged writes lost in sync mode). Every child issue was already done.