Hserver Failure Notes: Data Integrity and Publishing · advanced
Remote Terminal States Should Stop Re-Publishing Logic
If the provider already reports scheduled, publishing or published, the reconciler should observe rather than repeat the side effect.
A reconciler that finds an existing remote object still needs to decide whether to schedule, publish or do nothing. Reissuing a publish operation against an object already in flight can create confusing or provider-specific behavior. Discovery alone did not make the workflow idempotent; state-aware action selection was still required.
The publisher treats scheduled, publishing, published and partial as states where the existing remote object should be returned rather than re-triggered.
Mirror the provider state machine explicitly and write tests for every remote state your reconciler can observe, including unknown or newly introduced values. Idempotent workflows combine object identity with transition semantics. The same command can be safe in one state and wrong in another. The concrete hserver evidence is commit 9118e08, so this note is tied to an actual production change rather than a hypothetical failure.