test

authdb (20260712+4) unstable; urgency=high

  * ClusterBackend::fetchFromCluster(): only skip the real network fetch

    in favor of cached data while scrub/rebalance is running when there

    IS cached data — a node that has never successfully fetched a given

    domain (fresh restart, empty cache) now still attempts a real fetch

    even during scrub. Previously, a scrub/rebalance that never settles

    to idle (e.g. the just-fixed non-converging rebalance loop) would

    permanently block that node's first-ever fetch, leaving it stuck

    showing "cluster data unavailable — waiting for peers" forever.

  * Rebuild against libparitypp 20260711+9 (client::list() retry, which

    is the actual reason rebalance was never converging to idle in the

    first place).

 -- Jan Koester <jan.koester@tuxist.de>  Sun, 12 Jul 2026 13:15:00 +0200

authdb (20260712+3) unstable; urgency=high

  * Cluster::scrub(): drop the remove()+store() "cleanup" pass in the