| 文件 | 最后提交记录 | 最后更新时间 |
|---|---|---|
SPQ HA: cm_ctl-based DN callback and config cleanup; wait for cluster stable before updating pg_dist_node on_dn_role_change.sh: reconcile via cm_ctl query; poll until DN cluster is stable (exactly one Primary+Normal), then double-confirm new primary via pg_is_in_recovery()=false before spq_update_node, avoiding writes during switchover transitions. on_cn_role_change.sh: wait until local gaussdb exits recovery before spq_set_coordinator_host. spq_topology.conf.sample, README.md: simplify config and route logs/locks to the CM agent log dir. | 4 天前 |