DAOS-18633 rebuild: abort orphaned reclaim rpt after PS leader switch

After PS leader switch, ds_rebuild_regenerate_task() only regenerates
rebuild tasks for DOWN/DRAIN/UP targets. RECLAIM tasks are not
regenerated because reintegrated targets are already UPIN. This
leaves orphaned rpt on every target with a stale leader term, whose
IV updates are silently dropped by the new leader (no matching rgt).
The result is sp_rebuilding > 0 permanently, blocking EC aggregation
and causing system-wide performance degradation.

Fix: detect stale leader term in rebuild_tgt_status_check_ult() and
abort the orphaned rpt.

TODO: persist in-progress reclaim tasks in RDB so they can be
properly re-triggered on PS leader step_up.

Signed-off-by: Wang Shilong <shilong.wang@hpe.com>

Author: wangshilong

src/rebuild/srv.c

@@ -2972,6 +2972,23 @@ rebuild_tgt_status_check_ult(void *arg)
 		if (!rpt->rt_global_done) {
 			struct ds_iv_ns *ns = rpt->rt_pool->sp_iv_ns;
 
+			/* Abort orphaned rpt whose leader is gone.
+			 * After PS leader switch, reclaim tasks are
+			 * not regenerated (UPIN not in DOWN/UP/DRAIN),
+			 * so this rpt has no matching rgt on the new
+			 * leader and IV updates are silently dropped.
+			 */
+			if (rpt->rt_leader_term < ns->iv_master_term) {
+				D_WARN(DF_UUID " ver %d gen %u op %s: "
+					       "stale term " DF_U64 " < " DF_U64
+					       ", abort orphaned rpt\n",
+				       DP_UUID(rpt->rt_pool_uuid), rpt->rt_rebuild_ver,
+				       rpt->rt_rebuild_gen, RB_OP_STR(rpt->rt_rebuild_op),
+				       rpt->rt_leader_term, ns->iv_master_term);
+				rpt->rt_abort = 1;
+				break;
+			}
+
 			iv.riv_master_rank       = ns->iv_master_rank;
 			iv.riv_rank = rpt->rt_rank;
 			iv.riv_ver = rpt->rt_rebuild_ver;