Merge pull request ceph#60438 from xxhdx1985126/wip-68661

crimson/osd/pg: make sure ops are sent to backfill targets if and only if the object has been pushed.

Reviewed-by: Samuel Just <[email protected]>

src/crimson/osd/pg.cc

@@ -1621,14 +1621,21 @@ bool PG::should_send_op(
     return true;
   bool should_send =
     (hoid.pool != (int64_t)get_info().pgid.pool() ||
-    (has_backfill_state() && hoid <= get_last_backfill_started()) ||
-    hoid <= peering_state.get_peer_info(peer).last_backfill);
+    // An object has been fully pushed to the backfill target if and only if
+    // either of the following conditions is met:
+    // 1. peer_info.last_backfill has passed "hoid"
+    // 2. last_backfill_started has passed "hoid" and "hoid" is not in the peer
+    //    missing set
+    hoid <= peering_state.get_peer_info(peer).last_backfill ||
+    (has_backfill_state() && hoid <= get_last_backfill_started() &&
+     !peering_state.get_peer_missing(peer).is_missing(hoid)));
   if (!should_send) {
     ceph_assert(is_backfill_target(peer));
     logger().debug("{} issue_repop shipping empty opt to osd."
                    "{}, object {} beyond std::max(last_backfill_started, "
                    "peer_info[peer].last_backfill {})",
-                   peer, hoid, peering_state.get_peer_info(peer).last_backfill);
+                   __func__, peer, hoid,
+                   peering_state.get_peer_info(peer).last_backfill);
   }
   return should_send;
   // TODO: should consider async recovery cases in the future which are not supported