DAOS-14976 object: properly select collective punch leader for resend (…

…#13602) Before resending the collective punch RPC, we need to check whether the original leader shard is valid or not. It is possible the object layout has been shrinked after rebuild. Under such case, select a new shard as the collective punch leader. Signed-off-by: Fan Yong <[email protected]>
daos-stack · Jan 23, 2024 · 67723f1 · 67723f1
1 parent e1d0cd7
commit 67723f1
Showing 1 changed file with 11 additions and 5 deletions.
diff --git a/src/object/cli_obj.c b/src/object/cli_obj.c
@@ -7060,7 +7060,12 @@ dc_obj_coll_punch(tse_task_t *task, struct dc_object *obj, struct dtx_epoch *epo
 	if (rc != 0)
 		goto out;
 
+	leader = coa->coa_dct_nr;
+
 	if (auxi->io_retry) {
+		if (unlikely(spa->pa_auxi.shard >= obj->cob_shards_nr))
+			goto new_leader;
+
 		/* Try to reuse the same leader. */
 		rc = obj_shard_open(obj, spa->pa_auxi.shard, map_ver, &shard);
 		if (rc == 0) {
@@ -7078,10 +7083,13 @@ dc_obj_coll_punch(tse_task_t *task, struct dc_object *obj, struct dtx_epoch *epo
 		/* Then change to new leader for retry. */
 	}
 
-	/* Randomly select a rank as the leader. */
-	leader = d_rand() % coa->coa_dct_nr;
-
 new_leader:
+	if (leader == coa->coa_dct_nr)
+		/* Randomly select a rank as the leader. */
+		leader = d_rand() % coa->coa_dct_nr;
+	else
+		leader = (leader + 1) % coa->coa_dct_nr;
+
 	dct = &coa->coa_dcts[leader];
 	len = dct->dct_bitmap_sz << 3;
 
@@ -7098,8 +7106,6 @@ dc_obj_coll_punch(tse_task_t *task, struct dc_object *obj, struct dtx_epoch *epo
 		}
 	}
 
-	/* Try another for leader. */
-	leader = (leader + 1) % coa->coa_dct_nr;
 	goto new_leader;
 
 gen_mbs: