From 2c4b718162198e6db679113ce3196b391d8571e5 Mon Sep 17 00:00:00 2001
From: Seena Fallah <seenafallah@gmail.com>
Date: Mon, 3 Mar 2025 17:16:55 +0100
Subject: [PATCH] rgw: skip versioned entries to non-versioned buckets and
 vice-versa

As instance field is preserved while replicating, this can make
confusion for versioned buckets having non-versioned objects and
non-versioned buckets having versioned objects.

Fixes: https://tracker.ceph.com/issues/70486
Signed-off-by: Seena Fallah <seenafallah@gmail.com>
---
 src/rgw/driver/rados/rgw_data_sync.cc |  8 +++
 src/test/rgw/rgw_multi/tests.py       | 96 +++++++++++++++++++++++++++
 2 files changed, 104 insertions(+)

diff --git a/src/rgw/driver/rados/rgw_data_sync.cc b/src/rgw/driver/rados/rgw_data_sync.cc
index 7eb8db380889c..10fe3b227b38b 100644
--- a/src/rgw/driver/rados/rgw_data_sync.cc
+++ b/src/rgw/driver/rados/rgw_data_sync.cc
@@ -4520,6 +4520,14 @@ public:
             tn->log(0, "entry with empty obj name, skipping");
             goto done;
           }
+
+          // make sure versioned object only lands on versioned bucket and non-versioned object only lands on non-versioned bucket
+          if (key.instance.empty() == sync_pipe.dest_bucket_info.versioned()) {
+            set_status("skipping entry due to versioning mismatch");
+            tn->log(0, SSTR("skipping entry due to versioning mismatch: " << key));
+            goto done;
+          }
+
           if (error_injection &&
               rand() % 10000 < cct->_conf->rgw_sync_data_inject_err_probability * 10000.0) {
             tn->log(0, SSTR(": injecting data sync error on key=" << key.name));
diff --git a/src/test/rgw/rgw_multi/tests.py b/src/test/rgw/rgw_multi/tests.py
index c01937932b72c..a08f36ca4c6c1 100644
--- a/src/test/rgw/rgw_multi/tests.py
+++ b/src/test/rgw/rgw_multi/tests.py
@@ -3946,3 +3946,99 @@ def test_bucket_replicaion_reject_objectlock_identical():
                           }]
                       })
     assert e.response['ResponseMetadata']['HTTPStatusCode'] == 400
+
+@allow_bucket_replication
+def test_bucket_replication_non_versioned_to_versioned():
+    zonegroup = realm.master_zonegroup()
+    zonegroup_conns = ZonegroupConns(zonegroup)
+
+    source = zonegroup_conns.non_account_rw_zones[0]
+    dest = zonegroup_conns.non_account_rw_zones[1]
+
+    source_bucket = source.create_bucket(gen_bucket_name())
+    dest_bucket = dest.create_bucket(gen_bucket_name())
+    zonegroup_meta_checkpoint(zonegroup)
+
+    # create replication configuration
+    response = source.s3_client.put_bucket_replication(
+        Bucket=source_bucket.name,
+        ReplicationConfiguration={
+            'Role': '',
+            'Rules': [
+                {
+                    'ID': 'rule1',
+                    'Status': 'Enabled',
+                    'Destination': {
+                        'Bucket': f'arn:aws:s3:::{dest_bucket.name}',
+                    }
+                }
+            ]
+        }
+    )
+    assert response['ResponseMetadata']['HTTPStatusCode'] == 200
+    zonegroup_meta_checkpoint(zonegroup)
+
+    # enable versioning on destination bucket
+    dest.s3_client.put_bucket_versioning(
+        Bucket=dest_bucket.name,
+        VersioningConfiguration={'Status': 'Enabled'}
+    )
+    zonegroup_meta_checkpoint(zonegroup)
+
+    # upload an object and wait for sync.
+    objname = 'dummy'
+    k = new_key(source, source_bucket, objname)
+    k.set_contents_from_string('foo')
+    zone_data_checkpoint(dest.zone, source.zone)
+
+    # check that object not exists in destination bucket
+    e = assert_raises(ClientError, dest.s3_client.get_object, Bucket=dest_bucket.name, Key=objname)
+    assert e.response['Error']['Code'] == 'NoSuchKey'
+
+@allow_bucket_replication
+def test_bucket_replication_versioned_to_non_versioned():
+    zonegroup = realm.master_zonegroup()
+    zonegroup_conns = ZonegroupConns(zonegroup)
+
+    source = zonegroup_conns.non_account_rw_zones[0]
+    dest = zonegroup_conns.non_account_rw_zones[1]
+
+    source_bucket = source.create_bucket(gen_bucket_name())
+    dest_bucket = dest.create_bucket(gen_bucket_name())
+    zonegroup_meta_checkpoint(zonegroup)
+
+    # create replication configuration
+    response = source.s3_client.put_bucket_replication(
+        Bucket=source_bucket.name,
+        ReplicationConfiguration={
+            'Role': '',
+            'Rules': [
+                {
+                    'ID': 'rule1',
+                    'Status': 'Enabled',
+                    'Destination': {
+                        'Bucket': f'arn:aws:s3:::{dest_bucket.name}',
+                    }
+                }
+            ]
+        }
+    )
+    assert response['ResponseMetadata']['HTTPStatusCode'] == 200
+    zonegroup_meta_checkpoint(zonegroup)
+
+    # enable versioning on source bucket
+    source.s3_client.put_bucket_versioning(
+        Bucket=source_bucket.name,
+        VersioningConfiguration={'Status': 'Enabled'}
+    )
+    zonegroup_meta_checkpoint(zonegroup)
+
+    # upload an object and wait for sync.
+    objname = 'dummy'
+    k = new_key(source, source_bucket, objname)
+    k.set_contents_from_string('foo')
+    zone_data_checkpoint(dest.zone, source.zone)
+
+    # check that object not exists in destination bucket
+    e = assert_raises(ClientError, dest.s3_client.get_object, Bucket=dest_bucket.name, Key=objname)
+    assert e.response['Error']['Code'] == 'NoSuchKey'
-- 
2.39.5