rgw: fix handling of bucket layout in metadata sync

clear the bucket layout we get from the metadata master, and overwrite it
with our zone's defaults

without clearing the layout, init_default_bucket_layout() was adding another
log layout in addition to the one from the master. this caused the bilog
list API to provide a 'next_log' when only gen=0 exists

Signed-off-by: Casey Bodley <cbodley@redhat.com>

diff --git a/src/rgw/rgw_bucket.cc b/src/rgw/rgw_bucket.cc
index f6c0a280c0e25eece2013288d1cc30cf3d1bb594..30166c7820af2d5a04762a665ebb53f4f18ba5b8 100644 (file)
--- a/src/rgw/rgw_bucket.cc
+++ b/src/rgw/rgw_bucket.cc
@@ -2322,12 +2322,11 @@ int RGWMetadataHandlerPut_BucketInstance::put_check(const DoutPrefixProvider *dp
   if (from_remote_zone) {
     // don't sync bucket layout changes
     if (!exists) {
-      auto& bci_index = bci.info.layout.current_index.layout;
-      auto index_type = bci_index.type;
-      auto num_shards = bci_index.normal.num_shards;
+      // replace peer's layout with default-constructed, then apply our defaults
+      bci.info.layout = rgw::BucketLayout{};
       init_default_bucket_layout(cct, bci.info.layout,
                                 bihandler->svc.zone->get_zone(),
-                                num_shards, index_type);
+                                std::nullopt, std::nullopt);
     } else {
       bci.info.layout = old_bci->info.layout;
     }