c/topic_table: do not log duplicated lifecycle marker command

It is possible that the drop topic lifecycle marker replicate request
would timeout. In this case requester has no other option than to retry
dropping lifecycle marker. Logic in `cluster::topic_table` is idempotent
and duplicate marker drop command does not harm to Redpanda. Changed log
level of an entry notifying about duplicated drop command from an error
to info level.

Signed-off-by: Michal Maslanka <michal@redpanda.com>

src/v/cluster/topic_table.cc

@@ -173,9 +173,8 @@ topic_table::apply(topic_lifecycle_transition soft_del, model::offset offset) {
             _lifecycle_markers.erase(soft_del.topic);
             return ss::make_ready_future<std::error_code>(errc::success);
         } else {
-            // This is harmless but should not happen and indicates a bug.
             vlog(
-              clusterlog.error,
+              clusterlog.info,
               "Unexpected record at offset {} to drop non-existent lifecycle "
               "marker {} {}",
               offset,