Don't debug log when a cache update_counter fails

With recent reworkings of the message store this can fail
more often than it did before. That's fine because in that
case we will just try again. We don't need to keep the
debug log.

This only applies to fan-out scenarios.

Author: Loïc Hoguin

deps/rabbit/src/rabbit_msg_store.erl

@@ -1235,10 +1235,21 @@ contains_message(MsgId, From, State) ->
 
 update_msg_cache(CacheEts, MsgId, Msg) ->
     case ets:insert_new(CacheEts, {MsgId, Msg, 1}) of
-        true  -> ok;
-        false -> rabbit_misc:safe_ets_update_counter(
-                   CacheEts, MsgId, {3, +1}, fun (_) -> ok end,
-                   fun () -> update_msg_cache(CacheEts, MsgId, Msg) end)
+        true  ->
+            ok;
+        false ->
+            %% Note: This is basically rabbit_misc:safe_ets_update_counter/5,
+            %% but without the debug log that we don't want as the update is
+            %% more likely to fail following recent reworkings.
+            try
+                ets:update_counter(CacheEts, MsgId, {3, +1}),
+                ok
+            catch error:badarg ->
+                %% The entry must have been removed between
+                %% insert_new and update_counter, most likely
+                %% due to a file rollover or a confirm. Try again.
+                update_msg_cache(CacheEts, MsgId, Msg)
+            end
     end.
 
 adjust_valid_total_size(File, Delta, State = #msstate {