Merge pull request #1315 from rabbitmq/rabbitmq-server-1310

michaelklishin · web-flow · commit c58a15e7893b · 2017-08-01T03:10:38.000+03:00
Check if vhost supervisor is running when starting mirrors
diff --git a/src/rabbit_amqqueue_sup_sup.erl b/src/rabbit_amqqueue_sup_sup.erl
@@ -57,15 +57,15 @@ find_for_vhost(VHost) ->
 
 -spec find_for_vhost(rabbit_types:vhost(), atom()) -> {ok, pid()} | {error, term()}.
 find_for_vhost(VHost, Node) ->
-    {ok, VHostSup} = rabbit_vhost_sup_sup:vhost_sup(VHost, Node),
+    {ok, VHostSup} = rabbit_vhost_sup_sup:get_vhost_sup(VHost, Node),
     case supervisor2:find_child(VHostSup, rabbit_amqqueue_sup_sup) of
         [QSup] -> {ok, QSup};
         Result -> {error, {queue_supervisor_not_found, Result}}
     end.
 
 -spec start_for_vhost(rabbit_types:vhost()) -> {ok, pid()} | {error, term()}.
 start_for_vhost(VHost) ->
-    case rabbit_vhost_sup_sup:vhost_sup(VHost) of
+    case rabbit_vhost_sup_sup:get_vhost_sup(VHost) of
         {ok, VHostSup} ->
             supervisor2:start_child(
               VHostSup,
@@ -82,7 +82,7 @@ start_for_vhost(VHost) ->
 
 -spec stop_for_vhost(rabbit_types:vhost()) -> ok.
 stop_for_vhost(VHost) ->
-    case rabbit_vhost_sup_sup:vhost_sup(VHost) of
+    case rabbit_vhost_sup_sup:get_vhost_sup(VHost) of
         {ok, VHostSup} ->
             ok = supervisor2:terminate_child(VHostSup, rabbit_amqqueue_sup_sup),
             ok = supervisor2:delete_child(VHostSup, rabbit_amqqueue_sup_sup);
diff --git a/src/rabbit_connection_tracking_handler.erl b/src/rabbit_connection_tracking_handler.erl
@@ -82,11 +82,14 @@ handle_event(#event{type = vhost_deleted, props = Details}, State) ->
     close_connections(rabbit_connection_tracking:list(VHost),
                       rabbit_misc:format("vhost '~s' is deleted", [VHost])),
     {ok, State};
+%% Note: under normal circumstances this will be called immediately
+%% after the vhost_deleted above. Therefore we should be careful about
+%% what we log and be more defensive.
 handle_event(#event{type = vhost_down, props = Details}, State) ->
     VHost = pget(name, Details),
     Node = pget(node, Details),
-    rabbit_log_connection:info("Closing all connections in vhost '~s' at node '~s'"
-                               " because the vhost database has stopped working",
+    rabbit_log_connection:info("Closing all connections in vhost '~s' on node '~s'"
+                               " because the vhost is stopping",
                                [VHost, Node]),
     close_connections(rabbit_connection_tracking:list_on_node(Node, VHost),
                       rabbit_misc:format("vhost '~s' is down", [VHost])),
@@ -131,7 +134,17 @@ close_connections(Tracked, Message, Delay) ->
     ok.
 
 close_connection(#tracked_connection{pid = Pid, type = network}, Message) ->
-    rabbit_networking:close_connection(Pid, Message);
+    try
+        rabbit_networking:close_connection(Pid, Message)
+    catch error:{not_a_connection, _} ->
+            %% could has been closed concurrently, or the input
+            %% is bogus. In any case, we should not terminate
+            ok;
+          _:Err ->
+            %% ignore, don't terminate
+            rabbit_log:warning("Could not close connection ~p: ~p", [Pid, Err]),
+            ok
+    end;
 close_connection(#tracked_connection{pid = Pid, type = direct}, Message) ->
     %% Do an RPC call to the node running the direct client.
     Node = node(Pid),
diff --git a/src/rabbit_mirror_queue_misc.erl b/src/rabbit_mirror_queue_misc.erl
@@ -228,11 +228,21 @@ add_mirror(QName, MirrorNode, SyncMode) ->
             rabbit_misc:with_exit_handler(
               rabbit_misc:const(ok),
               fun () ->
-                      SPid = rabbit_amqqueue_sup_sup:start_queue_process(
-                               MirrorNode, Q, slave),
-                      log_info(QName, "Adding mirror on node ~p: ~p~n",
-                               [MirrorNode, SPid]),
-                      rabbit_mirror_queue_slave:go(SPid, SyncMode)
+                    #amqqueue{name = #resource{virtual_host = VHost}} = Q,
+                    case rabbit_vhost_sup_sup:get_vhost_sup(VHost, MirrorNode) of
+                        {ok, _} ->
+                            SPid = rabbit_amqqueue_sup_sup:start_queue_process(
+                                       MirrorNode, Q, slave),
+                            log_info(QName, "Adding mirror on node ~p: ~p~n",
+                                     [MirrorNode, SPid]),
+                            rabbit_mirror_queue_slave:go(SPid, SyncMode);
+                        {error, Error} ->
+                            log_warning(QName,
+                                        "Unable to start queue mirror on node '~p'. "
+                                        "Target virtual host is not running: ~p~n",
+                                        [MirrorNode, Error]),
+                            ok
+                    end
               end);
         {error, not_found} = E ->
             E
diff --git a/src/rabbit_recovery_terms.erl b/src/rabbit_recovery_terms.erl
@@ -48,7 +48,7 @@
 %%----------------------------------------------------------------------------
 
 start(VHost) ->
-    case rabbit_vhost_sup_sup:vhost_sup(VHost) of
+    case rabbit_vhost_sup_sup:get_vhost_sup(VHost) of
       {ok, VHostSup} ->
             {ok, _} = supervisor2:start_child(
                         VHostSup,
@@ -65,7 +65,7 @@ start(VHost) ->
     ok.
 
 stop(VHost) ->
-    case rabbit_vhost_sup_sup:vhost_sup(VHost) of
+    case rabbit_vhost_sup_sup:get_vhost_sup(VHost) of
         {ok, VHostSup} ->
             case supervisor:terminate_child(VHostSup, ?MODULE) of
                 ok -> supervisor:delete_child(VHostSup, ?MODULE);
diff --git a/src/rabbit_vhost_limit.erl b/src/rabbit_vhost_limit.erl
@@ -55,7 +55,12 @@ notify(VHost, <<"vhost-limits">>, <<"limits">>, Limits, ActingUser) ->
 notify_clear(VHost, <<"vhost-limits">>, <<"limits">>, ActingUser) ->
     rabbit_event:notify(vhost_limits_cleared, [{name, <<"limits">>},
                                                {user_who_performed_action, ActingUser}]),
-    update_vhost(VHost, undefined).
+    %% If the function is called as a part of vhost deletion, the vhost can
+    %% be already deleted.
+    case rabbit_vhost:exists(VHost) of
+        true  -> update_vhost(VHost, undefined);
+        false -> ok
+    end.
 
 connection_limit(VirtualHost) ->
     get_limit(VirtualHost, <<"max-connections">>).
diff --git a/src/rabbit_vhost_msg_store.erl b/src/rabbit_vhost_msg_store.erl
@@ -23,7 +23,7 @@
 
 start(VHost, Type, ClientRefs, StartupFunState) when is_list(ClientRefs);
                                                      ClientRefs == undefined  ->
-    case rabbit_vhost_sup_sup:vhost_sup(VHost) of
+    case rabbit_vhost_sup_sup:get_vhost_sup(VHost) of
         {ok, VHostSup} ->
             VHostDir = rabbit_vhost:msg_store_dir_path(VHost),
             supervisor2:start_child(VHostSup,
@@ -39,7 +39,7 @@ start(VHost, Type, ClientRefs, StartupFunState) when is_list(ClientRefs);
     end.
 
 stop(VHost, Type) ->
-    case rabbit_vhost_sup_sup:vhost_sup(VHost) of
+    case rabbit_vhost_sup_sup:get_vhost_sup(VHost) of
         {ok, VHostSup} ->
             ok = supervisor2:terminate_child(VHostSup, Type),
             ok = supervisor2:delete_child(VHostSup, Type);
@@ -65,7 +65,7 @@ with_vhost_store(VHost, Type, Fun) ->
     end.
 
 vhost_store_pid(VHost, Type) ->
-    {ok, VHostSup} = rabbit_vhost_sup_sup:vhost_sup(VHost),
+    {ok, VHostSup} = rabbit_vhost_sup_sup:get_vhost_sup(VHost),
     case supervisor2:find_child(VHostSup, Type) of
         [Pid] -> Pid;
         []    -> no_pid
diff --git a/src/rabbit_vhost_sup_sup.erl b/src/rabbit_vhost_sup_sup.erl
@@ -23,7 +23,7 @@
 -export([init/1]).
 
 -export([start_link/0, start/0]).
--export([init_vhost/1, vhost_sup/1, vhost_sup/2, save_vhost_sup/3]).
+-export([init_vhost/1, get_vhost_sup/1, get_vhost_sup/2, save_vhost_sup/3]).
 -export([delete_on_all_nodes/1]).
 -export([start_on_all_nodes/1]).
 
@@ -72,7 +72,7 @@ delete_on_all_nodes(VHost) ->
     ok.
 
 stop_and_delete_vhost(VHost) ->
-    case get_vhost_sup(VHost) of
+    StopResult = case lookup_vhost_sup_record(VHost) of
         not_found -> ok;
         #vhost_sup{wrapper_pid = WrapperPid,
                    vhost_sup_pid = VHostSupPid} ->
@@ -84,13 +84,15 @@ stop_and_delete_vhost(VHost) ->
                                     [VHostSupPid, VHost]),
                     case supervisor2:terminate_child(?MODULE, WrapperPid) of
                         ok ->
-                            ets:delete(?MODULE, VHost),
-                            ok = rabbit_vhost:delete_storage(VHost);
+                            true = ets:delete(?MODULE, VHost),
+                            ok;
                         Other ->
                             Other
                     end
             end
-    end.
+    end,
+    ok = rabbit_vhost:delete_storage(VHost),
+    StopResult.
 
 %% We take an optimistic approach whan stopping a remote VHost supervisor.
 stop_and_delete_vhost(VHost, Node) when Node == node(self()) ->
@@ -106,7 +108,7 @@ stop_and_delete_vhost(VHost, Node) ->
             {error, RpcErr}
     end.
 
--spec init_vhost(rabbit_types:vhost()) -> ok.
+-spec init_vhost(rabbit_types:vhost()) -> ok | {error, {no_such_vhost, rabbit_types:vhsot()}}.
 init_vhost(VHost) ->
     case start_vhost(VHost) of
         {ok, _} -> ok;
@@ -130,30 +132,32 @@ init_vhost(VHost) ->
             end
     end.
 
--spec vhost_sup(rabbit_types:vhost(), node()) -> {ok, pid()} | {error, {no_such_vhost, rabbit_types:vhost()} | term()}.
-vhost_sup(VHost, Node) ->
-    case rabbit_misc:rpc_call(Node, rabbit_vhost_sup_sup, vhost_sup, [VHost]) of
+-type vhost_error() :: {no_such_vhost, rabbit_types:vhost()} |
+                       {vhost_supervisor_not_running, rabbit_types:vhost()}.
+
+-spec get_vhost_sup(rabbit_types:vhost(), node()) -> {ok, pid()} | {error, vhost_error() | term()}.
+get_vhost_sup(VHost, Node) ->
+    case rabbit_misc:rpc_call(Node, rabbit_vhost_sup_sup, get_vhost_sup, [VHost]) of
         {ok, Pid} when is_pid(Pid) ->
             {ok, Pid};
+        {error, Err} ->
+            {error, Err};
         {badrpc, RpcErr} ->
             {error, RpcErr}
     end.
 
--spec vhost_sup(rabbit_types:vhost()) -> {ok, pid()} | {error, {no_such_vhost, rabbit_types:vhost()}}.
-vhost_sup(VHost) ->
-    case vhost_sup_pid(VHost) of
-        no_pid ->
-            case start_vhost(VHost) of
-                {ok, Pid} ->
-                    true = is_vhost_alive(VHost),
-                    {ok, Pid};
-                {error, {no_such_vhost, VHost}} ->
-                    {error, {no_such_vhost, VHost}};
-                Error ->
-                    throw(Error)
-            end;
-        {ok, Pid} when is_pid(Pid) ->
-            {ok, Pid}
+-spec get_vhost_sup(rabbit_types:vhost()) -> {ok, pid()} | {error, vhost_error()}.
+get_vhost_sup(VHost) ->
+    case rabbit_vhost:exists(VHost) of
+        false ->
+            {error, {no_such_vhost, VHost}};
+        true ->
+            case vhost_sup_pid(VHost) of
+                no_pid ->
+                    {error, {vhost_supervisor_not_running, VHost}};
+                {ok, Pid} when is_pid(Pid) ->
+                    {ok, Pid}
+            end
     end.
 
 -spec start_vhost(rabbit_types:vhost(), node()) -> {ok, pid()} | {error, term()}.
@@ -181,7 +185,7 @@ start_vhost(VHost) ->
 is_vhost_alive(VHost) ->
 %% A vhost is considered alive if it's supervision tree is alive and
 %% saved in the ETS table
-    case get_vhost_sup(VHost) of
+    case lookup_vhost_sup_record(VHost) of
         #vhost_sup{wrapper_pid = WrapperPid,
                    vhost_sup_pid = VHostSupPid,
                    vhost_process_pid = VHostProcessPid}
@@ -210,16 +214,16 @@ save_vhost_process(VHost, VHostProcessPid) ->
                               {#vhost_sup.vhost_process_pid, VHostProcessPid}),
     ok.
 
--spec get_vhost_sup(rabbit_types:vhost()) -> #vhost_sup{}.
-get_vhost_sup(VHost) ->
+-spec lookup_vhost_sup_record(rabbit_types:vhost()) -> #vhost_sup{} | not_found.
+lookup_vhost_sup_record(VHost) ->
     case ets:lookup(?MODULE, VHost) of
         [] -> not_found;
         [#vhost_sup{} = VHostSup] -> VHostSup
     end.
 
 -spec vhost_sup_pid(rabbit_types:vhost()) -> no_pid | {ok, pid()}.
 vhost_sup_pid(VHost) ->
-    case get_vhost_sup(VHost) of
+    case lookup_vhost_sup_record(VHost) of
         not_found ->
             no_pid;
         #vhost_sup{vhost_sup_pid = Pid} = VHostSup ->
diff --git a/src/rabbit_vhost_sup_wrapper.erl b/src/rabbit_vhost_sup_wrapper.erl
@@ -29,7 +29,12 @@
 start_link(VHost) ->
     %% Using supervisor, because supervisor2 does not stop a started child when
     %% another one fails to start. Bug?
-    supervisor:start_link(?MODULE, [VHost]).
+    case rabbit_vhost_sup_sup:get_vhost_sup(VHost) of
+        {ok, Pid}  ->
+            {error, {already_started, Pid}};
+        {error, _} ->
+            supervisor:start_link(?MODULE, [VHost])
+    end.
 
 init([VHost]) ->
     %% 2 restarts in 5 minutes. One per message store.
diff --git a/test/dynamic_ha_SUITE.erl b/test/dynamic_ha_SUITE.erl
@@ -57,13 +57,15 @@ groups() ->
       {clustered, [], [
           {cluster_size_2, [], [
               vhost_deletion,
-              promote_on_shutdown
+              promote_on_shutdown,
+              queue_survive_adding_dead_vhost_mirror
             ]},
           {cluster_size_3, [], [
               change_policy,
               rapid_change,
               nodes_policy_should_pick_master_from_its_params,
-              promote_slave_after_standalone_restart
+              promote_slave_after_standalone_restart,
+              queue_survive_adding_dead_vhost_mirror
               % FIXME: Re-enable those tests when the know issues are
               % fixed.
               % failing_random_policies,
@@ -218,6 +220,19 @@ rapid_loop(Config, Node, MRef) ->
             rapid_loop(Config, Node, MRef)
     end.
 
+queue_survive_adding_dead_vhost_mirror(Config) ->
+    rabbit_ct_broker_helpers:force_vhost_failure(Config, 1, <<"/">>),
+    NodeA = rabbit_ct_broker_helpers:get_node_config(Config, 0, nodename),
+    ChA = rabbit_ct_client_helpers:open_channel(Config, NodeA),
+    QName = <<"queue_survive_adding_dead_vhost_mirror-q-1">>,
+    amqp_channel:call(ChA, #'queue.declare'{queue = QName}),
+    Q = find_queue(QName, NodeA),
+    Pid = proplists:get_value(pid, Q),
+    rabbit_ct_broker_helpers:set_ha_policy_all(Config),
+    %% Queue should not fail
+    Q1 = find_queue(QName, NodeA),
+    Pid = proplists:get_value(pid, Q1).
+
 %% Vhost deletion needs to successfully tear down policies and queues
 %% with policies. At least smoke-test that it doesn't blow up.
 vhost_deletion(Config) ->
diff --git a/test/vhost_SUITE.erl b/test/vhost_SUITE.erl
