Migrating to per-vhost supervisor message store.

Support reading/saving recovery terms from global storage to per-vhost storages.

Author: Daniil Fedotov

src/rabbit_amqqueue_sup_sup.erl

@@ -19,7 +19,8 @@
 -behaviour(supervisor2).
 
 -export([start_link/0, start_queue_process/3]).
--export([start_for_vhost/1, stop_for_vhost/1, find_for_vhost/2]).
+-export([start_for_vhost/1, stop_for_vhost/1,
+         find_for_vhost/2, find_for_vhost/1]).
 
 -export([init/1]).
 
@@ -50,13 +51,19 @@ init([]) ->
           [{rabbit_amqqueue_sup, {rabbit_amqqueue_sup, start_link, []},
             temporary, ?SUPERVISOR_WAIT, supervisor, [rabbit_amqqueue_sup]}]}}.
 
+-spec find_for_vhost(rabbit_types:vhost()) -> {ok, pid()} | {error, term()}.
+find_for_vhost(VHost) ->
+    find_for_vhost(VHost, node()).
+
+-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),
     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) ->
     {ok, VHostSup} = rabbit_vhost_sup_sup:vhost_sup(VHost),
     supervisor2:start_child(
@@ -65,6 +72,7 @@ start_for_vhost(VHost) ->
          {rabbit_amqqueue_sup_sup, start_link, []},
          transient, infinity, supervisor, [rabbit_amqqueue_sup_sup]}).
 
+-spec stop_for_vhost(rabbit_types:vhost()) -> ok.
 stop_for_vhost(VHost) ->
     {ok, VHostSup} = rabbit_vhost_sup_sup:vhost_sup(VHost),
     ok = supervisor2:terminate_child(VHostSup, rabbit_amqqueue_sup_sup),

src/rabbit_mirror_queue_slave.erl

@@ -194,12 +194,13 @@ stop_pending_slaves(QName, Pids) ->
     [begin
          rabbit_mirror_queue_misc:log_warning(
            QName, "Detected stale HA slave, stopping it: ~p~n", [Pid]),
-        %TODO: per-vhost supervisor
          case erlang:process_info(Pid, dictionary) of
              undefined -> ok;
              {dictionary, Dict} ->
+                 Vhost = QName#resource.virtual_host,
+                 {ok, AmqQSup} = rabbit_amqqueue_sup_sup:find_for_vhost(Vhost),
                  case proplists:get_value('$ancestors', Dict) of
-                     [Sup, rabbit_amqqueue_sup_sup | _] ->
+                     [Sup, AmqQSup | _] ->
                          exit(Sup, kill),
                          exit(Pid, kill);
                      _ ->

src/rabbit_queue_index.erl

@@ -26,7 +26,10 @@
 -export([scan_queue_segments/3]).
 
 %% Migrates from global to per-vhost message stores
--export([move_to_per_vhost_stores/1, update_recovery_term/2]).
+-export([move_to_per_vhost_stores/1,
+         update_recovery_term/2,
+         read_global_recovery_terms/1,
+         cleanup_global_recovery_terms/0]).
 
 -define(CLEAN_FILENAME, "clean.dot").
 
@@ -516,6 +519,33 @@ start(VHost, DurableQueueNames) ->
     OrderedTerms = lists:reverse(DurableTerms),
     {OrderedTerms, {fun queue_index_walker/1, {start, DurableQueueNames}}}.
 
+
+read_global_recovery_terms(DurableQueueNames) ->
+    ok = rabbit_recovery_terms:open_global_table(),
+
+    DurableTerms =
+        lists:foldl(
+          fun(QName, RecoveryTerms) ->
+                  DirName = queue_name_to_dir_name(QName),
+                  RecoveryInfo = case rabbit_recovery_terms:read_global(DirName) of
+                                     {error, _}  -> non_clean_shutdown;
+                                     {ok, Terms} -> Terms
+                                 end,
+                  [RecoveryInfo | RecoveryTerms]
+          end, [], DurableQueueNames),
+
+    ok = rabbit_recovery_terms:close_global_table(),
+    %% The backing queue interface requires that the queue recovery terms
+    %% which come back from start/1 are in the same order as DurableQueueNames
+    OrderedTerms = lists:reverse(DurableTerms),
+    {OrderedTerms, {fun queue_index_walker/1, {start, DurableQueueNames}}}.
+
+cleanup_global_recovery_terms() ->
+    rabbit_file:recursive_delete([filename:join([queues_base_dir(), "queues"])]),
+    rabbit_recovery_terms:delete_global_table(),
+    ok.
+
+
 stop(VHost) -> rabbit_recovery_terms:stop(VHost).
 
 all_queue_directory_names(VHost) ->

src/rabbit_recovery_terms.erl

@@ -28,6 +28,9 @@
          terminate/2, code_change/3]).
 
 -export([upgrade_recovery_terms/0, persistent_bytes/0]).
+-export([open_global_table/0, close_global_table/0,
+         read_global/1, delete_global_table/0]).
+-export([open_table/1, close_table/1]).
 
 -rabbit_upgrade({upgrade_recovery_terms, local, []}).
 -rabbit_upgrade({persistent_bytes, local, [upgrade_recovery_terms]}).
@@ -119,12 +122,19 @@ open_global_table() ->
     File = filename:join(rabbit_mnesia:dir(), "recovery.dets"),
     {ok, _} = dets:open_file(?MODULE, [{file,      File},
                                        {ram_file,  true},
-                                       {auto_save, infinity}]).
+                                       {auto_save, infinity}]),
+    ok.
 
 close_global_table() ->
     ok = dets:sync(?MODULE),
     ok = dets:close(?MODULE).
 
+read_global(DirBaseName) ->
+    read(?MODULE, DirBaseName).
+
+delete_global_table() ->
+    file:delete(filename:join(rabbit_mnesia:dir(), "recovery.dets")).
+
 %%----------------------------------------------------------------------------
 
 init([VHost]) ->

src/rabbit_variable_queue.erl

@@ -2753,7 +2753,7 @@ transform_store(Store, TransformFun) ->
 
 move_messages_to_vhost_store() ->
     case list_persistent_queues() of
-        []     -> ok;
+        % []     -> ok;
         Queues -> move_messages_to_vhost_store(Queues)
     end.
 
@@ -2776,21 +2776,23 @@ move_messages_to_vhost_store(Queues) ->
     VHosts = rabbit_vhost:list(),
 
     %% New store should not be recovered.
-    ok = start_new_store(VHosts),
+    NewMsgStore = start_new_store(VHosts),
+    %% Recovery terms should be started for all vhosts for new store.
+    [{ok, _} = rabbit_recovery_terms:open_table(VHost) || VHost <- VHosts],
+
     MigrationBatchSize = application:get_env(rabbit, queue_migration_batch_size,
                                              ?QUEUE_MIGRATION_BATCH_SIZE),
     in_batches(MigrationBatchSize,
-        {rabbit_variable_queue, migrate_queue, [OldStore]},
+        {rabbit_variable_queue, migrate_queue, [OldStore, NewMsgStore]},
         QueuesWithTerms,
         "message_store upgrades: Migrating batch ~p of ~p queues. Out of total ~p ~n",
         "message_store upgrades: Batch ~p of ~p queues migrated ~n. ~p total left"),
 
     log_upgrade("Message store migration finished"),
-    delete_old_store(OldStore),
-
-    ok = rabbit_queue_index:stop(),
-    ok = stop_new_store(VHosts),
-    ok.
+    ok = delete_old_store(OldStore),
+    ok = rabbit_queue_index:cleanup_global_recovery_terms(),
+    [ok= rabbit_recovery_terms:close_table(VHost) || VHost <- VHosts],
+    ok = stop_new_store(NewMsgStore).
 
 in_batches(Size, MFA, List, MessageStart, MessageEnd) ->
     in_batches(Size, 1, MFA, List, MessageStart, MessageEnd).
@@ -2816,12 +2818,14 @@ in_batches(Size, BatchNum, MFA, List, MessageStart, MessageEnd) ->
     rabbit_log:info(MessageEnd, [BatchNum, Size, length(Tail)]),
     in_batches(Size, BatchNum + 1, MFA, Tail, MessageStart, MessageEnd).
 
-migrate_queue({QueueName = #resource{virtual_host = VHost, name = Name}, RecoveryTerm}, OldStore) ->
+migrate_queue({QueueName = #resource{virtual_host = VHost, name = Name},
+               RecoveryTerm},
+              OldStore, NewStore) ->
     log_upgrade_verbose(
         "Migrating messages in queue ~s in vhost ~s to per-vhost message store~n",
         [Name, VHost]),
     OldStoreClient = get_global_store_client(OldStore),
-    NewStoreClient = get_per_vhost_store_client(QueueName),
+    NewStoreClient = get_per_vhost_store_client(QueueName, NewStore),
     %% WARNING: During scan_queue_segments queue index state is being recovered
     %% and terminated. This can cause side effects!
     rabbit_queue_index:scan_queue_segments(
@@ -2857,11 +2861,10 @@ migrate_message(MsgId, OldC, NewC) ->
         _ -> OldC
     end.
 
-get_per_vhost_store_client(#resource{virtual_host = VHost}) ->
-    rabbit_vhost_msg_store:client_init(VHost, ?PERSISTENT_MSG_STORE,
-                                       rabbit_guid:gen(),
-                                       fun(_,_) -> ok end,
-                                       fun() -> ok end).
+get_per_vhost_store_client(#resource{virtual_host = VHost}, NewStore) ->
+    {VHost, StorePid} = lists:keyfind(VHost, 1, NewStore),
+    rabbit_msg_store:client_init(StorePid, rabbit_guid:gen(),
+                                 fun(_,_) -> ok end, fun() -> ok end).
 
 get_global_store_client(OldStore) ->
     rabbit_msg_store:client_init(OldStore,
@@ -2902,19 +2905,22 @@ run_old_persistent_store(Refs, StartFunState) ->
 
 start_new_store(VHosts) ->
     %% Ensure vhost supervisor is started, so we can add vhsots to it.
-    %% TODO: Start message store for vhost without a supervisor.
-    lists:foreach(fun(VHost) ->
-        % Start persistent store without recovery.
-        {ok, _} = rabbit_vhost_msg_store:start(VHost, ?PERSISTENT_MSG_STORE, undefined, ?EMPTY_START_FUN_STATE)
+    lists:map(fun(VHost) ->
+        VHostDir = rabbit_vhost:msg_store_dir_path(VHost),
+        {ok, Pid} = rabbit_msg_store:start_link(?PERSISTENT_MSG_STORE,
+                                                VHostDir,
+                                                undefined,
+                                                ?EMPTY_START_FUN_STATE),
+        {VHost, Pid}
     end,
-    VHosts),
-    ok.
+    VHosts).
 
-stop_new_store(VHosts) ->
-    lists:foreach(fun(VHost) ->
-        ok = rabbit_vhost_msg_store:stop(VHost, ?PERSISTENT_MSG_STORE)
+stop_new_store(NewStore) ->
+    lists:foreach(fun({_VHost, StorePid}) ->
+        unlink(StorePid),
+        exit(StorePid, shutdown)
     end,
-    VHosts),
+    NewStore),
     ok.
 
 delete_old_store(OldStore) ->
@@ -2923,7 +2929,8 @@ delete_old_store(OldStore) ->
         [filename:join([rabbit_mnesia:dir(), ?PERSISTENT_MSG_STORE])]),
     %% Delete old transient store as well
     rabbit_file:recursive_delete(
-        [filename:join([rabbit_mnesia:dir(), ?TRANSIENT_MSG_STORE])]).
+        [filename:join([rabbit_mnesia:dir(), ?TRANSIENT_MSG_STORE])]),
+    ok.
 
 log_upgrade(Msg) ->
     log_upgrade(Msg, []).

src/rabbit_vhost_sup_sup.erl

@@ -64,7 +64,6 @@ stop_and_delete_vhost(VHost) ->
     end.
 
 delete_on_all_nodes(VHost) ->
-%% TODO: failing nodes
     [ stop_and_delete_vhost(VHost, Node) || Node <- rabbit_nodes:all_running() ],
     ok.
 

test/unit_inbroker_SUITE.erl

@@ -829,7 +829,7 @@ bq_queue_recover1(Config) ->
         rabbit_amqqueue:declare(queue_name(Config, <<"bq_queue_recover-q">>),
                                 true, false, [], none, <<"acting-user">>),
     publish_and_confirm(Q, <<>>, Count),
-%% TODO: per-vhost supervisor
+
     SupPid = rabbit_ct_broker_helpers:get_queue_sup_pid(Q),
     true = is_pid(SupPid),
     exit(SupPid, kill),