Merge remote-tracking branch 'origin/release-v43' into release-v44

apps/emqx_rule_engine/src/emqx_rule_engine.erl

@@ -79,8 +79,6 @@
              , action_instance_params/0
              ]).
 
--define(T_RETRY, 60000).
-
 %% redefine this macro to confine the appup scope
 -undef(RAISE).
 -define(RAISE(_EXP_, _ERROR_CONTEXT_),
@@ -684,8 +682,7 @@ init_resource_with_retrier(Module, OnCreate, ResId, Config) ->
                                     status = #{is_alive => true}},
         emqx_rule_registry:add_resource_params(ResParams)
     catch Class:Reason:ST ->
-        Interval = persistent_term:get({emqx_rule_engine, resource_restart_interval}, ?T_RETRY),
+        emqx_rule_monitor:ensure_resource_retrier(ResId),
-        emqx_rule_monitor:ensure_resource_retrier(ResId, Interval),
         erlang:raise(Class, {init_resource, Reason}, ST)
     end.
 

apps/emqx_rule_engine/src/emqx_rule_monitor.erl

@@ -32,10 +32,18 @@
 -export([ start_link/0
         , stop/0
         , async_refresh_resources_rules/0
-        , ensure_resource_retrier/2
+        , ensure_resource_retrier/1
         , retry_loop/3
         ]).
 
+%% fot test
+-export([ put_retry_interval/1
+        , get_retry_interval/0
+        , erase_retry_interval/0
+        ]).
+
+-define(T_RETRY, 60000).
+
 start_link() ->
     gen_server:start_link({local, ?MODULE}, ?MODULE, [], []).
 
@@ -46,10 +54,22 @@ init([]) ->
     _ = erlang:process_flag(trap_exit, true),
     {ok, #{retryers => #{}}}.
 
+put_retry_interval(I) when is_integer(I) andalso I >= 10 ->
+    _ = persistent_term:put({?MODULE, resource_restart_interval}, I),
+    ok.
+
+erase_retry_interval() ->
+    _ = persistent_term:erase({?MODULE, resource_restart_interval}),
+    ok.
+
+get_retry_interval() ->
+    persistent_term:get({?MODULE, resource_restart_interval}, ?T_RETRY).
+
 async_refresh_resources_rules() ->
     gen_server:cast(?MODULE, async_refresh).
 
-ensure_resource_retrier(ResId, Interval) ->
+ensure_resource_retrier(ResId) ->
+    Interval = get_retry_interval(),
     gen_server:cast(?MODULE, {create_restart_handler, resource, ResId, Interval}).
 
 handle_call(_Msg, _From, State) ->
@@ -111,11 +131,12 @@ update_object(Tag, Obj, Retryer, State) ->
     }.
 
 create_restart_handler(Tag, Obj, Interval) ->
-    ?LOG(info, "keep restarting ~p ~p, interval: ~p", [Tag, Obj, Interval]),
+    ?LOG(info, "starting_a_retry_loop for ~p ~p, with delay interval: ~p", [Tag, Obj, Interval]),
     %% spawn a dedicated process to handle the restarting asynchronously
     spawn_link(?MODULE, retry_loop, [Tag, Obj, Interval]).
 
 retry_loop(resource, ResId, Interval) ->
+    timer:sleep(Interval),
     case emqx_rule_registry:find_resource(ResId) of
         {ok, #resource{type = Type, config = Config}} ->
             try
@@ -124,10 +145,15 @@ retry_loop(resource, ResId, Interval) ->
                 ok = emqx_rule_engine:init_resource(M, F, ResId, Config),
                 refresh_and_enable_rules_of_resource(ResId)
             catch
-                Err:Reason:ST ->
+                Err:Reason:Stacktrace ->
-                    ?LOG(warning, "init_resource failed: ~p, ~0p",
+                    %% do not log stacktrace if it's a throw
-                        [{Err, Reason}, ST]),
+                    LogContext =
-                    timer:sleep(Interval),
+                        case Err of
+                            throw -> Reason;
+                            _ -> {Reason, Stacktrace}
+                        end,
+                    ?LOG_SENSITIVE(warning, "init_resource_retry_failed ~p, ~0p", [ResId, LogContext]),
+                    %% keep looping
                     ?MODULE:retry_loop(resource, ResId, Interval)
             end;
         not_found ->

apps/emqx_rule_engine/test/emqx_rule_monitor_SUITE.erl

@@ -48,7 +48,7 @@ end_per_suite(_Config) ->
     ok.
 
 init_per_testcase(t_restart_resource, Config) ->
-    persistent_term:put({emqx_rule_engine, resource_restart_interval}, 100),
+    emqx_rule_monitor:put_retry_interval(100),
     Opts = [public, named_table, set, {read_concurrency, true}],
     _ = ets:new(?RES_PARAMS_TAB, [{keypos, #resource_params.id}|Opts]),
     ets:new(t_restart_resource, [named_table, public]),
@@ -77,7 +77,6 @@ init_per_testcase(_, Config) ->
     Config.
 
 end_per_testcase(t_restart_resource, Config) ->
-    persistent_term:put({emqx_rule_engine, resource_restart_interval}, 60000),
     ets:delete(t_restart_resource),
     common_end_per_testcases(),
     Config;
@@ -91,7 +90,9 @@ end_per_testcase(_, Config) ->
 
 common_init_per_testcase() ->
     {ok, _} = emqx_rule_monitor:start_link().
+
 common_end_per_testcases() ->
+    emqx_rule_monitor:erase_retry_interval(),
     emqx_rule_monitor:stop().
 
 t_restart_resource(_) ->

changes/v4.3.22-en.md

@@ -36,13 +36,15 @@
   For example: `acl_order = jwt,http`, this will make sure `jwt` is always checked before `http`,
   meaning if JWT is not found (or no `acl` cliam) for a client, then the ACL check will fallback to use the HTTP backend.
 
-
 - Added configurations to enable more `client.disconnected` events (and counter bumps) [#9267](https://github.com/emqx/emqx/pull/9267).
   Prior to this change, the `client.disconnected` event (and counter bump) is triggered when a client
   performs a 'normal' disconnect, or is 'kicked' by system admin, but NOT triggered when a
   stale connection had to be 'discarded' (for clean session) or 'takenover' (for non-clean session).
   Now it is possible to set configs `broker.client_disconnect_discarded` and `broker.client_disconnect_takenover` to `on` to enable the event in these scenarios.
 
+- For Rule-Engine resource creation failure, delay before the first retry [#9313](https://github.com/emqx/emqx/pull/9313).
+  Prior to this change, the retry delay was added *after* the retry failure.
+
 ## Bug fixes
 
 - Fix that after uploading a backup file with an non-ASCII filename, HTTP API `GET /data/export` fails with status code 500 [#9224](https://github.com/emqx/emqx/pull/9224).

changes/v4.3.22-zh.md

@@ -37,6 +37,9 @@
   但不会在旧 session 被废弃 (clean_session = true) 或旧 session 被接管 (clean_session = false) 时被触发。
   可将 `broker.client_disconnect_discarded` 和 `broker.client_disconnect_takovered` 选项设置为 `on` 来启用此场景下的客户端断连事件。
 
+- 规则引擎资源创建失败后，第一次重试前增加一个延迟 [#9313](https://github.com/emqx/emqx/pull/9313)。
+  在此之前，重试的延迟发生在重试失败之后。
+
 ## 修复
 
 - 修复若上传的备份文件名中包含非 ASCII 字符，`GET /data/export` HTTP 接口返回 500 错误 [#9224](https://github.com/emqx/emqx/pull/9224)。