Fix replication job start_link.

We don't want to register replication jobs with the module name as that
prevents us from running more than one job at a time.

This was introduced when we removed the dependence on global registrations and
switched to using pg instead.

Even more embarrassing is we didn't have tests to check multiple replication
jobs running at the same time.

src/couch_replicator/src/couch_replicator_scheduler_job.erl

@@ -87,7 +87,7 @@ start_link(#rep{id = Id = {BaseId, Ext}, source = Src, target = Tgt} = Rep) ->
     Target = couch_replicator_api_wrap:db_uri(Tgt),
     case couch_replicator_pg:should_start(Id, node()) of
         yes ->
-            case gen_server:start_link({local, ?MODULE}, ?MODULE, Rep, []) of
+            case gen_server:start_link(?MODULE, Rep, []) of
                 {ok, Pid} ->
                     couch_replicator_pg:join(Id, Pid),
                     {ok, Pid};

src/couch_replicator/test/eunit/couch_replicator_error_reporting_tests.erl

@@ -30,6 +30,7 @@ error_reporting_test_() ->
             ?TDEF_FE(t_fail_changes_manager),
             ?TDEF_FE(t_fail_changes_reader_proc),
             ?TDEF_FE(t_dont_start_duplicate_job),
+            ?TDEF_FE(t_can_start_multiple_jobs),
             ?TDEF_FE(t_stop_duplicate_job)
         ]
     }.
@@ -160,6 +161,30 @@ t_dont_start_duplicate_job({_Ctx, {Source, Target}}) ->
     ExpectErr = {error, {already_started, Pid}},
     ?assertEqual(ExpectErr, couch_replicator_scheduler_job:start_link(Rep)).
 
+t_can_start_multiple_jobs({_Ctx, {Source, Target1}}) ->
+    Target2 = couch_replicator_test_helper:setup_db(),
+    populate_db(Source, 1, 5),
+
+    {ok, RepId1} = replicate(Source, Target1),
+    {ok, RepId2} = replicate(Source, Target2),
+    RepPid1 = couch_replicator_test_helper:get_pid(RepId1),
+    RepPid2 = couch_replicator_test_helper:get_pid(RepId2),
+    ?assert(is_pid(RepPid1)),
+    ?assert(is_pid(RepPid2)),
+
+    ?assert(is_process_alive(RepPid1)),
+    ?assert(is_process_alive(RepPid2)),
+
+    wait_target_in_sync(Source, Target1),
+    wait_target_in_sync(Source, Target2),
+
+    ?assert(is_process_alive(RepPid1)),
+    ?assert(is_process_alive(RepPid2)),
+
+    exit(RepPid1, kill),
+    exit(RepPid2, kill),
+    couch_replicator_test_helper:teardown_db(Target2).
+
 t_stop_duplicate_job({_Ctx, {Source, Target}}) ->
     {ok, RepId} = replicate(Source, Target),
     wait_target_in_sync(Source, Target),