diff options
Diffstat (limited to 'lib/mnesia/src')
-rw-r--r-- | lib/mnesia/src/mnesia_controller.erl | 29 | ||||
-rw-r--r-- | lib/mnesia/src/mnesia_dumper.erl | 60 | ||||
-rw-r--r-- | lib/mnesia/src/mnesia_loader.erl | 7 | ||||
-rw-r--r-- | lib/mnesia/src/mnesia_locker.erl | 10 | ||||
-rw-r--r-- | lib/mnesia/src/mnesia_monitor.erl | 4 | ||||
-rw-r--r-- | lib/mnesia/src/mnesia_recover.erl | 31 |
6 files changed, 110 insertions, 31 deletions
diff --git a/lib/mnesia/src/mnesia_controller.erl b/lib/mnesia/src/mnesia_controller.erl index 5a9bae54da..aa72de7594 100644 --- a/lib/mnesia/src/mnesia_controller.erl +++ b/lib/mnesia/src/mnesia_controller.erl @@ -51,6 +51,7 @@ force_load_table/1, async_dump_log/1, sync_dump_log/1, + snapshot_dcd/1, connect_nodes/1, connect_nodes/2, wait_for_schema_commit_lock/0, @@ -139,7 +140,8 @@ max_loaders() -> -record(block_controller, {owner}). -record(dump_log, {initiated_by, - opt_reply_to + opt_reply_to, + operation = dump_log }). -record(net_load, {table, @@ -201,6 +203,15 @@ async_dump_log(InitBy) -> ?SERVER_NAME ! {async_dump_log, InitBy}, ok. +snapshot_dcd(Tables) when is_list(Tables) -> + case [T || T <- Tables, + mnesia_lib:storage_type_at_node(node(), T) =/= disc_copies] of + [] -> + call({snapshot_dcd, Tables}); + BadTabs -> + {error, {not_disc_copies, BadTabs}} + end. + %% Wait for tables to be active %% If needed, we will wait for Mnesia to start %% If Mnesia stops, we will wait for Mnesia to restart @@ -646,6 +657,15 @@ handle_call({sync_dump_log, InitBy}, From, State) -> State2 = add_worker(Worker, State), noreply(State2); +handle_call({snapshot_dcd, Tables}, From, State) -> + Worker = #dump_log{initiated_by = user, + opt_reply_to = From, + operation = fun() -> + mnesia_dumper:snapshot_dcd(Tables) + end}, + State2 = add_worker(Worker, State), + noreply(State2); + handle_call(wait_for_schema_commit_lock, From, State) -> Worker = #schema_commit_lock{owner = From}, State2 = add_worker(Worker, State), @@ -2089,7 +2109,12 @@ start_remote_sender(Node, Tab, Receiver, Storage) -> dump_and_reply(ReplyTo, Worker) -> %% No trap_exit, die intentionally instead - Res = mnesia_dumper:opt_dump_log(Worker#dump_log.initiated_by), + Res = case Worker#dump_log.operation of + dump_log -> + mnesia_dumper:opt_dump_log(Worker#dump_log.initiated_by); + F when is_function(F, 0) -> + F() + end, ReplyTo ! #dumper_done{worker_pid = self(), worker_res = Res}, unlink(ReplyTo), diff --git a/lib/mnesia/src/mnesia_dumper.erl b/lib/mnesia/src/mnesia_dumper.erl index 14665797a0..509b765dee 100644 --- a/lib/mnesia/src/mnesia_dumper.erl +++ b/lib/mnesia/src/mnesia_dumper.erl @@ -34,11 +34,13 @@ -export([ get_log_writes/0, incr_log_writes/0, + needs_dump_ets/1, raw_dump_table/2, raw_named_dump_table/2, start_regulator/0, opt_dump_log/1, - update/3 + update/3, + snapshot_dcd/1 ]). %% Internal stuff @@ -99,6 +101,19 @@ opt_dump_log(InitBy) -> end, perform_dump(InitBy, Reg). +snapshot_dcd(Tables) -> + lists:foreach( + fun(Tab) -> + case mnesia_lib:storage_type_at_node(node(), Tab) of + disc_copies -> + mnesia_log:ets2dcd(Tab); + _ -> + %% Storage type was checked before queueing the op, though + skip + end + end, Tables), + dumped. + %% Scan for decisions perform_dump(InitBy, Regulator) when InitBy == scan_decisions -> ?eval_debug_fun({?MODULE, perform_dump}, [InitBy]), @@ -981,28 +996,10 @@ open_files(_Tab, _Storage, _UpdateInPlace, _InitBy) -> false. open_disc_copies(Tab, InitBy) -> - DclF = mnesia_lib:tab2dcl(Tab), - DumpEts = - case file:read_file_info(DclF) of - {error, enoent} -> - false; - {ok, DclInfo} -> - DcdF = mnesia_lib:tab2dcd(Tab), - case file:read_file_info(DcdF) of - {error, Reason} -> - mnesia_lib:dbg_out("File ~p info_error ~p ~n", - [DcdF, Reason]), - true; - {ok, DcdInfo} -> - Mul = case ?catch_val(dc_dump_limit) of - {'EXIT', _} -> ?DumpToEtsMultiplier; - Val -> Val - end, - DcdInfo#file_info.size =< (DclInfo#file_info.size * Mul) - end - end, + DumpEts = needs_dump_ets(Tab), if DumpEts == false; InitBy == startup -> + DclF = mnesia_lib:tab2dcl(Tab), mnesia_log:open_log({?MODULE,Tab}, mnesia_log:dcl_log_header(), DclF, @@ -1017,6 +1014,27 @@ open_disc_copies(Tab, InitBy) -> false end. +needs_dump_ets(Tab) -> + DclF = mnesia_lib:tab2dcl(Tab), + case file:read_file_info(DclF) of + {error, enoent} -> + false; + {ok, DclInfo} -> + DcdF = mnesia_lib:tab2dcd(Tab), + case file:read_file_info(DcdF) of + {error, Reason} -> + mnesia_lib:dbg_out("File ~p info_error ~p ~n", + [DcdF, Reason]), + true; + {ok, DcdInfo} -> + Mul = case ?catch_val(dc_dump_limit) of + {'EXIT', _} -> ?DumpToEtsMultiplier; + Val -> Val + end, + DcdInfo#file_info.size =< (DclInfo#file_info.size * Mul) + end + end. + %% Always opens the dcl file for writing overriding already_dumped %% mechanismen, used for schema transactions. open_dcl(Tab) -> diff --git a/lib/mnesia/src/mnesia_loader.erl b/lib/mnesia/src/mnesia_loader.erl index 530317bcdd..cbb3d7e430 100644 --- a/lib/mnesia/src/mnesia_loader.erl +++ b/lib/mnesia/src/mnesia_loader.erl @@ -69,9 +69,10 @@ do_get_disc_copy2(Tab, Reason, Storage, Type) when Storage == disc_copies -> ignore; _ -> mnesia_monitor:mktab(Tab, Args), - Count = mnesia_log:dcd2ets(Tab, Repair), - case ets:info(Tab, size) of - X when X < Count * 4 -> + _Count = mnesia_log:dcd2ets(Tab, Repair), + case mnesia_monitor:get_env(dump_disc_copies_at_startup) + andalso mnesia_dumper:needs_dump_ets(Tab) of + true -> ok = mnesia_log:ets2dcd(Tab); _ -> ignore diff --git a/lib/mnesia/src/mnesia_locker.erl b/lib/mnesia/src/mnesia_locker.erl index e27396731f..1efb939e00 100644 --- a/lib/mnesia/src/mnesia_locker.erl +++ b/lib/mnesia/src/mnesia_locker.erl @@ -982,8 +982,14 @@ sticky_flush(Ns=[Node | Tail], Store) -> flush_remaining([], _SkipNode, Res) -> del_debug(), exit(Res); -flush_remaining([SkipNode | Tail ], SkipNode, Res) -> - flush_remaining(Tail, SkipNode, Res); +flush_remaining(Ns=[SkipNode | Tail ], SkipNode, Res) -> + add_debug(Ns), + receive + {?MODULE, SkipNode, _} -> + flush_remaining(Tail, SkipNode, Res) + after 0 -> + flush_remaining(Tail, SkipNode, Res) + end; flush_remaining(Ns=[Node | Tail], SkipNode, Res) -> add_debug(Ns), receive diff --git a/lib/mnesia/src/mnesia_monitor.erl b/lib/mnesia/src/mnesia_monitor.erl index 6fc1a394a6..a0e0e630ec 100644 --- a/lib/mnesia/src/mnesia_monitor.erl +++ b/lib/mnesia/src/mnesia_monitor.erl @@ -664,6 +664,7 @@ env() -> backup_module, debug, dir, + dump_disc_copies_at_startup, dump_log_load_regulation, dump_log_time_threshold, dump_log_update_in_place, @@ -692,6 +693,8 @@ default_env(debug) -> default_env(dir) -> Name = lists:concat(["Mnesia.", node()]), filename:absname(Name); +default_env(dump_disc_copies_at_startup) -> + true; default_env(dump_log_load_regulation) -> false; default_env(dump_log_time_threshold) -> @@ -741,6 +744,7 @@ do_check_type(debug, trace) -> trace; do_check_type(debug, true) -> debug; do_check_type(debug, verbose) -> verbose; do_check_type(dir, V) -> filename:absname(V); +do_check_type(dump_disc_copies_at_startup, B) -> bool(B); do_check_type(dump_log_load_regulation, B) -> bool(B); do_check_type(dump_log_time_threshold, I) when is_integer(I), I > 0 -> I; do_check_type(dump_log_update_in_place, B) -> bool(B); diff --git a/lib/mnesia/src/mnesia_recover.erl b/lib/mnesia/src/mnesia_recover.erl index b6492707e2..eeb4fa0ced 100644 --- a/lib/mnesia/src/mnesia_recover.erl +++ b/lib/mnesia/src/mnesia_recover.erl @@ -689,12 +689,29 @@ handle_call({connect_nodes, Ns}, From, State) -> %% called from handle_info gen_server:reply(From, {[], AlreadyConnected}), {noreply, State}; - GoodNodes -> + ProbablyGoodNodes -> %% Now we have agreed upon a protocol with some new nodes - %% and we may use them when we recover transactions + %% and we may use them when we recover transactions. + %% + %% Just in case Mnesia was stopped on some of those nodes + %% between the protocol negotiation and now, we check one + %% more time the state of Mnesia. + %% + %% Of course, there is still a chance that mnesia_down + %% events occur during this check and we miss them. To + %% prevent it, handle_cast({mnesia_down, ...}, ...) removes + %% the down node again, in addition to mnesia_down/1. + %% + %% See a comment in handle_cast({mnesia_down, ...}, ...). + Verify = fun(N) -> + Run = mnesia_lib:is_running(N), + Run =:= yes orelse Run =:= starting + end, + GoodNodes = [N || N <- ProbablyGoodNodes, Verify(N)], + mnesia_lib:add_list(recover_nodes, GoodNodes), cast({announce_all, GoodNodes}), - case get_master_nodes(schema) of + case get_master_nodes(schema) of [] -> Context = starting_partitioned_network, mnesia_monitor:detect_inconcistency(GoodNodes, Context); @@ -842,6 +859,14 @@ handle_cast({what_decision, Node, OtherD}, State) -> {noreply, State}; handle_cast({mnesia_down, Node}, State) -> + %% The node was already removed from recover_nodes in mnesia_down/1, + %% but we do it again here in the mnesia_recover process, in case + %% another event incorrectly added it back. This can happen during + %% Mnesia startup which takes time betweenthe connection, the + %% protocol negotiation and the merge of the schema. + %% + %% See a comment in handle_call({connect_nodes, ...), ...). + mnesia_lib:del(recover_nodes, Node), case State#state.unclear_decision of undefined -> {noreply, State}; |