aboutsummaryrefslogtreecommitdiffstats
path: root/lib/mnesia
diff options
context:
space:
mode:
authorDan Gudmundsson <dgud@erlang.org>2011-03-25 13:14:35 +0100
committerDan Gudmundsson <dgud@erlang.org>2011-04-04 14:26:14 +0200
commit02e09f135b2ef75ff440ee008e3db437b8fc3ff7 (patch)
tree9fd8ae0c522df77248402818aab87d931b097d29 /lib/mnesia
parent91b2e57ea0e3ab794d4b57a12ef10205383525a5 (diff)
downloadotp-02e09f135b2ef75ff440ee008e3db437b8fc3ff7.tar.gz
otp-02e09f135b2ef75ff440ee008e3db437b8fc3ff7.tar.bz2
otp-02e09f135b2ef75ff440ee008e3db437b8fc3ff7.zip
Mnesia sometimes failed to tell all nodes that it had started.
Diffstat (limited to 'lib/mnesia')
-rw-r--r--lib/mnesia/src/mnesia_controller.erl83
1 files changed, 50 insertions, 33 deletions
diff --git a/lib/mnesia/src/mnesia_controller.erl b/lib/mnesia/src/mnesia_controller.erl
index 021be8af2a..0254769758 100644
--- a/lib/mnesia/src/mnesia_controller.erl
+++ b/lib/mnesia/src/mnesia_controller.erl
@@ -1,7 +1,7 @@
%%
%% %CopyrightBegin%
%%
-%% Copyright Ericsson AB 1996-2010. All Rights Reserved.
+%% Copyright Ericsson AB 1996-2011. All Rights Reserved.
%%
%% The contents of this file are subject to the Erlang Public License,
%% Version 1.1, (the "License"); you may not use this file except in
@@ -457,7 +457,7 @@ connect_nodes2(Father, Ns, UserFun) ->
New1 = mnesia_lib:intersect(Ns, Connected),
New = New1 -- Current,
process_flag(trap_exit, true),
- Res = try_merge_schema(New, UserFun),
+ Res = try_merge_schema(New, [], UserFun),
Msg = {schema_is_merged, [], late_merge, []},
multicall([node()|Ns], Msg),
After = val({current, db_nodes}),
@@ -471,7 +471,7 @@ connect_nodes2(Father, Ns, UserFun) ->
merge_schema() ->
AllNodes = mnesia_lib:all_nodes(),
- case try_merge_schema(AllNodes, fun default_merge/1) of
+ case try_merge_schema(AllNodes, [node()], fun default_merge/1) of
ok ->
schema_is_merged();
{aborted, {throw, Str}} when is_list(Str) ->
@@ -483,11 +483,17 @@ merge_schema() ->
default_merge(F) ->
F([]).
-try_merge_schema(Nodes, UserFun) ->
+try_merge_schema(Nodes, Told0, UserFun) ->
case mnesia_schema:merge_schema(UserFun) of
{atomic, not_merged} ->
%% No more nodes that we need to merge the schema with
- ok;
+ %% Ensure we have told everybody that we are running
+ case val({current,db_nodes}) -- mnesia_lib:uniq(Told0) of
+ [] -> ok;
+ Tell ->
+ im_running(Tell, [node()]),
+ ok
+ end;
{atomic, {merged, OldFriends, NewFriends}} ->
%% Check if new nodes has been added to the schema
Diff = mnesia_lib:all_nodes() -- [node() | Nodes],
@@ -496,12 +502,18 @@ try_merge_schema(Nodes, UserFun) ->
%% Tell everybody to adopt orphan tables
im_running(OldFriends, NewFriends),
im_running(NewFriends, OldFriends),
-
- try_merge_schema(Nodes, UserFun);
+ Told = case lists:member(node(), NewFriends) of
+ true -> Told0 ++ OldFriends;
+ false -> Told0 ++ NewFriends
+ end,
+ try_merge_schema(Nodes, Told, UserFun);
{atomic, {"Cannot get cstructs", Node, Reason}} ->
dbg_out("Cannot get cstructs, Node ~p ~p~n", [Node, Reason]),
- timer:sleep(1000), % Avoid a endless loop look alike
- try_merge_schema(Nodes, UserFun);
+ timer:sleep(300), % Avoid a endless loop look alike
+ try_merge_schema(Nodes, Told0, UserFun);
+ {aborted, {shutdown, _}} -> %% One of the nodes is going down
+ timer:sleep(300), % Avoid a endless loop look alike
+ try_merge_schema(Nodes, Told0, UserFun);
Other ->
Other
end.
@@ -915,6 +927,7 @@ handle_cast(unblock_controller, State) ->
handle_cast({mnesia_down, Node}, State) ->
maybe_log_mnesia_down(Node),
mnesia_lib:del({current, db_nodes}, Node),
+ mnesia_lib:unset({node_up, Node}),
mnesia_checkpoint:tm_mnesia_down(Node),
Alltabs = val({schema, tables}),
reconfigure_tables(Node, Alltabs),
@@ -977,11 +990,12 @@ handle_cast(Msg, State) when State#state.schema_is_merged /= true ->
%% This must be done after schema_is_merged otherwise adopt_orphan
%% might trigger a table load from wrong nodes as a result of that we don't
%% know which tables we can load safly first.
-handle_cast({im_running, _Node, NewFriends}, State) ->
+handle_cast({im_running, Node, NewFriends}, State) ->
LocalTabs = mnesia_lib:local_active_tables() -- [schema],
RemoveLocalOnly = fun(Tab) -> not val({Tab, local_content}) end,
Tabs = lists:filter(RemoveLocalOnly, LocalTabs),
- Ns = mnesia_lib:intersect(NewFriends, val({current, db_nodes})),
+ Nodes = mnesia_lib:union([Node],val({current, db_nodes})),
+ Ns = mnesia_lib:intersect(NewFriends, Nodes),
abcast(Ns, {adopt_orphans, node(), Tabs}),
noreply(State);
@@ -1042,30 +1056,33 @@ handle_cast({master_nodes_updated, Tab, Masters}, State) ->
end;
handle_cast({adopt_orphans, Node, Tabs}, State) ->
-
State2 = node_has_tabs(Tabs, Node, State),
- %% Register the other node as up and running
- mnesia_recover:log_mnesia_up(Node),
- verbose("Logging mnesia_up ~w~n",[Node]),
- mnesia_lib:report_system_event({mnesia_up, Node}),
-
- %% Load orphan tables
- LocalTabs = val({schema, local_tables}) -- [schema],
- Nodes = val({current, db_nodes}),
- {LocalOrphans, RemoteMasters} =
- orphan_tables(LocalTabs, Node, Nodes, [], []),
- Reason = {adopt_orphan, node()},
- mnesia_late_loader:async_late_disc_load(node(), LocalOrphans, Reason),
-
- Fun =
- fun(N) ->
- RemoteOrphans =
- [Tab || {Tab, Ns} <- RemoteMasters,
- lists:member(N, Ns)],
- mnesia_late_loader:maybe_async_late_disc_load(N, RemoteOrphans, Reason)
- end,
- lists:foreach(Fun, Nodes),
+ case ?catch_val({node_up,Node}) of
+ true -> ignore;
+ _ ->
+ %% Register the other node as up and running
+ set({node_up, Node}, true),
+ mnesia_recover:log_mnesia_up(Node),
+ verbose("Logging mnesia_up ~w~n",[Node]),
+ mnesia_lib:report_system_event({mnesia_up, Node}),
+ %% Load orphan tables
+ LocalTabs = val({schema, local_tables}) -- [schema],
+ Nodes = val({current, db_nodes}),
+ {LocalOrphans, RemoteMasters} =
+ orphan_tables(LocalTabs, Node, Nodes, [], []),
+ Reason = {adopt_orphan, node()},
+ mnesia_late_loader:async_late_disc_load(node(), LocalOrphans, Reason),
+
+ Fun =
+ fun(N) ->
+ RemoteOrphans =
+ [Tab || {Tab, Ns} <- RemoteMasters,
+ lists:member(N, Ns)],
+ mnesia_late_loader:maybe_async_late_disc_load(N, RemoteOrphans, Reason)
+ end,
+ lists:foreach(Fun, Nodes)
+ end,
noreply(State2);
handle_cast(Msg, State) ->