EMQ学习 ---集群

时间:2024-10-02 08:08:01

emqttd集群设置管理

一、先来看EMQ的文档定义:http://emqtt.com/docs/v1/cluster.html

emqttd集群设置管理

假设部署两台服务器s1.emqtt.io, s2.emqtt.io上部署集群:

节点名 主机名(FQDN) IP地址

emqttd@s1.emqtt.io 或emqttd@192.168.0.10	s1.emqtt.io	192.168.0.10
emqttd@s2.emqtt.io 或emqttd@192.168.0.20 s2.emqtt.io 192.168.0.20

Warning

节点名格式: Name@Host, Host必须是IP地址或FQDN(主机名.域名)

emqttd@s1.emqtt.io节点设置

emqttd/etc/vm.args:

-name emqttd@s1.emqtt.io

-name emqttd@192.168.0.10

节点加入集群

启动两台节点后,emqttd@s2.emqtt.io上执行:

$ ./bin/emqttd_ctl cluster join emqttd@s1.emqtt.io

Join the cluster successfully.
Cluster status: [{running_nodes,['emqttd@s1.emqtt.io','emqttd@s2.emqtt.io']}]

或,emqttd@s1.emqtt.io上执行:

$ ./bin/emqttd_ctl cluster join emqttd@s2.emqtt.io

Join the cluster successfully.
Cluster status: [{running_nodes,['emqttd@s1.emqtt.io','emqttd@s2.emqtt.io']}]

  

任意节点上查询集群状态:

$ ./bin/emqttd_ctl cluster status

Cluster status: [{running_nodes,['emqttd@s1.emqtt.io','emqttd@s2.emqtt.io']}]

节点退出集群

节点退出集群,两种方式:

  • leave: 本节点退出集群
  • remove: 从集群删除其他节点

emqttd@s2.emqtt.io主动退出集群:

$ ./bin/emqttd_ctl cluster leave

或emqttd@s1.emqtt.io节点上,从集群删除emqttd@s2.emqtt.io节点:

$ ./bin/emqttd_ctl cluster remove emqttd@s2.emqtt.io

emqttd_ctl是怎么使用的?

-module(emqttd_cli).有定义要加载的命令

-export([status/1, broker/1, cluster/1, users/1, clients/1, sessions/1,
routes/1, topics/1, subscriptions/1, plugins/1, bridges/1,
listeners/1, vm/1, mnesia/1, trace/1]).
load() ->
Cmds = [Fun || {Fun, _} <- ?MODULE:module_info(exports), is_cmd(Fun)],
[emqttd_ctl:register_cmd(Cmd, {?MODULE, Cmd}, []) || Cmd <- Cmds].

加入集群后,子节点mnesia数据库怎么办?

mnesia数据库天然支持分布式集群。子节点加入之后就类似MySQL数据库主从备份一样,主节点和子节点mnesia会保持同步。来看源码:

-module(emqttd_mnesia).

%% @doc Join the mnesia cluster
-spec(join_cluster(node()) -> ok).
join_cluster(Node) when Node =/= node() ->
%% Stop mnesia and delete schema first
ensure_ok(ensure_stopped()),
ensure_ok(delete_schema()),
%% Start mnesia and cluster to node
ensure_ok(ensure_started()),
ensure_ok(connect(Node)),
ensure_ok(copy_schema(node())),
%% Copy tables
copy_tables(),
ensure_ok(wait_for(tables)).

子节点加入之后,会先删除自己的mnesia数据库和各个表,然后copy一份主节点的库,再copy各个表数据。

%% @doc Cluster with node.
-spec(connect(node()) -> ok | {error, any()}).
connect(Node) ->
case mnesia:change_config(extra_db_nodes, [Node]) of
{ok, [Node]} -> ok;
{ok, []} -> {error, {failed_to_connect_node, Node}};
Error -> Error
end. %% @doc Copy schema.
copy_schema(Node) ->
case mnesia:change_table_copy_type(schema, Node, disc_copies) of
{atomic, ok} ->
ok;
{aborted, {already_exists, schema, Node, disc_copies}} ->
ok;
{aborted, Error} ->
{error, Error}
end. %% @doc Copy mnesia tables.
copy_tables() ->
emqttd_boot:apply_module_attributes(copy_mnesia).

函数copy_tables(),会检索和执行emq工程目录下所有erl模块里面的mnesia(copy)函数,模块要求含有"-copy_mnesia({mnesia, [copy]})."关键字

例如:

-module(emqttd_backend).

mnesia(copy) ->
ok = emqttd_mnesia:copy_table(retained_message),
ok = emqttd_mnesia:copy_table(backend_subscription).

-module(emqttd_router).

-copy_mnesia({mnesia, [copy]}).

mnesia(copy) ->
ok = emqttd_mnesia:copy_table(route, ram_copies).

EMQ工程目录下,有关键字-boot_mnesia({mnesia, [boot]}).和-copy_mnesia({mnesia, [copy]}).的模块是:

-module(emqttd_backend).
-module(emqttd_pubsub).
-module(emqttd_router).
-module(emqttd_server).
-module(emqttd_sm).
-module(emqttd_trie).

其中,emqttd_backend模块新建的数据库retained_message和backend_subscription是disc_copies类型的,其他模块是ram_copies类型的。

emqttd的mnesia初始化

1、-module(emqttd_app).

start(_StartType, _StartArgs) ->
print_banner(),
emqttd_mnesia:start(),

  

2、-module(emqttd_mnesia)

start() ->
ensure_ok(ensure_data_dir()),
ensure_ok(init_schema()),
ok = mnesia:start(),
init_tables(),
wait_for(tables). %% @doc Init mnesia schema or tables.
init_schema() ->
case mnesia:system_info(extra_db_nodes) of
[] -> mnesia:create_schema([node()]);
[_|_] -> ok
end. %% @private
%% @doc Init mnesia tables.
init_tables() ->
case mnesia:system_info(extra_db_nodes) of
[] -> create_tables();
[_|_] -> copy_tables()
end.

3、数据库启动和拷贝的例子-module(emqttd_backend).

-boot_mnesia({mnesia, [boot]}).
-copy_mnesia({mnesia, [copy]}). %% Mnesia callbacks
%%-------------------------------------------------------------------- mnesia(boot) ->
ok = emqttd_mnesia:create_table(retained_message, [
{type, ordered_set},
{disc_copies, [node()]},
{record_name, retained_message},
{attributes, record_info(fields, retained_message)},
{storage_properties, [{ets, [compressed]},
{dets, [{auto_save, 1000}]}]}]),
ok = emqttd_mnesia:create_table(backend_subscription, [
{type, bag},
{disc_copies, [node()]},
{record_name, mqtt_subscription},
{attributes, record_info(fields, mqtt_subscription)},
{storage_properties, [{ets, [compressed]},
{dets, [{auto_save, 5000}]}]}]); mnesia(copy) ->
ok = emqttd_mnesia:copy_table(retained_message),
ok = emqttd_mnesia:copy_table(backend_subscription).
4、数据库启动和拷贝的例子-module(emqttd_router). -boot_mnesia({mnesia, [boot]}).
-copy_mnesia({mnesia, [copy]}). mnesia(boot) ->
ok = emqttd_mnesia:create_table(route, [
{type, bag},
{ram_copies, [node()]},
{record_name, mqtt_route},
{attributes, record_info(fields, mqtt_route)}]); mnesia(copy) ->
ok = emqttd_mnesia:copy_table(route, ram_copies).

4、数据库启动和拷贝的例子-module(emqttd_router).

-boot_mnesia({mnesia, [boot]}).
-copy_mnesia({mnesia, [copy]}). mnesia(boot) ->
ok = emqttd_mnesia:create_table(route, [
{type, bag},
{ram_copies, [node()]},
{record_name, mqtt_route},
{attributes, record_info(fields, mqtt_route)}]); mnesia(copy) ->
ok = emqttd_mnesia:copy_table(route, ram_copies).

注意事项

1、如果EMQ所在服务器的IP地址是192.168.0.10

那么节点名称A:emqttd@192.168.0.10和节点名称B:emqttd@127.0.0.1是相同的意思,如果EMQ以节点A启动服务器,那么再以节点B启动是会失败的。

此时只能把A或B其中一个更名一下。即节点名格式: Name@Host里面的Name要加以区分。

2、集群的信息会记录在工程目录下,/rel/emqttd/data/mnesia/emqttd@172.16.6.161/schema.DAT

即,当子节点A连接了主节点B,集群信息会分别记录在schema.DAT。如果子节点A没有主动断开集群,下次重启时,仍然会主动连接主节点B。

★有几个遗留问题待确认,不知道EMQ V2版本有无修正:

问题(1)如果子节点A没有主动断开集群,下次重启时,如果B不存在,那么A就会启动失败!好可怕!

问题(2)A连接上B之后。A目录下的文件/rel/emqttd/data/mnesia/emqttd@172.16.6.161/retained_message.DCD和backend_subscription.DCD就自我删除了。以后也见不着了,彻底消失了,奇怪!请注意,这两个Mnesia数据库表类型是持久化,disc_copies。

★2018/05/17实测emq2.3.7,A主B从,结论如下:

(1)B join A之后,会主动删除B的Mnesia表,然后从A拷贝一份过来。B leave A之后,也会删除B的Mnesia表。

(2)B join A之后,A和B的Mnesia表会始终保持一致性。

添加或删除或更新A的表数据,B会同步。

添加或删除或更新B的表数据,A会同步。

(3)集群的信息会记录在工程目录下,/rel/emqttd/data/mnesia/emqttd@172.16.6.161/schema.DAT。A或B进程退出后,再次启动时,仍然保持集群状态。

3、常用命令

./emqttd console
./emqttd start
./emqttd stop
./emqttd_ctl cluster join emqttd@172.16.6.161
./emqttd_ctl cluster status

./emqttd_ctl cluster leave

werl -name firecat@127.0.0.1 -setcookie emqsecretcookie
observer:start().
./_rel/emqttd/bin/emqttd console
./_rel/emqttd/bin/emqttd start
./_rel/emqttd/bin/emqttd_ctl status
./_rel/emqttd/bin/emqttd stop
./_rel/emqttd/bin/emqttd_ctl cluster join emq@192.168.0.116
./_rel/emqttd/bin/emqttd_ctl cluster status
./_rel/emqttd/bin/emqttd_ctl cluster leave