global

A Global Name Registration Facility

This documentation describes the Global module which consists of the following functionalities:

registration of global names; global locks; maintenance of the fully connected network.

These services are controlled via the process global_name_server which exists on every node. The global name server is started automatically when a node is started. With the term global is meant over a system consisting of several Erlang nodes.

The ability to globally register names is a central concept in the programming of distributed Erlang systems. In this module, the equivalent of the register/2 and whereis/1 BIFs (for local name registration) are implemented, but for a network of Erlang nodes. A registered name is an alias for a process identifier (pid). The global name server monitors globally registered pids. If a process terminates, the name will also be globally unregistered.

The registered names are stored in replica global name tables on every node. There is no central storage point. Thus, the translation of a name to a pid is fast, as it is always done locally. When any action in taken which results in a change to the global name table, all tables on other nodes are automatically updated.

Global locks have lock identities and are set on a specific resource. For instance, the specified resource could be a pid. When a global lock is set, access to the locked resource is denied for all other resources other than the lock requester.

Both the registration and lock functionalities are atomic. All nodes involved in these actions will have the same view of the information.

The global name server also performs the critical task of continuously monitoring changes in node configuration: if a node which runs a globally registered process goes down, the name will be globally unregistered. To this end the global name server subscribes to nodeup and nodedown messages sent from the net_kernel module. Relevant Kernel application variables in this context are net_setuptime, net_ticktime, and dist_auto_connect. See also kernel(6).

The name server will also maintain a fully connected network. For example, if node N1 connects to node N2 (which is already connected to N3), the global name servers on the nodes N1 and N3 will make sure that also N1 and N3 are connected. If this is not desired, the command line flag -connect_all false can be used (see also erl(1)). In this case the name registration facility cannot be used, but the lock mechanism will still work.

If the global name server fails to connect nodes (N1 and N3 in the example above) a warning event is sent to the error logger. The presence of such an event does not exclude the possibility that the nodes will later connect--one can for example try the command rpc:call(N1, net_adm, ping, [N2]) in the Erlang shell--but it indicates some kind of problem with the network.

Note!

If the fully connected network is not set up properly, the first thing to try is to increase the value of net_setuptime.

Types


id() = {ResourceId :: term(), LockRequesterId :: term()}

Functions


del_lock(Id) -> true

del_lock(Id, Nodes) -> true

  • Id = id()
  • Nodes = [node()]

Deletes the lock Id synchronously.

notify_all_name(Name, Pid1, Pid2) -> none

  • Name = term()
  • Pid1 = Pid2 = pid()

This function can be used as a name resolving function for register_name/3 and re_register_name/3. It unregisters both pids, and sends the message {global_name_conflict, Name, OtherPid} to both processes.

random_exit_name(Name, Pid1, Pid2) -> none

  • Name = term()
  • Pid1 = Pid2 = pid()

This function can be used as a name resolving function for register_name/3 and re_register_name/3. It randomly chooses one of the pids for registration and kills the other one.

random_notify_name(Name, Pid1, Pid2) -> none

  • Name = term()
  • Pid1 = Pid2 = pid()

This function can be used as a name resolving function for register_name/3 and re_register_name/3. It randomly chooses one of the pids for registration, and sends the message {global_name_conflict, Name} to the other pid.

register_name(Name, Pid) -> yes | no

  • Name = term()
  • Pid = pid()

register_name(Name, Pid, Resolve) -> yes | no

  • Name = term()
  • Pid = pid()
  • Resolve = method()
  • method() =
        fun((Name :: term(), Pid :: pid(), Pid2 :: pid()) ->
                pid() | none)
{Module, Function} is also allowed

Globally associates the name Name with a pid, that is, Globally notifies all nodes of a new global name in a network of Erlang nodes.

When new nodes are added to the network, they are informed of the globally registered names that already exist. The network is also informed of any global names in newly connected nodes. If any name clashes are discovered, the Resolve function is called. Its purpose is to decide which pid is correct. If the function crashes, or returns anything other than one of the pids, the name is unregistered. This function is called once for each name clash.

There are three pre-defined resolve functions: random_exit_name/3, random_notify_name/3, and notify_all_name/3. If no Resolve function is defined, random_exit_name is used. This means that one of the two registered processes will be selected as correct while the other is killed.

This function is completely synchronous. This means that when this function returns, the name is either registered on all nodes or none.

The function returns yes if successful, no if it fails. For example, no is returned if an attempt is made to register an already registered process or to register a process with a name that is already in use.

Note!

Releases up to and including OTP R10 did not check if the process was already registered. As a consequence the global name table could become inconsistent. The old (buggy) behavior can be chosen by giving the Kernel application variable global_multi_name_action the value allow.

If a process with a registered name dies, or the node goes down, the name is unregistered on all nodes.

registered_names() -> [Name]

  • Name = term()

Returns a lists of all globally registered names.

re_register_name(Name, Pid) -> term()

  • Name = term()
  • Pid = pid()

re_register_name(Name, Pid, Resolve) -> term()

  • Name = term()
  • Pid = pid()
  • Resolve = method()
  • method() =
        fun((Name :: term(), Pid :: pid(), Pid2 :: pid()) ->
                pid() | none)
{Module, Function} is also allowed

Atomically changes the registered name Name on all nodes to refer to Pid.

The Resolve function has the same behavior as in register_name/2,3.

send(Name, Msg) -> Pid

  • Name = Msg = term()
  • Pid = pid()

Sends the message Msg to the pid globally registered as Name.

Failure: If Name is not a globally registered name, the calling function will exit with reason {badarg, {Name, Msg}}.

set_lock(Id) -> boolean()

set_lock(Id, Nodes) -> boolean()

  • Id = id()
  • Nodes = [node()]

set_lock(Id, Nodes, Retries) -> boolean()

  • id() = {ResourceId :: term(), LockRequesterId :: term()}
  • retries() = integer() >= 0 | infinity

Sets a lock on the specified nodes (or on all nodes if none are specified) on ResourceId for LockRequesterId. If a lock already exists on ResourceId for another requester than LockRequesterId, and Retries is not equal to 0, the process sleeps for a while and will try to execute the action later. When Retries attempts have been made, false is returned, otherwise true. If Retries is infinity, true is eventually returned (unless the lock is never released).

If no value for Retries is given, infinity is used.

This function is completely synchronous.

If a process which holds a lock dies, or the node goes down, the locks held by the process are deleted.

The global name server keeps track of all processes sharing the same lock, that is, if two processes set the same lock, both processes must delete the lock.

This function does not address the problem of a deadlock. A deadlock can never occur as long as processes only lock one resource at a time. But if some processes try to lock two or more resources, a deadlock may occur. It is up to the application to detect and rectify a deadlock.

Note!

Some values of ResourceId should be avoided or Erlang/OTP will not work properly. A list of resources to avoid: global, dist_ac, mnesia_table_lock, mnesia_adjust_log_writes, pg2.

sync() -> ok | {error, Reason :: term()}

Synchronizes the global name server with all nodes known to this node. These are the nodes which are returned from erlang:nodes(). When this function returns, the global name server will receive global information from all nodes. This function can be called when new nodes are added to the network.

The only possible error reason Reason is {"global_groups definition error", Error}.

trans(Id, Fun) -> Res | aborted

trans(Id, Fun, Nodes) -> Res | aborted

trans(Id, Fun, Nodes, Retries) -> Res | aborted

  • retries() = integer() >= 0 | infinity
  • trans_fun() = function() | {module(), atom()}

Sets a lock on Id (using set_lock/3). If this succeeds, Fun() is evaluated and the result Res is returned. Returns aborted if the lock attempt failed. If Retries is set to infinity, the transaction will not abort.

infinity is the default setting and will be used if no value is given for Retries.

unregister_name(Name) -> term()

  • Name = term()

Removes the globally registered name Name from the network of Erlang nodes.

whereis_name(Name) -> pid() | undefined

  • Name = term()

Returns the pid with the globally registered name Name. Returns undefined if the name is not globally registered.