leader_cron provides a distributed task scheduler for executing task periodically in an Erlang cluster
Participating members of the cluster elect a leader node. The leader node schedules and executes given tasks as per their schedules. Should the leader node become unavailable a new leader is elected who resumes task execution responsibilities.
Tasks are defined by specifying a function in a particular module with given arguments to run according to a schedule. The schedule types are:
- sleeper - sleep a specified number of milliseconds between task executions
- one shot - execute task once at a given date and time or after a number of milliseconds
- cron - define a schedule very similar to Unix cron
Startup leader_cron on each participating node (do this on all nodes):
leader_cron:start_link(['[email protected]', '[email protected]']).Schedule tasks from any node. Here a cron style schedule is defined.
leader_cron:schedule_task({cron, {[5], all, all, all, all}},
{io, format, [user, "It is 5 past the hour", []]}).That's it. In this example the task prints, "It is 5 past the hour" on the leader node at 5 minutes past every hour.
You can also schedule anonymous functions:
F = fun(Device, Format, Args) -> io:format(Device, Format, Args) end,
leader_cron:schedule_task({cron, {[5], all, all, all, all}},
{F, [user, "It is 5 past the hour", []]}).See the leader_cron_task module for full scheduling details (or make doc).
Run make or include as a rebar dependency
in your project.
To run all tests run make check. This runs both the eunit tests as well as the
common tests. These can also be run individually via make eunit and make ct.
To perform the Dialyzer
static analysis of the code run make dialyzer. Run make build_plt
once before performing the static analysis to build the plt file required
by dialyzer.
To run the Xref
cross reference tool run make xref.