ceph-mgr plugin author guide ============================ Creating a plugin ----------------- In pybind/mgr/, create a python module. Within your module, create a class named ``Module`` that inherits from ``MgrModule``. The most important methods to override are: * a ``serve`` member function for server-type modules. This function should block forever. * a ``notify`` member function if your module needs to take action when new cluster data is available. * a ``handle_command`` member function if your module exposes CLI commands. Installing a plugin ------------------- Once your module is present in the location set by the ``mgr module path`` configuration setting, you can enable it via the ``ceph mgr module enable`` command:: ceph mgr module enable mymodule Note that the MgrModule interface is not stable, so any modules maintained outside of the Ceph tree are liable to break when run against any newer or older versions of Ceph. Logging ------- MgrModule instances have a ``log`` property which is a logger instance that sends log messages into the Ceph logging layer where they will be recorded in the mgr daemon's log file. Use it the same way you would any other python logger. The python log levels debug, info, warn, err are mapped into the Ceph severities 20, 4, 1 and 0 respectively. Exposing commands ----------------- Set the ``COMMANDS`` class attribute of your plugin to a list of dicts like this:: COMMANDS = [ { "cmd": "foobar name=myarg,type=CephString", "desc": "Do something awesome", "perm": "rw" } ] The ``cmd`` part of each entry is parsed in the same way as internal Ceph mon and admin socket commands (see mon/MonCommands.h in the Ceph source for examples) Config settings --------------- Modules have access to a simple key/value store (keys and values are byte strings) for storing configuration. Don't use this for storing large amounts of data. Config values are stored using the mon's config-key commands. Hints for using these: * Reads are fast: ceph-mgr keeps a local in-memory copy * Don't set things by hand with "ceph config-key", the mgr doesn't update at runtime (only set things from within modules). * Writes block until the value is persisted, but reads from another thread will see the new value immediately. Any config settings you want to expose to users from your module will need corresponding hooks in ``COMMANDS`` to expose a setter. Accessing cluster data ---------------------- Modules have access to the in-memory copies of the Ceph cluster's state that the mgr maintains. Accessor functions as exposed as members of MgrModule. Calls that access the cluster or daemon state are generally going from Python into native C++ routines. There is some overhead to this, but much less than for example calling into a REST API or calling into an SQL database. There are no consistency rules about access to cluster structures or daemon metadata. For example, an OSD might exist in OSDMap but have no metadata, or vice versa. On a healthy cluster these will be very rare transient states, but plugins should be written to cope with the possibility. Note that these accessors must not be called in the modules ``__init__`` function. This will result in a circular locking exception. ``get(self, data_name)`` Fetch named cluster-wide objects such as the OSDMap. Valid things to fetch are osd_crush_map_text, osd_map, osd_map_tree, osd_map_crush, config, mon_map, fs_map, osd_metadata, pg_summary, df, osd_stats, health, mon_status. All these structures have their own JSON representations: experiment or look at the C++ dump() methods to learn about them. ``get_server(self, hostname)`` Fetch metadata about a particular hostname. This is information that ceph-mgr has gleaned from the daemon metadata reported by daemons running on a particular server. ``list_servers(self)`` Like ``get_server``, but gives information about all servers (i.e. all unique hostnames that have been mentioned in daemon metadata) ``get_metadata(self, svc_type, svc_id)`` Fetch the daemon metadata for a particular service. svc_type is one of osd or mds, and svc_id is a string (convert OSD integer IDs to strings when calling this). ``get_counter(self, svc_type, svc_name, path)`` Fetch the latest performance counter data for a particular counter. The path is a period-separated concatenation of the subsystem and the counter name, for example "mds.inodes". A list of two-tuples of (timestamp, value) is returned. This may be empty if no data is available. Sending commands ---------------- A non-blocking facility is provided for sending monitor commands to the cluster. ``send_command(self, result, command_str, tag)`` The ``result`` parameter should be an instance of the CommandResult class, defined in the same module as MgrModule. This acts as a completion and stores the output of the command. Use CommandResult.wait() if you want to block on completion. The ``command_str`` parameter is a JSON-serialized command. This uses the same format as the ceph command line, which is a dictionary of command arguments, with the extra ``prefix`` key containing the command name itself. Consult MonCommands.h for available commands and their expected arguments. The ``tag`` parameter is used for nonblocking operation: when a command completes, the ``notify()`` callback on the MgrModule instance is triggered, with notify_type set to "command", and notify_id set to the tag of the command. Logging ------- Use your module's ``log`` attribute as your logger. This is a logger configured to output via the ceph logging framework, to the local ceph-mgr log files. Python log severities are mapped to ceph severities as follows: * DEBUG is 20 * INFO is 4 * WARN is 1 * ERR is 0 Shutting down cleanly --------------------- If a module implements the ``serve()`` method, it should also implement the ``shutdown()`` method to shutdown cleanly: misbehaving modules may otherwise prevent clean shutdown of ceph-mgr. Is something missing? --------------------- The ceph-mgr python interface is not set in stone. If you have a need that is not satisfied by the current interface, please bring it up on the ceph-devel mailing list. While it is desired to avoid bloating the interface, it is not generally very hard to expose existing data to the Python code when there is a good reason.