frappy/frappy/server.py
Markus Zolliker eee63ee3df config: do not override equipment_id with name
In the previous code, the equipment_id was overridden by the
server name when the interface argument was given over
the commandline. This was leftover from the previous config
file format, where the config files not neccessarly needed
an equipment_id.

Change-Id: I2fc248372a7d2f61cc0690804268d6d066a0a9fa
Reviewed-on: https://forge.frm2.tum.de/review/c/secop/frappy/+/35391
Tested-by: Jenkins Automated Tests <pedersen+jenkins@frm2.tum.de>
Reviewed-by: Markus Zolliker <markus.zolliker@psi.ch>
2025-01-17 15:00:20 +01:00

356 lines
14 KiB
Python

# *****************************************************************************
#
# This program is free software; you can redistribute it and/or modify it under
# the terms of the GNU General Public License as published by the Free Software
# Foundation; either version 2 of the License, or (at your option) any later
# version.
#
# This program is distributed in the hope that it will be useful, but WITHOUT
# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
# FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
# details.
#
# You should have received a copy of the GNU General Public License along with
# this program; if not, write to the Free Software Foundation, Inc.,
# 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
#
# Module authors:
# Enrico Faulhaber <enrico.faulhaber@frm2.tum.de>
# Alexander Lenz <alexander.lenz@frm2.tum.de>
# Markus Zolliker <markus.zolliker@psi.ch>
#
# *****************************************************************************
"""Define helpers"""
import os
import signal
import sys
import threading
import time
import mlzlog
from frappy.config import load_config
from frappy.errors import ConfigError
from frappy.lib import formatException, generalConfig, get_class, mkthread
from frappy.lib.multievent import MultiEvent
from frappy.logging import init_remote_logging
from frappy.params import PREDEFINED_ACCESSIBLES
from frappy.secnode import SecNode
from frappy.protocol.discovery import UDPListener
try:
from daemon import DaemonContext
try:
from daemon import pidlockfile
except ImportError:
import daemon.pidfile as pidlockfile
except ImportError:
DaemonContext = None
try:
# pylint: disable=unused-import
import systemd.daemon
except ImportError:
systemd = None
class Server:
INTERFACES = {
'tcp': 'protocol.interface.tcp.TCPServer',
'ws': 'protocol.interface.ws.WSServer',
}
_restart = True
def __init__(self, name, parent_logger, *, cfgfiles=None, interface=None, testonly=False):
"""initialize server
Arguments:
- name: the node name
- parent_logger: the logger to inherit from. a handler is installed by
the server to provide remote logging
- cfgfiles: if not given, defaults to [name]
may be a list of cfg files
items ending with .py are taken as paths, else _cfg.py is appended and
files are looked up in the config path retrieved from the general config
- interface: an uri of the from tcp://<port> or a bare port number for tcp
if not given, the interface is taken from the config file. In case of
multiple cfg files, the interface is taken from the first cfg file
- testonly: test mode. tries to build all modules, but the server is not started
Config file:
Format: Example:
Node('<equipment_id>', Node('ex.frappy.demo',
<description>, 'short description\n\nlong descr.',
<main interface>, 'tcp://10769',
secondary=[ secondary=['ws://10770'], # optional
<interfaces>
],
) )
Mod('<module name>', Mod('temp',
<param config> value = Param(unit='K'),
) )
...
"""
self._testonly = testonly
if not cfgfiles:
cfgfiles = [name]
# sanitize name (in case it is a cfgfile)
self.name = name = os.path.splitext(os.path.basename(name))[0]
if isinstance(parent_logger, mlzlog.MLZLogger):
self.log = parent_logger.getChild(name, True)
else:
self.log = parent_logger.getChild(name)
init_remote_logging(self.log)
merged_cfg = load_config(cfgfiles, self.log)
self.node_cfg = merged_cfg.pop('node')
self.module_cfg = merged_cfg
if interface:
self.node_cfg['interface'] = str(interface)
elif not self.node_cfg.get('interface'):
raise ConfigError('No interface specified in configuration or arguments!')
self._cfgfiles = cfgfiles
self._pidfile = generalConfig.piddir / (name + '.pid')
signal.signal(signal.SIGINT, self.signal_handler)
signal.signal(signal.SIGTERM, self.signal_handler)
self.discovery = None
def signal_handler(self, num, frame):
if hasattr(self, 'interfaces') and self.interfaces:
self.shutdown()
else:
# TODO: we should probably clean up the already initialized modules
# when getting an interrupt while the server is starting
signal.default_int_handler(num, frame)
def start(self):
if not DaemonContext:
raise ConfigError('can not daemonize, as python-daemon is not installed')
piddir = self._pidfile.parent
if not piddir.is_dir():
piddir.mkdir(parents=True)
pidfile = pidlockfile.TimeoutPIDLockFile(self._pidfile)
if pidfile.is_locked():
self.log.error('Pidfile already exists. Exiting')
with DaemonContext(
pidfile=pidfile,
files_preserve=self.log.getLogfileStreams()):
self.run()
def unknown_options(self, cls, options):
return f"{cls.__name__} class don't know how to handle option(s): {', '.join(options)}"
def restart_hook(self):
"""Actions to be done on restart. May be overridden by a subclass."""
def run(self):
global systemd # pylint: disable=global-statement
while self._restart:
self._restart = False
try:
# TODO: make systemd notifications configurable
if systemd:
systemd.daemon.notify("STATUS=initializing")
except Exception:
systemd = None
try:
self._processCfg()
if self._testonly:
return
except Exception:
print(formatException(verbose=True))
raise
# client interfaces
self.interfaces = {}
iface_threads = []
# default_timeout 12 sec: TCPServer might need up to 10 sec to wait for Address no longer in use
interfaces_started = MultiEvent(default_timeout=12)
lock = threading.Lock()
failed = {}
interfaces = [self.node_cfg['interface']] + self.node_cfg.get('secondary', [])
# allow missing "tcp://"
interfaces = [iface if '://' in iface else f'tcp://{iface}' for iface in interfaces]
with lock:
for interface in interfaces:
opts = {'uri': interface}
t = mkthread(
self._interfaceThread,
opts,
lock,
failed,
interfaces,
interfaces_started.get_trigger(),
)
iface_threads.append(t)
if not interfaces_started.wait():
for iface in interfaces:
if iface not in failed and iface not in self.interfaces:
self.log.error('timeout starting interface %s', iface)
while failed:
iface, err = failed.popitem()
self.log.error('starting interface %s failed with %r', iface, err)
if not self.interfaces:
self.log.error('no interface started')
return
self.secnode.add_secnode_property('_interfaces', list(self.interfaces))
self.log.info('startup done with interface(s) %s',
', '.join(self.interfaces))
# start discovery interface when we know where we listen
self.discovery = UDPListener(
self.secnode.equipment_id,
self.secnode.get_secnode_property('description'),
list(self.interfaces),
self.log.getChild('discovery')
)
mkthread(self.discovery.run)
if systemd:
systemd.daemon.notify("READY=1\nSTATUS=accepting requests")
if os.name == 'nt':
# workaround: thread.join() on Windows blocks and is not
# interruptible by the signal handler, so loop and check
# periodically whether the interfaces are still running.
while True:
time.sleep(1)
if not interfaces:
break
else:
for t in iface_threads:
t.join()
while failed:
iface, err = failed.popitem()
self.log.error('interface %s failed with %r', iface, err)
self.log.info('stopped listening, cleaning up %d modules',
len(self.secnode.modules))
if systemd:
if self._restart:
systemd.daemon.notify('RELOADING=1')
else:
systemd.daemon.notify('STOPPING=1')
self.secnode.shutdown_modules()
if self._restart:
self.restart_hook()
self.log.info('restarting')
self.log.info('shut down')
def restart(self):
if not self._restart:
self._restart = True
for iface in self.interfaces.values():
iface.shutdown()
def shutdown(self):
self._restart = False
if self.discovery:
self.discovery.shutdown()
for iface in self.interfaces.values():
iface.shutdown()
def _interfaceThread(self, opts, lock, failed, interfaces, start_cb):
iface = opts['uri']
scheme = iface.split('://')[0]
cls = get_class(self.INTERFACES[scheme])
try:
with cls(scheme, self.log.getChild(scheme), opts, self) as interface:
if opts:
raise ConfigError(self.unknown_options(cls, opts))
with lock:
self.interfaces[iface] = interface
start_cb()
interface.serve_forever()
# server_close() called by 'with'
except Exception as e:
with lock:
failed[iface] = e
interfaces.remove(iface)
start_cb() # callback should also be called on failure
else:
with lock:
interfaces.remove(iface)
self.log.info(f'stopped {iface}')
def _processCfg(self):
"""Processes the module configuration.
All modules specified in the config file and read recursively from
Pinata class Modules are instantiated, initialized and started by the
end of this function.
If there are errors that occur, they will be collected and emitted
together in the end.
"""
errors = []
opts = dict(self.node_cfg)
cls = get_class(opts.pop('cls'))
self.secnode = SecNode(self.name, self.log.getChild('secnode'), opts, self)
self.dispatcher = cls(self.name, self.log.getChild('dispatcher'), opts, self)
# add other options as SECNode properties, those with '_' prefixed will
# get exported
for k in list(opts):
self.secnode.add_secnode_property(k, opts.pop(k))
self.secnode.create_modules()
# initialize all modules by getting them with Dispatcher.get_module,
# which is done in the get_descriptive data
# TODO: caching, to not make this extra work
self.secnode.get_descriptive_data('')
# =========== All modules are initialized ===========
# all errors from initialization process
errors = self.secnode.errors
if not self._testonly:
start_events = MultiEvent(default_timeout=30)
for modname, modobj in self.secnode.modules.items():
# startModule must return either a timeout value or None (default 30 sec)
start_events.name = f'module {modname}'
modobj.startModule(start_events)
if not modobj.startModuleDone:
errors.append(f'{modobj.startModule.__qualname__} was not called, probably missing super call')
if errors:
for errtxt in errors:
for line in errtxt.split('\n'):
self.log.error(line)
# print a list of config errors to stderr
sys.stderr.write('\n'.join(errors))
sys.stderr.write('\n')
sys.exit(1)
if self._testonly:
return
self.log.info('waiting for modules being started')
start_events.name = None
if not start_events.wait():
# some timeout happened
for name in start_events.waiting_for():
self.log.warning('timeout when starting %s', name)
self.log.info('all modules started')
history_path = os.environ.get('FRAPPY_HISTORY')
if history_path:
from frappy_psi.historywriter import \
FrappyHistoryWriter # pylint: disable=import-outside-toplevel
writer = FrappyHistoryWriter(history_path, PREDEFINED_ACCESSIBLES.keys(), self.dispatcher)
# treat writer as a connection
self.dispatcher.add_connection(writer)
writer.init(self.dispatcher.handle_describe(writer, None, None))
# TODO: if ever somebody wants to implement an other history writer:
# - a general config file /etc/secp/frappy.conf or <frappy repo>/etc/frappy.conf
# might be introduced, which contains the log, pid and cfg directory path and
# the class path implementing the history
# - or we just add here an other if statement:
# history_path = os.environ.get('ALTERNATIVE_HISTORY')
# if history_path:
# from frappy_<xx>.historywriter import ... etc.