2 """A very simple logger that tries to be concurrency-safe."""
10 LOG_FILE = '/var/log/nodemanager'
11 LOG_SLIVERS = '/var/lib/nodemanager/getslivers.txt'
12 LOG_DATABASE = '/var/lib/nodemanager/database.txt'
14 # basically define 3 levels
18 # default is to log a reasonable amount of stuff for when running on operational nodes
23 assert level in [LOG_NONE,LOG_NODE,LOG_VERBOSE]
27 log('(v) '+msg,LOG_VERBOSE)
29 def log(msg,level=LOG_NODE):
30 """Write <msg> to the log file if level >= current log level (default LOG_NODE)."""
31 if (level > LOG_LEVEL):
34 fd = os.open(LOG_FILE, os.O_WRONLY | os.O_CREAT | os.O_APPEND, 0600)
35 if not msg.endswith('\n'): msg += '\n'
36 os.write(fd, '%s: %s' % (time.asctime(time.gmtime()), msg))
42 def log_exc(msg="",name=None):
43 """Log the traceback resulting from an exception."""
45 log("%s: EXCEPTION caught <%s> \n %s" %(name, msg, traceback.format_exc()))
47 log("EXCEPTION caught <%s> \n %s" %(msg, traceback.format_exc()))
49 ########## snapshot data to a file
50 # for some reason the various modules are still triggered even when the
51 # data from PLC cannot be reached
52 # we show this message instead of the exception stack instead in this case
53 def log_missing_data (msg,key):
54 log("%s: could not find the %s key in data (PLC connection down?) - IGNORED"%(msg,key))
56 def log_data_in_file (data, file, message="",level=LOG_NODE):
57 if (level > LOG_LEVEL):
62 now=time.strftime("Last update: %Y.%m.%d at %H:%M:%S %Z", time.localtime())
64 if message: f.write('Message:'+message+'\n')
65 pp=pprint.PrettyPrinter(stream=f,indent=2)
68 verbose("logger:.log_data_in_file Owerwrote %s"%file)
70 log_exc('logger.log_data_in_file failed - file=%s - message=%r'%(file,message))
72 def log_slivers (data):
73 log_data_in_file (data, LOG_SLIVERS, "raw GetSlivers")
74 def log_database (db):
75 log_data_in_file (db, LOG_DATABASE, "raw database")
77 #################### child processes
78 # avoid waiting until the process returns;
79 # that makes debugging of hanging children hard
82 def __init__ (self,message='log_call: '):
88 if c=='\n': self.flush()
92 log (self.message + self.buffer)
95 # time out in seconds - avoid hanging subprocesses - default is 5 minutes
96 default_timeout_minutes=5
98 # returns a bool that is True when everything goes fine and the retcod is 0
99 def log_call(command,timeout=default_timeout_minutes*60,poll=1):
100 message=" ".join(command)
101 log("log_call: running command %s" % message)
102 verbose("log_call: timeout=%r s" % timeout)
103 verbose("log_call: poll=%r s" % poll)
104 trigger=time.time()+timeout
107 child = subprocess.Popen(command, bufsize=1,
108 stdout=subprocess.PIPE, stderr=subprocess.STDOUT, close_fds=True)
111 # see if anything can be read within the poll interval
112 (r,w,x)=select.select([child.stdout],[],[],poll)
113 if r: buffer.add(child.stdout.read(1))
115 returncode=child.poll()
117 if returncode != None:
119 # child is done and return 0
121 log("log_call:end command (%s) completed" % message)
126 log("log_call:end command (%s) returned with code %d" %(message,returncode))
128 # no : still within timeout ?
129 if time.time() >= trigger:
132 log("log_call:end terminating command (%s) - exceeded timeout %d s"%(message,timeout))
134 except: log_exc("failed to run command %s" % message)