#!/usr/bin/env python
#
# $Id:$
#
# hd_raidmonitor.py
#
# This script will check the remaining disk space on the
# RAID disk on the gluon cluster and set the EPICS variable
# HD:coda:daq:availableRAID to the number of TB remaining. The
# EPICS alarm system is then used to raise an alarm if the
# value drops below a certain level. This script will loop
# infinitely in order to keep this value updated by checking
# every 2 seconds. The heartbeat variable HD:coda:daq:heartbeat will
# also be updated so the EPICS system can know that this program
# is still reporting.
#
# This file is kept under version control at the following location:
#
# https://halldsvn.jlab.org/repos/trunk/online/daq/tools/raidutils/hd_raidmonitor.py
#
#
#  It should be run on gluonraid1 regardless of where the ER is
# run. This allows us to know where to look for it and to ensure
# only one of these is running. It will observe other disks via
# NFS.
#
# This can be started using the procServ program so that it will
# be automatically restarted if the process should die. It is
# currently not set up to start automatically on reboot so one
# should start it "by hand" from the hdops account on gluonraid1
# like this:
#
#   procServ -n "DAQ RAID Monitor" 26001 /gluex/builds/devel/Linux_RHEL6-x86_64-gcc4.9.2/bin/hd_raidmonitor.py
#

import os
import sys
import time
import getopt
from epics import caget, caput

MONITOR_DIR = '/gluex/data/rawdata/curr' # directory to use for checking the remaining disk space
VERBOSE = False
PERIOD  = 2 # seconds between checks (and heartbeats)
EPICS_VAR = 'HD:coda:daq:availableRAID'
EPICS_HEARTBEAT = 'HD:coda:daq:heartbeat'

def Usage(err):
	print 'Usage:'
	print '   hd_raidmonitor.py [-v] [-p period_secs] [-d directory]'
	print ' '
	sys.exit(err)


# Parse command line arguments
try:
	opts, args = getopt.getopt(sys.argv[1:], 'p:d:vh', ['pause=','dir=','verbose','help'])
except getopt.GetoptError:
	Usage(2)

for opt,arg in opts:
	if opt in ("-h", "--help")    : Usage(0)
	if opt in ("-p", "--pause")   : PERIOD = int(arg)
	if opt in ("-d", "--dir")     : MONITOR_DIR = arg
	if opt in ("-v", "--verbose") : VERBOSE = True



# Loop forever
first_iteration = True
heartbeat = 0
while True:

	# Make sure MONITOR_DIR exists
	if not os.path.exists(MONITOR_DIR) :
		print '-------------------------------------'
		print 'path "%s" does not exist!' % MONITOR_DIR
		print 'hd_raidmonitor.py exiting ..'
		print '-------------------------------------'
		break

	# Get remaining disk space
	s = os.statvfs(MONITOR_DIR)
	Nbytes = float(s.f_bavail * s.f_frsize)
	NTB = Nbytes/1024/1024/1024/1024

	# Print info only on first iteration
	if first_iteration :
		print '-------------------------------------'
		print 'hd_raidmonitor.py starting'
		print ''
		print '        VERBOSE: %s' % VERBOSE
		print '         PERIOD: %d' % PERIOD
		print '    MONITOR_DIR: %s' % MONITOR_DIR
		print ' EPICS PV names: %s' % EPICS_VAR
		print '                 %s' % EPICS_HEARTBEAT
		print 'available space: %6.3f TB' % NTB
		print '-------------------------------------'
		first_iteration = False

	# Write remaining space to EPICS
	caput(EPICS_VAR, NTB)

	# Update heartbeat
	myheartbeat = caget(EPICS_HEARTBEAT)
	if myheartbeat == None:
		print 'Failed to read hearbeat value from EPICS variable: %s setting to %d' % (EPICS_HEARTBEAT, heartbeat)
	else:
		heartbeat = myheartbeat

	heartbeat += 1
	if heartbeat>1: heartbeat = 0
	caput(EPICS_HEARTBEAT, heartbeat)

	if VERBOSE :
		print '%s : %6.3f TB' % (EPICS_VAR, NTB)

	# Sleep
	time.sleep(PERIOD)