From: František Dvořák <valtri@civ.zcu.cz>
Date: Wed, 4 Mar 2015 23:04:47 +0000 (+0100)
Subject: Initial database consumer.
X-Git-Url: http://scientific.zcu.cz/git/?a=commitdiff_plain;h=e6b5689757849edab401e131b163e650d6353606;p=hadoop-jobstat.git

Initial database consumer.
---

diff --git a/stat2pbs.py b/stat2pbs.py
new file mode 100755
index 0000000..7986ffb
--- /dev/null
+++ b/stat2pbs.py
@@ -0,0 +1,94 @@
+#! /usr/bin/python2
+
+import getopt
+import sys
+import MySQLdb
+
+dbhost = 'localhost'
+dbname = 'hadoop'
+dbuser = 'hadoop'
+dbpassword = ''
+statfile = 'stat2pbs.txt'
+
+try:
+	opts, args = getopt.getopt(sys.argv[1:], 'hc:s:', ['help', 'config=', 'dbhost=', 'dbname=', 'dbuser=', 'dbpassword=', 'statfile=' ])
+except getopt.GetoptError:
+	print 'Args error'
+	sys.exit(2)
+for opt, arg in opts:
+	if opt in ('-h', '--help'):
+		print "stat2pbs.py [OPTIONS]\n\
+OPTIONS are:\n\
+  -h, --help ............ help message\n\
+  -c, --config .......... config file\n\
+  --dbhost\n\
+  --dbname\n\
+  --dbuser\n\
+  --dbpassword\n\
+  -s, --statfile FILE ... status file with the last timestamp"
+		sys.exit(0)
+	elif opt in ('-c', '--config'):
+		f = open(arg, 'r')
+		for line in f:
+			cfg=line.rstrip().split('=')
+			if cfg[0] == 'dbhost':
+				dbhost = cfg[1]
+			elif cfg[0] == 'dbname':
+				dbname = cfg[1]
+			elif cfg[0] == 'dbuser':
+				dbuser = cfg[1]
+			elif cfg[0] == 'dbpassword':
+				dbpassword = cfg[1]
+			elif cfg[0] == 'debug':
+				debug = int(cfg[1])
+			elif cfg[0] == 'statfile':
+				statfile = int(cfg[1])
+		f.close()
+	elif opt in ('--dbhost'):
+		dbhost = arg
+	elif opt in ('--dbname'):
+		dbname = arg
+	elif opt in ('--dbuser'):
+		dbuser = arg
+	elif opt in ('--dbpassword'):
+		dbpassword = arg
+	elif opt in ('-s', '--statfile'):
+		statfile = arg
+	else:
+		print 'Args error'
+		sys.exit(2)
+
+try:
+	f = open(statfile, 'r')
+	ts = f.readline()
+	f.close()
+	ts = ts.rstrip()
+except IOError:
+	ts = None
+
+print 'Status file: %s' % statfile
+print 'Timestamp: %s' % ts
+
+db = MySQLdb.connect(dbhost, dbuser, dbpassword, dbname)
+st = db.cursor()
+
+# beware of actually running second - rather get only >5 seconds older changes
+if ts:
+	st.execute('SELECT id, name, user, changed FROM job WHERE changed > %s AND TIMESTAMPDIFF(SECOND,changed,NOW()) > 5 ORDER BY CHANGED', ts)
+else:
+	st.execute('SELECT id, name, user, changed FROM job WHERE TIMESTAMPDIFF(SECOND,changed,NOW()) > 5 ORDER BY changed')
+while 1:
+	data = st.fetchone()
+	if data:
+		print '%s, %s, %s' % (data[0], data[2], data[3])
+		ts = data[3]
+	else:
+		break
+
+db.close()
+
+if ts:
+	f = open(statfile, 'w')
+	f.write(str(ts) + '\n')
+	f.close()
+	print 'Written: %s' % ts