X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;f=findbad.py;h=6420f99c1efdafb498c67b68deb47c7193ba31a0;hb=refs%2Fheads%2F1.0;hp=fa4c76a5c67b2346f87b4e9226004e0a9a9c8392;hpb=b93294f1bc40cdfe7f5e632577f8d19b24f97c37;p=monitor.git

diff --git a/findbad.py b/findbad.py
index fa4c76a..6420f99 100755
--- a/findbad.py
+++ b/findbad.py
@@ -4,27 +4,15 @@ import os
 import sys
 import string
 import time
+import config
+import util.file
 
-from config import config
-from optparse import OptionParser
-parser = OptionParser()
-parser.set_defaults(filename=None, increment=False, dbname="findbadnodes", cachenodes=False)
-parser.add_option("-f", "--nodelist", dest="filename", metavar="FILE", 
-					help="Provide the input file for the node list")
-parser.add_option("", "--cachenodes", action="store_true",
-					help="Cache node lookup from PLC")
-parser.add_option("", "--dbname", dest="dbname", metavar="FILE", 
-					help="Specify the name of the database to which the information is saved")
-parser.add_option("-i", "--increment", action="store_true", dest="increment", 
-					help="Increment round number to force refresh or retry")
-config = config(parser)
-config.parse_args()
 
 # QUERY all nodes.
 COMON_COTOPURL= "http://summer.cs.princeton.edu/status/tabulator.cgi?" + \
-					"table=table_nodeview&" + \
-				    "dumpcols='name,resptime,sshstatus,uptime,lastcotop'&" + \
-				    "formatcsv"
+				"table=table_nodeview&" + \
+				"dumpcols='name,resptime,sshstatus,uptime,lastcotop,cpuspeed,memsize,disksize'&" + \
+				"formatcsv"
 				    #"formatcsv&" + \
 					#"select='lastcotop!=0'"
 
@@ -35,16 +23,20 @@ externalState = {'round': round, 'nodes': {}}
 count = 0
 
 
-import soltesz
-import plc
+import database
+import moncommands 
 import comon
 import threadpool
 import syncplcdb
+from nodequery import verify,query_to_dict,node_select
+import traceback
+import plc
+api = plc.getAuthAPI()
 
 def collectPingAndSSH(nodename, cohash):
 	### RUN PING ######################
-	ping = soltesz.CMD()
-	(oval,eval) = ping.run_noexcept("ping -c 1 -q %s | grep rtt" % nodename)
+	ping = moncommands.CMD()
+	(oval,errval) = ping.run_noexcept("ping -c 1 -q %s | grep rtt" % nodename)
 
 	values = {}
 
@@ -54,31 +46,70 @@ def collectPingAndSSH(nodename, cohash):
 	else:
 		values['ping'] = "PING"
 
-	#uptime = soltesz.SSH('root', nodename)
-	#(oval,eval) = uptime.run_noexcept("uptime | awk '{print $3,$4}' | tr , ' '")
+	try:
+		for port in [22, 806]: 
+			ssh = moncommands.SSH('root', nodename, port)
+
+			(oval, errval) = ssh.run_noexcept2(""" <<\EOF
+				echo "{"
+				echo '  "kernel":"'`uname -a`'",'
+				echo '  "bmlog":"'`ls /tmp/bm.log`'",'
+				echo '  "bootcd":"'`cat /mnt/cdrom/bootme/ID`'",'
+				echo '  "nm":"'`ps ax | grep nm.py | grep -v grep`'",'
+				echo '  "readonlyfs":"'`touch /var/log/monitor 2>&1 ; touch /vservers/monitor.log 2>&1`'",'
+				echo '  "dns":"'`host boot.planet-lab.org 2>&1`'",'
+				echo '  "princeton_comon":"'`ls -d /vservers/princeton_comon`'",'
+
+				ID=`grep princeton_comon /etc/passwd | awk -F : '{if ( $3 > 500 ) { print $3}}'` 
+
+				echo '  "princeton_comon_running":"'`ls -d /proc/virtual/$ID`'",'
+				echo '  "princeton_comon_procs":"'`vps ax | grep $ID | grep -v grep | wc -l`'",'
+				echo "}"
+EOF			""")
+			
+			if len(oval) > 0:
+				values.update(eval(oval))
+				values['sshport'] = port
+				break
+			else:
+				values.update({'kernel': "", 'bmlog' : "", 'bootcd' : '', 
+							 	'nm' : '', 
+								'readonlyfs' : '',
+								'dns' : '',
+								'princeton_comon' : '', 
+								'princeton_comon_running' : '', 
+								'princeton_comon_procs' : '', 'sshport' : None})
+	except:
+		print traceback.print_exc()
+		from nodecommon import email_exception
+		email_exception()
+		sys.exit(1)
 
 	### RUN SSH ######################
 	b_getbootcd_id = True
-	ssh = soltesz.SSH('root', nodename)
-	oval = ""
-	eval = ""
-	(oval, eval) = ssh.run_noexcept('echo `uname -a ; ls /tmp/bm.log`')
-	val = oval
+	#ssh = moncommands.SSH('root', nodename)
+	#oval = ""
+	#errval = ""
+	#(oval, errval) = ssh.run_noexcept('echo `uname -a ; ls /tmp/bm.log`')
+
+	oval = values['kernel']
 	if "2.6.17" in oval or "2.6.2" in oval:
 		values['ssh'] = 'SSH'
-		values['category'] = 'ALPHA'
-		if "bm.log" in oval:
+		values['category'] = 'PROD'
+		if "bm.log" in values['bmlog']:
 			values['state'] = 'DEBUG'
 		else:
 			values['state'] = 'BOOT'
 	elif "2.6.12" in oval or "2.6.10" in oval:
 		values['ssh'] = 'SSH'
-		values['category'] = 'PROD'
-		if "bm.log" in oval:
+		values['category'] = 'OLDPROD'
+		if "bm.log" in values['bmlog']:
 			values['state'] = 'DEBUG'
 		else:
 			values['state'] = 'BOOT'
-	elif "2.4" in oval:
+	
+	# NOTE: on 2.6.8 kernels, with 4.2 bootstrapfs, the chroot command fails.  I have no idea why.
+	elif "2.4" in oval or "2.6.8" in oval:
 		b_getbootcd_id = False
 		values['ssh'] = 'SSH'
 		values['category'] = 'OLDBOOTCD'
@@ -86,7 +117,7 @@ def collectPingAndSSH(nodename, cohash):
 	elif oval != "":
 		values['ssh'] = 'SSH'
 		values['category'] = 'UNKNOWN'
-		if "bm.log" in oval:
+		if "bm.log" in values['bmlog']:
 			values['state'] = 'DEBUG'
 		else:
 			values['state'] = 'BOOT'
@@ -96,17 +127,18 @@ def collectPingAndSSH(nodename, cohash):
 		values['ssh'] = 'NOSSH'
 		values['category'] = 'ERROR'
 		values['state'] = 'DOWN'
-		val = eval.strip()
+		val = errval.strip()
+		values['kernel'] = val
 
-	values['kernel'] = val
+	#values['kernel'] = val
 
 	if b_getbootcd_id:
 		# try to get BootCD for all nodes that are not 2.4 nor inaccessible
-		(oval, eval) = ssh.run_noexcept('cat /mnt/cdrom/bootme/ID')
-		val = oval
-		if "BootCD" in val:
-			values['bootcd'] = val
-			if "v2" in val and \
+		#(oval, errval) = ssh.run_noexcept('cat /mnt/cdrom/bootme/ID')
+		oval = values['bootcd']
+		if "BootCD" in oval:
+			values['bootcd'] = oval
+			if "v2" in oval and \
 				( nodename is not "planetlab1.cs.unc.edu" and \
 				  nodename is not "planetlab2.cs.unc.edu" ):
 				values['category'] = 'OLDBOOTCD'
@@ -117,6 +149,41 @@ def collectPingAndSSH(nodename, cohash):
 
 	# TODO: get bm.log for debug nodes.
 	# 'zcat /tmp/bm.log'
+	
+	#(oval, errval) = ssh.run_noexcept('ps ax | grep nm.py | grep -v grep')
+	oval = values['nm']
+	if "nm.py" in oval:
+		values['nm'] = "Y"
+	else:
+		values['nm'] = "N"
+
+	continue_slice_check = True
+	#(oval, errval) = ssh.run_noexcept('ls -d /vservers/princeton_comon')
+	oval = values['princeton_comon']
+	if "princeton_comon" in oval:
+		values['princeton_comon'] = "Y"
+	else:
+		values['princeton_comon'] = "N"
+		continue_slice_check = False
+
+	if continue_slice_check:
+		#(oval, errval) = ssh.run_noexcept('ID=`grep princeton_comon /etc/passwd | awk -F : "{if ( \\\$3 > 500 ) { print \\\$3}}"`; ls -d /proc/virtual/$ID')
+		oval = values['princeton_comon_running']
+		if len(oval) > len('/proc/virtual/'):
+			values['princeton_comon_running'] = "Y"
+		else:
+			values['princeton_comon_running'] = "N"
+			continue_slice_check = False
+	else:
+		values['princeton_comon_running'] = "-"
+		
+	if continue_slice_check:
+		#(oval, errval) = ssh.run_noexcept('ID=`grep princeton_comon /etc/passwd | awk -F : "{if ( \\\$3 > 500 ) { print \\\$3}}"`; vps ax | grep $ID | grep -v grep | wc -l')
+		oval = values['princeton_comon_procs']
+		values['princeton_comon_procs'] = oval
+	else:
+		values['princeton_comon_procs'] = "-"
+
 		
 	if nodename in cohash: 
 		values['comonstats'] = cohash[nodename]
@@ -124,20 +191,22 @@ def collectPingAndSSH(nodename, cohash):
 		values['comonstats'] = {'resptime':  '-1', 
 								'uptime':    '-1',
 								'sshstatus': '-1', 
-								'lastcotop': '-1'}
+								'lastcotop': '-1',
+								'cpuspeed' : "null",
+								'disksize' : 'null',
+								'memsize'  : 'null'}
 	# include output value
 	### GET PLC NODE ######################
 	b_except = False
 	plc_lock.acquire()
 
 	try:
-		d_node = plc.getNodes({'hostname': nodename}, ['pcu_ids', 'site_id', 'last_contact', 'boot_state', 'nodegroup_ids'])
+		d_node = plc.getNodes({'hostname': nodename}, ['pcu_ids', 'site_id', 'date_created', 'last_updated', 'last_contact', 'boot_state', 'nodegroup_ids'])
 	except:
 		b_except = True
-		import traceback
-		b_except = True
-		import traceback
 		traceback.print_exc()
+		from nodecommon import email_exception
+		email_exception()
 
 	plc_lock.release()
 	if b_except: return (None, None)
@@ -151,13 +220,15 @@ def collectPingAndSSH(nodename, cohash):
 			values['pcu'] = "NOPCU"
 		site_id = d_node[0]['site_id']
 		last_contact = d_node[0]['last_contact']
-		nodegroups = d_node[0]['nodegroup_ids']
+		nodegroups = [ i['groupname'] for i in api.GetNodeGroups(d_node[0]['nodegroup_ids']) ]
 		values['plcnode'] = {'status' : 'SUCCESS', 
 							'pcu_ids': pcu, 
 							'boot_state' : d_node[0]['boot_state'],
 							'site_id': site_id,
 							'nodegroups' : nodegroups,
-							'last_contact': last_contact}
+							'last_contact': last_contact,
+							'date_created': d_node[0]['date_created'],
+							'last_updated': d_node[0]['last_updated']}
 	else:
 		values['pcu']     = "UNKNOWN"
 		values['plcnode'] = {'status' : "GN_FAILED"}
@@ -172,8 +243,9 @@ def collectPingAndSSH(nodename, cohash):
 							['max_slices', 'slice_ids', 'node_ids', 'login_base'])
 	except:
 		b_except = True
-		import traceback
 		traceback.print_exc()
+		from nodecommon import email_exception
+		email_exception()
 
 	plc_lock.release()
 	if b_except: return (None, None)
@@ -207,7 +279,8 @@ def recordPingAndSSH(request, result):
 
 		count += 1
 		print "%d %s %s" % (count, nodename, externalState['nodes'][nodename]['values'])
-		soltesz.dbDump(config.dbname, externalState)
+		if count % 20 == 0:
+			database.dbDump(config.dbname, externalState)
 
 # this will be called when an exception occurs within a thread
 def handle_exception(request, result):
@@ -242,10 +315,16 @@ def checkAndRecordState(l_nodes, cohash):
 			pass
 
 	# WAIT while all the work requests are processed.
+	begin = time.time()
 	while 1:
 		try:
 			time.sleep(1)
 			tp.poll()
+			# if more than two hours
+			if time.time() - begin > (60*60*1.5):
+				print "findbad.py has run out of time!!!!!!"
+				database.dbDump(config.dbname, externalState)
+				os._exit(1)
 		except KeyboardInterrupt:
 			print "Interrupted!"
 			break
@@ -253,12 +332,14 @@ def checkAndRecordState(l_nodes, cohash):
 			print "All results collected."
 			break
 
+	database.dbDump(config.dbname, externalState)
+
 
 
 def main():
 	global externalState
 
-	externalState = soltesz.if_cached_else(1, config.dbname, lambda : externalState) 
+	externalState = database.if_cached_else(1, config.dbname, lambda : externalState) 
 
 	if config.increment:
 		# update global round number to force refreshes across all nodes
@@ -270,14 +351,30 @@ def main():
 	cotop_url = COMON_COTOPURL
 
 	# history information for all nodes
+	#cohash = {}
 	cohash = cotop.coget(cotop_url)
 	l_nodes = syncplcdb.create_plcdb()
-	if config.filename:
-		f_nodes = config.getListFromFile(config.filename)
+	if config.nodelist:
+		f_nodes = util.file.getListFromFile(config.nodelist)
 		l_nodes = filter(lambda x: x['hostname'] in f_nodes, l_nodes)
-
+	elif config.node:
+		f_nodes = [config.node]
+		l_nodes = filter(lambda x: x['hostname'] in f_nodes, l_nodes)
+	elif config.nodegroup:
+		ng = api.GetNodeGroups({'groupname' : config.nodegroup})
+		l_nodes = api.GetNodes(ng[0]['node_ids'])
+	elif config.site:
+		site = api.GetSites(config.site)
+		l_nodes = api.GetNodes(site[0]['node_ids'], ['hostname'])
+		
 	l_nodes = [node['hostname'] for node in l_nodes]
 
+	# perform this query after the above options, so that the filter above
+	# does not break.
+	if config.nodeselect:
+		fb = database.dbLoad("findbad")
+		l_nodes = node_select(config.nodeselect, fb['nodes'].keys(), fb)
+
 	print "fetching %s hosts" % len(l_nodes)
 
 	checkAndRecordState(l_nodes, cohash)
@@ -286,12 +383,29 @@ def main():
 
 
 if __name__ == '__main__':
+	import parser as parsermodule
+
+	parser = parsermodule.getParser(['nodesets'])
+
+	parser.set_defaults( increment=False, dbname="findbad", cachenodes=False)
+	parser.add_option("", "--cachenodes", action="store_true",
+						help="Cache node lookup from PLC")
+	parser.add_option("", "--dbname", dest="dbname", metavar="FILE", 
+						help="Specify the name of the database to which the information is saved")
+	parser.add_option("-i", "--increment", action="store_true", dest="increment", 
+						help="Increment round number to force refresh or retry")
+
+	parser = parsermodule.getParser(['defaults'], parser)
+	
+	cfg = parsermodule.parse_args(parser)
+
 	try:
 		main()
 	except Exception, err:
-		import traceback
 		print traceback.print_exc()
+		from nodecommon import email_exception
+		email_exception()
 		print "Exception: %s" % err
 		print "Saving data... exitting."
-		soltesz.dbDump(config.dbname, externalState)
+		database.dbDump(config.dbname, externalState)
 		sys.exit(0)