from monitor.sources import comon
from monitor.wrapper import plc, plccache
+from monitor.scanapi import *
from nodequery import verify,query_to_dict,node_select
import traceback
+from nodecommon import nmap_port_status
#print "starting sqlfindbad.py"
# QUERY all nodes.
global_round = round
count = 0
-def collectPingAndSSH(nodename, cohash):
- ### RUN PING ######################
- ping = command.CMD()
- (oval,errval) = ping.run_noexcept("ping -c 1 -q %s | grep rtt" % nodename)
-
- try:
- values = {}
-
- if oval == "":
- # An error occurred
- values['ping'] = "NOPING"
- else:
- values['ping'] = "PING"
-
- try:
- for port in [22, 806]:
- ssh = command.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`'",'
- 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 """)
-
- values['ssherror'] = errval
- if len(oval) > 0:
- #print "OVAL: %s" % oval
- 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()
- sys.exit(1)
-
- ### RUN SSH ######################
- b_getbootcd_id = True
- #ssh = command.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'] = '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'] = 'OLDPROD'
- if "bm.log" in values['bmlog']:
- values['state'] = 'DEBUG'
- else:
- values['state'] = 'BOOT'
-
- # 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'
- values['state'] = 'DEBUG'
- elif oval != "":
- values['ssh'] = 'SSH'
- values['category'] = 'UNKNOWN'
- if "bm.log" in values['bmlog']:
- values['state'] = 'DEBUG'
- else:
- values['state'] = 'BOOT'
- else:
- # An error occurred.
- b_getbootcd_id = False
- values['ssh'] = 'NOSSH'
- values['category'] = 'ERROR'
- values['state'] = 'DOWN'
- val = errval.strip()
- values['ssherror'] = val
- values['kernel'] = ""
-
- #values['kernel'] = val
-
- if b_getbootcd_id:
- # try to get BootCD for all nodes that are not 2.4 nor inaccessible
- #(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'
- else:
- values['bootcd'] = ""
- else:
- values['bootcd'] = ""
-
- # 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'] = True
- else:
- values['princeton_comon'] = False
- 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'] = True
- else:
- values['princeton_comon_running'] = False
- continue_slice_check = False
- else:
- values['princeton_comon_running'] = False
-
- 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'] = int(oval)
- else:
- values['princeton_comon_procs'] = None
-
-
- if nodename in cohash:
- values['comonstats'] = cohash[nodename]
- else:
- values['comonstats'] = {'resptime': '-1',
- 'uptime': '-1',
- 'sshstatus': '-1',
- 'lastcotop': '-1',
- 'cpuspeed' : "null",
- 'disksize' : 'null',
- 'memsize' : 'null'}
- # include output value
- ### GET PLC NODE ######################
- plc_lock.acquire()
- d_node = None
- try:
- d_node = plc.getNodes({'hostname': nodename}, ['pcu_ids', 'site_id', 'date_created',
- 'last_updated', 'last_contact', 'boot_state', 'nodegroup_ids'])[0]
- except:
- traceback.print_exc()
- plc_lock.release()
- values['plcnode'] = d_node
-
- ### GET PLC PCU ######################
- site_id = -1
- d_pcu = None
- if d_node:
- pcu = d_node['pcu_ids']
- if len(pcu) > 0:
- d_pcu = pcu[0]
-
- site_id = d_node['site_id']
-
- values['pcu'] = d_pcu
-
- ### GET PLC SITE ######################
- plc_lock.acquire()
- d_site = None
- values['loginbase'] = ""
- try:
- d_site = plc.getSites({'site_id': site_id},
- ['max_slices', 'slice_ids', 'node_ids', 'login_base'])[0]
- values['loginbase'] = d_site['login_base']
- except:
- traceback.print_exc()
- plc_lock.release()
-
- values['plcsite'] = d_site
- values['date_checked'] = time.time()
- except:
- print traceback.print_exc()
-
- return (nodename, values)
-
-def recordPingAndSSH(request, result):
- global global_round
- global count
- (nodename, values) = result
-
- try:
- if values is not None:
- fbsync = FindbadNodeRecordSync.findby_or_create(hostname="global",
- if_new_set={'round' : global_round})
- global_round = fbsync.round
- fbnodesync = FindbadNodeRecordSync.findby_or_create(hostname=nodename,
- if_new_set={'round' : global_round})
-
- # NOTE: This code will either add a new record for the new global_round,
- # OR it will find the previous value, and update it
- # with new information.
- # The data that is 'lost' is not that important, b/c older
- # history still exists.
- fbrec = FindbadNodeRecord.findby_or_create(
- round=global_round,
- hostname=nodename)
- before = fbrec.to_dict()
- print "BEFORE, ", before
- fbrec.flush()
- time.sleep(2)
- print "Setting VALUES"
- fbrec.set( date_checked=datetime.fromtimestamp(values['date_checked']),
- loginbase=values['loginbase'],
- kernel_version=values['kernel'],
- bootcd_version=values['bootcd'],
- nm_status=values['nm'],
- fs_status=values['readonlyfs'],
- dns_status=values['dns'],
- princeton_comon_dir=values['princeton_comon'],
- princeton_comon_running=values['princeton_comon_running'],
- princeton_comon_procs=values['princeton_comon_procs'],
- plc_node_stats = values['plcnode'],
- plc_site_stats = values['plcsite'],
- plc_pcuid = values['pcu'],
- comon_stats = values['comonstats'],
- ping_status = (values['ping'] == "PING"),
- ssh_portused = values['sshport'],
- ssh_status = (values['ssh'] == "SSH"),
- ssh_error = values['ssherror'],
- observed_status = values['state'],
- observed_category = values['category'])
- after = fbrec.to_dict()
- print "AFTER , ", after
-
- for v in before.keys():
- if before[v] == after[v]:
- print "SAME FOR KEY %s" % v
- print "%s : %s\t%s" % ( v, before[v], after[v] )
-
- fbrec.flush()
- fbnodesync.round = global_round
- fbnodesync.flush()
- fbsync.flush()
-
- count += 1
- print "%d %s %s" % (count, nodename, values)
- except:
- print "ERROR:"
- print traceback.print_exc()
-
# this will be called when an exception occurs within a thread
def handle_exception(request, result):
print "Exception occured in request %s" % request.requestID
for i in result:
print "Result: %s" % i
-
-def probe(hostname):
- try:
- (nodename, values) = collectPingAndSSH(hostname, {})
- recordPingAndSSH(None, (nodename, values))
- session.flush()
- return True
- except:
- print traceback.print_exc()
- return False
def checkAndRecordState(l_nodes, cohash):
global count
tp = threadpool.ThreadPool(20)
+ scannode = ScanNodeInternal(global_round)
# CREATE all the work requests
for nodename in l_nodes:
node_round = fbnodesync.round
fbnodesync.flush()
- if node_round < global_round:
+ if node_round < global_round or config.force:
# recreate node stats when refreshed
#print "%s" % nodename
- req = threadpool.WorkRequest(collectPingAndSSH, [nodename, cohash], {},
- None, recordPingAndSSH, handle_exception)
+ req = threadpool.WorkRequest(scannode.collectInternal, [nodename, cohash], {},
+ None, scannode.record, handle_exception)
tp.putRequest(req)
else:
# We just skip it, since it's "up to date"
if config.increment:
# update global round number to force refreshes across all nodes
global_round += 1
- fbsync.round = global_round
-
- fbsync.flush()
cotop = comon.Comon()
# lastcotop measures whether cotop is actually running. this is a better
checkAndRecordState(l_nodes, cohash)
+ if config.increment:
+ # update global round number to force refreshes across all nodes
+ fbsync.round = global_round
+ fbsync.flush()
+
return 0
parser = parsermodule.getParser(['nodesets'])
- parser.set_defaults( increment=False, dbname="findbad", cachenodes=False)
+ parser.set_defaults( increment=False, dbname="findbad", cachenodes=False,
+ force=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.add_option("", "--force", action="store_true", dest="force",
+ help="Force probe without incrementing global 'round'.")
parser = parsermodule.getParser(['defaults'], parser)
git://git.onelab.eu / monitor.git / blobdiff