7 from datetime import datetime,timedelta
12 from pcucontrol import reboot
14 from pcucontrol.util import command
15 from monitor import config
17 from monitor.database.info.model import *
19 from monitor.sources import comon
20 from monitor.wrapper import plc, plccache
23 from monitor.common import nmap_port_status, email_exception
25 COMON_COTOPURL= "http://summer.cs.princeton.edu/status/tabulator.cgi?" + \
26 "table=table_nodeview&" + \
27 "dumpcols='name,resptime,sshstatus,uptime,lastcotop,cpuspeed,memsize,disksize'&" + \
30 api = plc.getAuthAPI()
31 plc_lock = threading.Lock()
40 #print "GetPCU from PLC %s" % pcuname
41 l_pcu = plc.GetPCUs({'pcu_id' : pcuname})
47 #print "GetPCU from file %s" % pcuname
48 l_pcus = plccache.l_pcus
50 if i['pcu_id'] == pcuname:
59 def get_nodes(node_ids):
63 l_node = plc.getNodes(node_ids, ['hostname', 'last_contact', 'node_id', 'ports'])
66 plc_nodes = plccache.l_nodes
68 if n['node_id'] in node_ids:
80 def get_plc_pcu_values(pcuname):
82 Try to contact PLC to get the PCU info.
83 If that fails, try a backup copy from the last run.
84 If that fails, return None
88 l_pcu = get_pcu(pcuname)
91 site_id = l_pcu['site_id']
92 node_ids = l_pcu['node_ids']
93 l_node = get_nodes(node_ids)
95 if l_node is not None:
97 values[node['hostname']] = node['ports'][0]
99 values['nodenames'] = [node['hostname'] for node in l_node]
101 # NOTE: this is for a dry run later. It doesn't matter which node.
102 values['node_id'] = l_node[0]['node_id']
110 class ScanInterface(object):
113 primarykey = 'hostname'
115 def __init__(self, round=1):
119 def __getattr__(self, name):
120 if 'collect' in name or 'record' in name:
121 method = getattr(self, name, None)
123 raise Exception("No such method %s" % name)
126 raise Exception("No such method %s" % name)
128 def collect(self, nodename, data):
131 def record(self, request, (nodename, values) ):
138 fbnodesync = self.syncclass.findby_or_create(
139 #if_new_set={'round' : self.round},
140 **{ self.primarykey : nodename})
141 # NOTE: This code will either add a new record for the new self.round,
142 # OR it will find the previous value, and update it with new information.
143 # The data that is 'lost' is not that important, b/c older
144 # history still exists.
145 fbrec = self.recordclass.findby_or_create(
146 **{ self.primarykey:nodename})
148 fbrec.set( **values )
152 fbnodesync.round = self.round
155 print "%d %s %s" % (self.count, nodename, values)
160 email_exception(str(nodename))
161 print traceback.print_exc()
164 class ScanNodeInternal(ScanInterface):
165 recordclass = FindbadNodeRecord
166 #syncclass = FindbadNodeRecordSync
168 primarykey = 'hostname'
170 def collectNMAP(self, nodename, cohash):
171 #### RUN NMAP ###############################
174 print "nmap -oG - -P0 -p22,80,806 %s | grep Host:" % nodename
175 (oval,eval) = nmap.run_noexcept("nmap -oG - -P0 -p22,80,806 %s | grep Host:" % nodename)
176 # NOTE: an empty / error value for oval, will still work.
177 (values['port_status'], continue_probe) = nmap_port_status(oval)
179 values['date_checked'] = datetime.now()
181 return (nodename, values)
183 def collectInternal(self, nodename, cohash):
184 ### RUN PING ######################
186 (oval,errval) = ping.run_noexcept("ping -c 1 -q %s | grep rtt" % nodename)
193 values['ping_status'] = False
195 values['ping_status'] = True
198 for port in [22, 806]:
199 ssh = command.SSH('root', nodename, port)
201 (oval, errval) = ssh.run_noexcept2(""" <<\EOF
203 echo ' "kernel_version":"'`uname -a`'",'
204 echo ' "bmlog":"'`ls /tmp/bm.log`'",'
205 echo ' "bootcd_version":"'`cat /mnt/cdrom/bootme/ID`'",'
206 echo ' "nm_status":"'`ps ax | grep nm.py | grep -v grep`'",'
207 echo ' "dns_status":"'`host boot.planet-lab.org 2>&1`'",'
208 echo ' "princeton_comon_dir":"'`ls -d /vservers/princeton_comon`'",'
210 ID=`grep princeton_comon /etc/passwd | awk -F : '{if ( $3 > 500 ) { print $3}}'`
211 echo ' "princeton_comon_running":"'`ls -d /proc/virtual/$ID`'",'
212 echo ' "princeton_comon_procs":"'`vps ax | grep $ID | grep -v grep | wc -l`'",'
213 echo ' "fs_status":"'`grep proc /proc/mounts | grep ro, ; if [ -x /usr/bin/timeout.pl ] ; then timeout.pl 20 touch /var/log/monitor 2>&1 ; if [ -d /vservers/ ] ; then timeout.pl 20 touch /vservers/monitor.log 2>&1 ; fi ; fi`'",'
214 echo ' "rpm_version":"'`if [ -x /usr/bin/timeout.pl ] ; then timeout.pl 30 rpm -q NodeManager ; fi`'",'
215 echo ' "rpm_versions":"'`if [ -x /usr/bin/timeout.pl ] ; then timeout.pl 45 rpm -q -a ; fi`'",'
219 values['ssh_error'] = errval
221 #print "OVAL: %s" % oval
222 values.update(eval(oval))
223 values['ssh_portused'] = port
226 values.update({'kernel_version': "", 'bmlog' : "", 'bootcd_version' : '',
232 'princeton_comon_dir' : "",
233 'princeton_comon_running' : "",
234 'princeton_comon_procs' : "", 'ssh_portused' : None})
236 print traceback.print_exc()
239 values['fs_status'] = ""
240 print "ALLVERSIONS: %s %s" % (nodename, values['rpm_versions'])
242 print "RPMVERSION: %s %s" % (nodename, values['rpm_version'])
243 ### RUN SSH ######################
244 b_getbootcd_id = True
246 oval = values['kernel_version']
247 if "2.6.17" in oval or "2.6.2" in oval:
248 values['ssh_status'] = True
249 values['observed_category'] = 'PROD'
250 if "bm.log" in values['bmlog']:
251 values['observed_status'] = 'DEBUG'
253 values['observed_status'] = 'BOOT'
254 elif "2.6.12" in oval or "2.6.10" in oval:
255 values['ssh_status'] = True
256 values['observed_category'] = 'OLDPROD'
257 if "bm.log" in values['bmlog']:
258 values['observed_status'] = 'DEBUG'
260 values['observed_status'] = 'BOOT'
262 # NOTE: on 2.6.8 kernels, with 4.2 bootstrapfs, the chroot
263 # command fails. I have no idea why.
264 elif "2.4" in oval or "2.6.8" in oval:
265 b_getbootcd_id = False
266 values['ssh_status'] = True
267 values['observed_category'] = 'OLDBOOTCD'
268 values['observed_status'] = 'DEBUG'
270 values['ssh_status'] = True
271 values['observed_category'] = 'UNKNOWN'
272 if "bm.log" in values['bmlog']:
273 values['observed_status'] = 'DEBUG'
275 values['observed_status'] = 'BOOT'
278 b_getbootcd_id = False
279 values['ssh_status'] = False
280 values['observed_category'] = 'ERROR'
281 values['observed_status'] = 'DOWN'
283 values['ssh_error'] = val
284 values['kernel_version'] = ""
287 # try to get BootCD for all nodes that are not 2.4 nor inaccessible
288 oval = values['bootcd_version']
290 values['bootcd_version'] = oval
291 if "v2" in oval and \
292 ( nodename is not "planetlab1.cs.unc.edu" and \
293 nodename is not "planetlab2.cs.unc.edu" ):
294 values['observed_category'] = 'OLDBOOTCD'
296 values['bootcd_version'] = ""
298 values['bootcd_version'] = ""
300 oval = values['nm_status']
302 values['nm_status'] = "Y"
304 values['nm_status'] = "N"
306 continue_slice_check = True
307 oval = values['princeton_comon_dir']
308 if "princeton_comon_dir" in oval:
309 values['princeton_comon_dir'] = True
311 values['princeton_comon_dir'] = False
312 continue_slice_check = False
314 if continue_slice_check:
315 oval = values['princeton_comon_running']
316 if len(oval) > len('/proc/virtual/'):
317 values['princeton_comon_running'] = True
319 values['princeton_comon_running'] = False
320 continue_slice_check = False
322 values['princeton_comon_running'] = False
324 if continue_slice_check:
325 oval = values['princeton_comon_procs']
326 values['princeton_comon_procs'] = int(oval)
328 values['princeton_comon_procs'] = None
331 if nodename in cohash:
332 values['comon_stats'] = cohash[nodename]
334 values['comon_stats'] = {'resptime': '-1',
341 # include output value
342 ### GET PLC NODE ######################
346 d_node = plccache.GetNodeByName(nodename)
347 #d_node = plc.getNodes({'hostname': nodename}, ['pcu_ids', 'site_id',
348 # 'date_created', 'last_updated',
349 # 'last_contact', 'boot_state', 'nodegroup_ids'])[0]
351 traceback.print_exc()
353 values['plc_node_stats'] = d_node
355 ##### NMAP ###################
356 (n, v) = self.collectNMAP(nodename, None)
359 ### GET PLC PCU ######################
363 pcu = d_node['pcu_ids']
367 site_id = d_node['site_id']
369 values['plc_pcuid'] = d_pcu
371 ### GET PLC SITE ######################
374 values['loginbase'] = ""
376 d_site = plccache.GetSitesById([ site_id ])[0]
377 #d_site = plc.getSites({'site_id': site_id},
378 # ['max_slices', 'slice_ids', 'node_ids', 'login_base'])[0]
379 values['loginbase'] = d_site['login_base']
381 traceback.print_exc()
384 values['plc_site_stats'] = d_site
385 values['date_checked'] = datetime.now()
387 print traceback.print_exc()
389 return (nodename, values)
391 def internalprobe(hostname):
392 #fbsync = FindbadNodeRecordSync.findby_or_create(hostname="global",
393 # if_new_set={'round' : 1})
394 scannode = ScanNodeInternal() # fbsync.round)
396 (nodename, values) = scannode.collectInternal(hostname, {})
397 scannode.record(None, (nodename, values))
401 print traceback.print_exc()
404 def externalprobe(hostname):
405 #fbsync = FindbadNodeRecordSync.findby_or_create(hostname="global",
406 # if_new_set={'round' : 1})
407 scannode = ScanNodeInternal() # fbsync.round)
409 (nodename, values) = scannode.collectNMAP(hostname, {})
410 scannode.record(None, (nodename, values))
414 print traceback.print_exc()
417 class ScanPCU(ScanInterface):
418 recordclass = FindbadPCURecord
420 primarykey = 'plc_pcuid'
422 def collectInternal(self, pcuname, cohash):
424 continue_probe = True
426 values = {'reboot_trial_status' : 'novalue'}
427 ### GET PCU ######################
431 v = get_plc_pcu_values(pcuname)
432 if v['hostname'] is not None: v['hostname'] = v['hostname'].strip()
433 if v['ip'] is not None: v['ip'] = v['ip'].strip()
436 values['plc_pcu_stats'] = v
438 continue_probe = False
441 traceback.print_exc()
442 continue_probe = False
444 if b_except or not continue_probe: return (None, None, None)
446 #### RUN NMAP ###############################
449 print "nmap -oG - -P0 -p22,23,80,443,5869,9100,16992 %s | grep Host:" % reboot.pcu_name(values['plc_pcu_stats'])
450 (oval,eval) = nmap.run_noexcept("nmap -oG - -P0 -p22,23,80,443,5869,9100,16992 %s | grep Host:" % reboot.pcu_name(values['plc_pcu_stats']))
451 # NOTE: an empty / error value for oval, will still work.
452 (values['port_status'], continue_probe) = nmap_port_status(oval)
454 values['port_status'] = None
456 #### COMPLETE ENTRY #######################
458 values['entry_complete'] = []
459 #if values['protocol'] is None or values['protocol'] is "":
460 # values['entry_complete'] += ["protocol"]
461 if values['plc_pcu_stats']['model'] is None or values['plc_pcu_stats']['model'] is "":
462 values['entry_complete'] += ["model"]
463 # Cannot continue due to this condition
464 continue_probe = False
466 if values['plc_pcu_stats']['password'] is None or values['plc_pcu_stats']['password'] is "":
467 values['entry_complete'] += ["password"]
468 # Cannot continue due to this condition
469 continue_probe = False
471 if len(values['entry_complete']) > 0:
472 continue_probe = False
474 if values['plc_pcu_stats']['hostname'] is None or values['plc_pcu_stats']['hostname'] is "":
475 values['entry_complete'] += ["hostname"]
476 if values['plc_pcu_stats']['ip'] is None or values['plc_pcu_stats']['ip'] is "":
477 values['entry_complete'] += ["ip"]
479 # If there are no nodes associated with this PCU, then we cannot continue.
480 if len(values['plc_pcu_stats']['node_ids']) == 0:
481 continue_probe = False
482 values['entry_complete'] += ['nodeids']
485 #### DNS and IP MATCH #######################
486 if values['plc_pcu_stats']['hostname'] is not None and values['plc_pcu_stats']['hostname'] is not "" and \
487 values['plc_pcu_stats']['ip'] is not None and values['plc_pcu_stats']['ip'] is not "":
489 ipaddr = socket.gethostbyname(values['plc_pcu_stats']['hostname'])
490 if ipaddr == values['plc_pcu_stats']['ip']:
491 values['dns_status'] = "DNS-OK"
493 values['dns_status'] = "DNS-MISMATCH"
494 values['plc_pcu_stats']['hostname'] = values['plc_pcu_stats']['ip']
496 except Exception, err:
497 values['dns_status'] = "DNS-NOENTRY"
498 values['plc_pcu_stats']['hostname'] = values['plc_pcu_stats']['ip']
500 if values['plc_pcu_stats']['ip'] is not None and values['plc_pcu_stats']['ip'] is not "":
501 values['dns_status'] = "NOHOSTNAME"
502 values['plc_pcu_stats']['hostname'] = values['plc_pcu_stats']['ip']
504 values['dns_status'] = "NO-DNS-OR-IP"
505 values['plc_pcu_stats']['hostname'] = "No_entry_in_DB"
506 continue_probe = False
509 ###### DRY RUN ############################
510 if continue_probe and 'node_ids' in values['plc_pcu_stats'] and \
511 len(values['plc_pcu_stats']['node_ids']) > 0:
512 rb_ret = reboot.reboot_test_new(values['plc_pcu_stats']['nodenames'][0],
515 rb_ret = "Not_Run" # No nodes to test"
517 values['reboot_trial_status'] = rb_ret
520 print "____________________________________"
523 print "____________________________________"
524 errors['traceback'] = traceback.format_exc()
525 print errors['traceback']
526 values['reboot_trial_status'] = str(errors['traceback'])
529 values['entry_complete']=" ".join(values['entry_complete'])
531 values['date_checked'] = datetime.now()
532 return (pcuname, values)