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_plcnodes
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(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 ' "fs_status":"'`touch /var/log/monitor 2>&1`'",'
208 echo ' "dns_status":"'`host boot.planet-lab.org 2>&1`'",'
209 echo ' "princeton_comon_dir":"'`ls -d /vservers/princeton_comon`'",'
211 ID=`grep princeton_comon /etc/passwd | awk -F : '{if ( $3 > 500 ) { print $3}}'`
212 echo ' "princeton_comon_running":"'`ls -d /proc/virtual/$ID`'",'
213 echo ' "princeton_comon_procs":"'`vps ax | grep $ID | grep -v grep | wc -l`'",'
217 values['ssh_error'] = errval
219 #print "OVAL: %s" % oval
220 values.update(eval(oval))
221 values['ssh_portused'] = port
224 values.update({'kernel_version': "", 'bmlog' : "", 'bootcd_version' : '',
228 'princeton_comon_dir' : "",
229 'princeton_comon_running' : "",
230 'princeton_comon_procs' : "", 'ssh_portused' : None})
232 print traceback.print_exc()
235 ### RUN SSH ######################
236 b_getbootcd_id = True
238 oval = values['kernel_version']
239 if "2.6.17" in oval or "2.6.2" in oval:
240 values['ssh_status'] = True
241 values['observed_category'] = 'PROD'
242 if "bm.log" in values['bmlog']:
243 values['observed_status'] = 'DEBUG'
245 values['observed_status'] = 'BOOT'
246 elif "2.6.12" in oval or "2.6.10" in oval:
247 values['ssh_status'] = True
248 values['observed_category'] = 'OLDPROD'
249 if "bm.log" in values['bmlog']:
250 values['observed_status'] = 'DEBUG'
252 values['observed_status'] = 'BOOT'
254 # NOTE: on 2.6.8 kernels, with 4.2 bootstrapfs, the chroot
255 # command fails. I have no idea why.
256 elif "2.4" in oval or "2.6.8" in oval:
257 b_getbootcd_id = False
258 values['ssh_status'] = True
259 values['observed_category'] = 'OLDBOOTCD'
260 values['observed_status'] = 'DEBUG'
262 values['ssh_status'] = True
263 values['observed_category'] = 'UNKNOWN'
264 if "bm.log" in values['bmlog']:
265 values['observed_status'] = 'DEBUG'
267 values['observed_status'] = 'BOOT'
270 b_getbootcd_id = False
271 values['ssh_status'] = False
272 values['observed_category'] = 'ERROR'
273 values['observed_status'] = 'DOWN'
275 values['ssh_error'] = val
276 values['kernel_version'] = ""
279 # try to get BootCD for all nodes that are not 2.4 nor inaccessible
280 oval = values['bootcd_version']
282 values['bootcd_version'] = oval
283 if "v2" in oval and \
284 ( nodename is not "planetlab1.cs.unc.edu" and \
285 nodename is not "planetlab2.cs.unc.edu" ):
286 values['observed_category'] = 'OLDBOOTCD'
288 values['bootcd_version'] = ""
290 values['bootcd_version'] = ""
292 oval = values['nm_status']
294 values['nm_status'] = "Y"
296 values['nm_status'] = "N"
298 continue_slice_check = True
299 oval = values['princeton_comon_dir']
300 if "princeton_comon_dir" in oval:
301 values['princeton_comon_dir'] = True
303 values['princeton_comon_dir'] = False
304 continue_slice_check = False
306 if continue_slice_check:
307 oval = values['princeton_comon_running']
308 if len(oval) > len('/proc/virtual/'):
309 values['princeton_comon_running'] = True
311 values['princeton_comon_running'] = False
312 continue_slice_check = False
314 values['princeton_comon_running'] = False
316 if continue_slice_check:
317 oval = values['princeton_comon_procs']
318 values['princeton_comon_procs'] = int(oval)
320 values['princeton_comon_procs'] = None
323 if nodename in cohash:
324 values['comon_stats'] = cohash[nodename]
326 values['comon_stats'] = {'resptime': '-1',
333 # include output value
334 ### GET PLC NODE ######################
338 d_node = plccache.GetNodeByName(nodename)
339 #d_node = plc.getNodes({'hostname': nodename}, ['pcu_ids', 'site_id',
340 # 'date_created', 'last_updated',
341 # 'last_contact', 'boot_state', 'nodegroup_ids'])[0]
343 traceback.print_exc()
345 values['plc_node_stats'] = d_node
347 ##### NMAP ###################
348 (n, v) = self.collectNMAP(nodename, None)
351 ### GET PLC PCU ######################
355 pcu = d_node['pcu_ids']
359 site_id = d_node['site_id']
361 values['plc_pcuid'] = d_pcu
363 ### GET PLC SITE ######################
366 values['loginbase'] = ""
368 d_site = plccache.GetSitesById([ site_id ])[0]
369 #d_site = plc.getSites({'site_id': site_id},
370 # ['max_slices', 'slice_ids', 'node_ids', 'login_base'])[0]
371 values['loginbase'] = d_site['login_base']
373 traceback.print_exc()
376 values['plc_site_stats'] = d_site
377 values['date_checked'] = datetime.now()
379 print traceback.print_exc()
381 return (nodename, values)
383 def internalprobe(hostname):
384 #fbsync = FindbadNodeRecordSync.findby_or_create(hostname="global",
385 # if_new_set={'round' : 1})
386 scannode = ScanNodeInternal() # fbsync.round)
388 (nodename, values) = scannode.collectInternal(hostname, {})
389 scannode.record(None, (nodename, values))
393 print traceback.print_exc()
396 def externalprobe(hostname):
397 #fbsync = FindbadNodeRecordSync.findby_or_create(hostname="global",
398 # if_new_set={'round' : 1})
399 scannode = ScanNodeInternal() # fbsync.round)
401 (nodename, values) = scannode.collectNMAP(hostname, {})
402 scannode.record(None, (nodename, values))
406 print traceback.print_exc()
409 class ScanPCU(ScanInterface):
410 recordclass = FindbadPCURecord
412 primarykey = 'plc_pcuid'
414 def collectInternal(self, pcuname, cohash):
416 continue_probe = True
418 values = {'reboot_trial_status' : 'novalue'}
419 ### GET PCU ######################
423 v = get_plc_pcu_values(pcuname)
424 if v['hostname'] is not None: v['hostname'] = v['hostname'].strip()
425 if v['ip'] is not None: v['ip'] = v['ip'].strip()
428 values['plc_pcu_stats'] = v
430 continue_probe = False
433 traceback.print_exc()
434 continue_probe = False
436 if b_except or not continue_probe: return (None, None, None)
438 #### RUN NMAP ###############################
441 print "nmap -oG - -P0 -p22,23,80,443,5869,9100,16992 %s | grep Host:" % reboot.pcu_name(values['plc_pcu_stats'])
442 (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']))
443 # NOTE: an empty / error value for oval, will still work.
444 (values['port_status'], continue_probe) = nmap_port_status(oval)
446 values['port_status'] = None
448 #### COMPLETE ENTRY #######################
450 values['entry_complete'] = []
451 #if values['protocol'] is None or values['protocol'] is "":
452 # values['entry_complete'] += ["protocol"]
453 if values['plc_pcu_stats']['model'] is None or values['plc_pcu_stats']['model'] is "":
454 values['entry_complete'] += ["model"]
455 # Cannot continue due to this condition
456 continue_probe = False
458 if values['plc_pcu_stats']['password'] is None or values['plc_pcu_stats']['password'] is "":
459 values['entry_complete'] += ["password"]
460 # Cannot continue due to this condition
461 continue_probe = False
463 if len(values['entry_complete']) > 0:
464 continue_probe = False
466 if values['plc_pcu_stats']['hostname'] is None or values['plc_pcu_stats']['hostname'] is "":
467 values['entry_complete'] += ["hostname"]
468 if values['plc_pcu_stats']['ip'] is None or values['plc_pcu_stats']['ip'] is "":
469 values['entry_complete'] += ["ip"]
471 # If there are no nodes associated with this PCU, then we cannot continue.
472 if len(values['plc_pcu_stats']['node_ids']) == 0:
473 continue_probe = False
474 values['entry_complete'] += ['nodeids']
477 #### DNS and IP MATCH #######################
478 if values['plc_pcu_stats']['hostname'] is not None and values['plc_pcu_stats']['hostname'] is not "" and \
479 values['plc_pcu_stats']['ip'] is not None and values['plc_pcu_stats']['ip'] is not "":
481 ipaddr = socket.gethostbyname(values['plc_pcu_stats']['hostname'])
482 if ipaddr == values['plc_pcu_stats']['ip']:
483 values['dns_status'] = "DNS-OK"
485 values['dns_status'] = "DNS-MISMATCH"
486 continue_probe = False
488 except Exception, err:
489 values['dns_status'] = "DNS-NOENTRY"
490 values['plc_pcu_stats']['hostname'] = values['plc_pcu_stats']['ip']
492 if values['plc_pcu_stats']['ip'] is not None and values['plc_pcu_stats']['ip'] is not "":
493 values['dns_status'] = "NOHOSTNAME"
494 values['plc_pcu_stats']['hostname'] = values['plc_pcu_stats']['ip']
496 values['dns_status'] = "NO-DNS-OR-IP"
497 values['plc_pcu_stats']['hostname'] = "No_entry_in_DB"
498 continue_probe = False
501 ###### DRY RUN ############################
502 if continue_probe and 'node_ids' in values['plc_pcu_stats'] and \
503 len(values['plc_pcu_stats']['node_ids']) > 0:
504 rb_ret = reboot.reboot_test_new(values['plc_pcu_stats']['nodenames'][0],
507 rb_ret = "Not_Run" # No nodes to test"
509 values['reboot_trial_status'] = rb_ret
512 print "____________________________________"
515 print "____________________________________"
516 errors['traceback'] = traceback.format_exc()
517 print errors['traceback']
518 values['reboot_trial_status'] = str(errors['traceback'])
521 values['entry_complete']=" ".join(values['entry_complete'])
523 values['date_checked'] = datetime.now()
524 return (pcuname, values)