3 # Reboot specified nodes
9 import errno, time, traceback
11 import threading, popen2
17 from subprocess import PIPE, Popen
18 import ssh.pxssh as pxssh
19 import ssh.pexpect as pexpect
22 # Use our versions of telnetlib and pyssh
23 sys.path.insert(0, os.path.dirname(sys.argv[0]))
25 sys.path.insert(0, os.path.dirname(sys.argv[0]) + "/pyssh")
31 # Event class ID from pcu events
32 #NODE_POWER_CONTROL = 3
35 #MONITOR_USER_ID = 11142
38 logger = logging.getLogger("monitor")
42 class ExceptionNoTransport(Exception): pass
43 class ExceptionNotFound(Exception): pass
44 class ExceptionPassword(Exception): pass
45 class ExceptionTimeout(Exception): pass
46 class ExceptionPrompt(Exception): pass
47 class ExceptionSequence(Exception): pass
48 class ExceptionReset(Exception): pass
49 class ExceptionPort(Exception): pass
50 class ExceptionUsername(Exception): pass
52 def telnet_answer(telnet, expected, buffer):
55 output = telnet.read_until(expected, TELNET_TIMEOUT)
57 # logger.debug(output)
58 if output.find(expected) == -1:
59 raise ExceptionNotFound, "'%s' not found" % expected
61 telnet.write(buffer + "\r\n")
64 # PCU has model, host, preferred-port, user, passwd,
66 # This is an object derived directly form the PLCAPI DB fields
68 def __init__(self, plc_pcu_dict):
69 for field in ['username', 'password', 'site_id',
72 'node_ids', 'ports', ]:
73 if field in plc_pcu_dict:
74 self.__setattr__(field, plc_pcu_dict[field])
76 raise Exception("No such field %s in PCU object" % field)
78 # These are the convenience functions build around the PCU object.
80 def __init__(self, plc_pcu_dict):
81 PCU.__init__(self, plc_pcu_dict)
82 self.host = self.pcu_name()
85 if self.hostname is not None and self.hostname is not "":
87 elif self.ip is not None and self.ip is not "":
92 def nodeidToPort(self, node_id):
93 if node_id in self.node_ids:
94 for i in range(0, len(self.node_ids)):
95 if node_id == self.node_ids[i]:
98 raise Exception("No such Node ID: %d" % node_id)
100 # This class captures the observed pcu records from FindBadPCUs.py
102 def __init__(self, pcu_record_dict):
103 for field in ['nodenames', 'portstatus',
106 if field in pcu_record_dict:
107 if field == "reboot":
108 self.__setattr__("reboot_str", pcu_record_dict[field])
110 self.__setattr__(field, pcu_record_dict[field])
112 raise Exception("No such field %s in pcu record dict" % field)
122 def __init__(self, type, verbose):
124 self.verbose = verbose
125 self.transport = None
127 def open(self, host, username=None, password=None, prompt="User Name"):
130 if self.type == self.TELNET:
131 transport = telnetlib.Telnet(host, timeout=self.TELNET_TIMEOUT)
132 transport.set_debuglevel(self.verbose)
133 if username is not None:
134 self.transport = transport
135 self.ifThenSend(prompt, username, ExceptionUsername)
137 elif self.type == self.SSH:
138 if username is not None:
139 transport = pyssh.Ssh(username, host)
140 transport.set_debuglevel(self.verbose)
142 # TODO: have an ssh set_debuglevel() also...
144 raise Exception("Username cannot be None for ssh transport.")
145 elif self.type == self.HTTP:
146 self.url = "http://%s:%d/" % (host,80)
147 uri = "%s:%d" % (host,80)
150 authinfo = urllib2.HTTPPasswordMgrWithDefaultRealm()
151 authinfo.add_password (None, uri, username, password)
152 authhandler = urllib2.HTTPBasicAuthHandler( authinfo )
154 transport = urllib2.build_opener(authhandler)
157 raise Exception("Unknown transport type: %s" % self.type)
159 self.transport = transport
163 if self.type == self.TELNET:
164 self.transport.close()
165 elif self.type == self.SSH:
166 self.transport.close()
167 elif self.type == self.HTTP:
170 raise Exception("Unknown transport type %s" % self.type)
171 self.transport = None
173 def sendHTTP(self, resource, data):
175 print "POSTing '%s' to %s" % (data,self.url + resource)
178 f = self.transport.open(self.url + resource ,data)
183 except urllib2.URLError,err:
184 logger.info('Could not open http connection', err)
185 return "http transport error"
189 def sendPassword(self, password, prompt=None):
190 if self.type == self.TELNET:
192 self.ifThenSend("Password", password, ExceptionPassword)
194 self.ifThenSend(prompt, password, ExceptionPassword)
195 elif self.type == self.SSH:
196 self.ifThenSend("password:", password, ExceptionPassword)
197 elif self.type == self.HTTP:
200 raise Exception("Unknown transport type: %s" % self.type)
202 def ifThenSend(self, expected, buffer, ErrorClass=ExceptionPrompt):
204 if self.transport != None:
205 output = self.transport.read_until(expected, self.TELNET_TIMEOUT)
206 if output.find(expected) == -1:
207 raise ErrorClass, "'%s' not found" % expected
209 self.transport.write(buffer + "\r\n")
211 raise ExceptionNoTransport("transport object is type None")
213 def ifElse(self, expected, ErrorClass):
215 self.transport.read_until(expected, self.TELNET_TIMEOUT)
217 raise ErrorClass("Could not find '%s' within timeout" % expected)
220 class PCUControl(Transport,PCUModel,PCURecord):
221 def __init__(self, plc_pcu_record, verbose, supported_ports=[]):
222 PCUModel.__init__(self, plc_pcu_record)
223 PCURecord.__init__(self, plc_pcu_record)
226 if '22' in supported_ports and self.portstatus['22'] == "open":
228 elif '23' in supported_ports and self.portstatus['23'] == "open":
229 type = Transport.TELNET
230 elif '80' in supported_ports and self.portstatus['80'] == "open":
231 type = Transport.HTTP
232 elif '443' in supported_ports and self.portstatus['443'] == "open":
233 type = Transport.HTTP
234 elif '5869' in supported_ports and self.portstatus['5869'] == "open":
235 # For DRAC cards. Racadm opens this port.
236 type = Transport.HTTP
237 elif '9100' in supported_ports and self.portstatus['9100'] == "open":
238 type = Transport.IPAL
239 elif '16992' in supported_ports and self.portstatus['16992'] == "open":
240 type = Transport.HTTP
242 raise ExceptionPort("Unsupported Port: No transport from open ports")
244 raise Exception("No Portstatus: No transport because no open ports")
245 Transport.__init__(self, type, verbose)
247 def run(self, node_port, dryrun):
248 """ This function is to be defined by the specific PCU instance. """
251 def reboot(self, node_port, dryrun):
253 return self.run(node_port, dryrun)
254 except ExceptionNotFound, err:
255 return "error: " + str(err)
256 except ExceptionPassword, err:
257 return "password exception: " + str(err)
258 except ExceptionTimeout, err:
259 return "timeout exception: " + str(err)
260 except ExceptionUsername, err:
261 return "exception: no username prompt: " + str(err)
262 except ExceptionSequence, err:
263 return "sequence error: " + str(err)
264 except ExceptionPrompt, err:
265 return "prompt exception: " + str(err)
266 except ExceptionPort, err:
267 return "no ports exception: " + str(err)
268 except socket.error, err:
269 return "socket error: timeout: " + str(err)
270 except EOFError, err:
272 logger.debug("reboot: EOF")
274 self.transport.close()
276 traceback.print_exc()
277 return "EOF connection reset" + str(err)
279 class IPAL(PCUControl):
281 This now uses a proprietary format for communicating with the PCU. I
282 prefer it to Telnet, and Web access, since it's much lighter weight
283 and, more importantly, IT WORKS!! HHAHHHAHAHAHAHAHA!
286 def format_msg(self, data, cmd):
287 esc = chr(int('1b',16))
288 return "%c%s%c%s%c" % (esc, self.password, esc, data, cmd) # esc, 'q', chr(4))
290 def recv_noblock(self, s, count):
294 # TODO: make sleep backoff, before stopping.
296 ret = s.recv(count, socket.MSG_DONTWAIT)
297 except socket.error, e:
298 if e[0] == errno.EAGAIN:
299 return Exception(e[1])
301 # TODO: not other exceptions.
305 def run(self, node_port, dryrun):
311 s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
314 s.connect((self.host, 9100))
315 except socket.error, e:
317 if e[0] == errno.ECONNREFUSED:
318 # cannot connect to remote host
319 return Exception(e[1])
321 # TODO: what other conditions are there?
325 print "Checking status"
326 s.send(self.format_msg("", 'O'))
327 ret = self.recv_noblock(s, 8)
328 print "Current status is '%s'" % ret
330 if node_port < len(ret):
331 status = ret[node_port]
339 raise Exception("Unknown status for PCU socket %s : %s" % (node_port, ret))
341 raise Exception("Mismatch between configured port and PCU status: %s %s" % (node_port, ret))
345 print "Pulsing %s" % node_port
347 s.send(self.format_msg("%s" % node_port, 'P'))
349 # NOTE: turn power on before pulsing the port.
350 print "power was off, so turning on then pulsing..."
351 s.send(self.format_msg("%s" % node_port, 'E'))
352 s.send(self.format_msg("%s" % node_port, 'P'))
354 print "Receiving response."
355 ret = self.recv_noblock(s, 8)
356 print "Current status is '%s'" % ret
358 if node_port < len(ret):
359 status = ret[node_port]
367 raise Exception("Unknown status for PCU socket %s : %s" % (node_port, ret))
369 raise Exception("Mismatch between configured port and PCU status: %s %s" % (node_port, ret))
374 return "Failed Power On"
379 # TELNET version of protocol...
380 # #self.open(self.host)
381 # ## XXX Some iPals require you to hit Enter a few times first
382 # #self.ifThenSend("Password >", "\r\n\r\n", ExceptionNotFound)
384 # self.ifThenSend("Password >", self.password, ExceptionPassword)
385 # self.transport.write("\r\n\r\n")
386 # if not dryrun: # P# - Pulse relay
387 # print "node_port %s" % node_port
388 # self.ifThenSend("Enter >",
389 # "P7", # % node_port,
391 # print "send newlines"
392 # self.transport.write("\r\n\r\n")
393 # print "after new lines"
394 # # Get the next prompt
395 # print "wait for enter"
396 # self.ifElse("Enter >", ExceptionTimeout)
401 class APCEurope(PCUControl):
402 def run(self, node_port, dryrun):
403 self.open(self.host, self.username)
404 self.sendPassword(self.password)
406 self.ifThenSend("\r\n> ", "1", ExceptionPassword)
407 self.ifThenSend("\r\n> ", "2")
408 self.ifThenSend("\r\n> ", str(node_port))
409 # 3- Immediate Reboot
410 self.ifThenSend("\r\n> ", "3")
413 self.ifThenSend("Enter 'YES' to continue or <ENTER> to cancel",
417 self.ifThenSend("Enter 'YES' to continue or <ENTER> to cancel",
418 "", ExceptionSequence)
419 self.ifThenSend("Press <ENTER> to continue...", "", ExceptionSequence)
424 class APCBrazil(PCUControl):
425 def run(self, node_port, dryrun):
426 self.open(self.host, self.username)
427 self.sendPassword(self.password)
429 self.ifThenSend("\r\n> ", "1", ExceptionPassword)
430 self.ifThenSend("\r\n> ", str(node_port))
431 # 4- Immediate Reboot
432 self.ifThenSend("\r\n> ", "4")
435 self.ifThenSend("Enter 'YES' to continue or <ENTER> to cancel",
439 self.ifThenSend("Enter 'YES' to continue or <ENTER> to cancel",
440 "", ExceptionSequence)
441 self.ifThenSend("Press <ENTER> to continue...", "", ExceptionSequence)
446 class APCBerlin(PCUControl):
447 def run(self, node_port, dryrun):
448 self.open(self.host, self.username)
449 self.sendPassword(self.password)
451 self.ifThenSend("\r\n> ", "1", ExceptionPassword)
452 self.ifThenSend("\r\n> ", "2")
453 self.ifThenSend("\r\n> ", "1")
454 self.ifThenSend("\r\n> ", str(node_port))
455 # 3- Immediate Reboot
456 self.ifThenSend("\r\n> ", "3")
459 self.ifThenSend("Enter 'YES' to continue or <ENTER> to cancel",
463 self.ifThenSend("Enter 'YES' to continue or <ENTER> to cancel",
464 "", ExceptionSequence)
465 self.ifThenSend("Press <ENTER> to continue...", "", ExceptionSequence)
470 class APCFolsom(PCUControl):
471 def run(self, node_port, dryrun):
472 self.open(self.host, self.username)
473 self.sendPassword(self.password)
475 self.ifThenSend("\r\n> ", "1", ExceptionPassword)
476 self.ifThenSend("\r\n> ", "2")
477 self.ifThenSend("\r\n> ", "1")
478 self.ifThenSend("\r\n> ", str(node_port))
479 self.ifThenSend("\r\n> ", "1")
481 # 3- Immediate Reboot
482 self.ifThenSend("\r\n> ", "3")
485 self.ifThenSend("Enter 'YES' to continue or <ENTER> to cancel",
489 self.ifThenSend("Enter 'YES' to continue or <ENTER> to cancel",
490 "", ExceptionSequence)
491 self.ifThenSend("Press <ENTER> to continue...", "", ExceptionSequence)
496 class APCMaster(PCUControl):
497 def run(self, node_port, dryrun):
498 print "Rebooting %s" % self.host
499 self.open(self.host, self.username)
500 self.sendPassword(self.password)
503 self.ifThenSend("\r\n> ", "1", ExceptionPassword)
504 # 3- Outlet Control/Config
505 self.ifThenSend("\r\n> ", "3")
507 self.ifThenSend("\r\n> ", str(node_port))
509 self.ifThenSend("\r\n> ", "1")
510 # 3- Immediate Reboot
511 self.ifThenSend("\r\n> ", "3")
514 self.ifThenSend("Enter 'YES' to continue or <ENTER> to cancel",
518 self.ifThenSend("Enter 'YES' to continue or <ENTER> to cancel",
519 "", ExceptionSequence)
520 self.ifThenSend("Press <ENTER> to continue...", "", ExceptionSequence)
525 class APC(PCUControl):
526 def __init__(self, plc_pcu_record, verbose):
527 PCUControl.__init__(self, plc_pcu_record, verbose)
529 self.master = APCMaster(plc_pcu_record, verbose)
530 self.folsom = APCFolsom(plc_pcu_record, verbose)
531 self.europe = APCEurope(plc_pcu_record, verbose)
533 def run(self, node_port, dryrun):
537 for pcu in [self.master, self.europe, self.folsom]:
540 print "-*_*_*_*_*_*_*_*_*_*_*_*_*_*_*_*_*_*_*_*_*"
543 time.sleep(sleep_time)
544 ret = pcu.reboot(node_port, dryrun)
545 except ExceptionSequence, err:
551 return "Unknown reboot sequence for APC PCU"
555 class IntelAMT(PCUControl):
556 def run(self, node_port, dryrun):
560 cmd_str = "IntelAMTSDK/Samples/RemoteControl/remoteControl"
563 # NOTE: -p checks the power state of the host.
564 # TODO: parse the output to find out if it's ok or not.
565 cmd_str += " -p http://%s:16992/RemoteControlService -user admin -pass '%s' " % (self.host, self.password )
567 cmd_str += " -A http://%s:16992/RemoteControlService -user admin -pass '%s' " % (self.host, self.password )
570 return cmd.system(cmd_str, self.TELNET_TIMEOUT)
572 class DRACRacAdm(PCUControl):
573 def run(self, node_port, dryrun):
575 print "trying racadm_reboot..."
576 racadm_reboot(self.host, self.username, self.password, node_port, dryrun)
580 class DRAC(PCUControl):
581 def run(self, node_port, dryrun):
582 self.open(self.host, self.username)
583 self.sendPassword(self.password)
585 print "logging in..."
586 self.transport.write("\r\n")
589 self.ifThenSend("[%s]#" % self.username, "getsysinfo")
592 self.ifThenSend("[%s]#" % self.username, "serveraction powercycle")
594 self.ifThenSend("[%s]#" % self.username, "exit")
599 class HPiLO(PCUControl):
600 def run(self, node_port, dryrun):
601 self.open(self.host, self.username)
602 self.sendPassword(self.password)
605 self.ifThenSend("</>hpiLO->", "cd system1")
607 # Reboot Outlet N (Y/N)?
609 self.ifThenSend("</system1>hpiLO->", "POWER")
612 self.ifThenSend("</system1>hpiLO->", "reset")
614 self.ifThenSend("</system1>hpiLO->", "exit")
620 class HPiLOHttps(PCUControl):
621 def run(self, node_port, dryrun):
624 locfg = soltesz.CMD()
625 cmd = "cmdhttps/locfg.pl -s %s -f %s -u %s -p '%s' | grep 'MESSAGE' | grep -v 'No error'" % (
626 self.host, "iloxml/Get_Network.xml",
627 self.username, self.password)
628 sout, serr = locfg.run_noexcept(cmd)
630 if sout.strip() != "":
631 print "sout: %s" % sout.strip()
635 locfg = soltesz.CMD()
636 cmd = "cmdhttps/locfg.pl -s %s -f %s -u %s -p '%s' | grep 'MESSAGE' | grep -v 'No error'" % (
637 self.host, "iloxml/Reset_Server.xml",
638 self.username, self.password)
639 sout, serr = locfg.run_noexcept(cmd)
641 if sout.strip() != "":
642 print "sout: %s" % sout.strip()
646 class BayTechAU(PCUControl):
647 def run(self, node_port, dryrun):
648 self.open(self.host, self.username, None, "Enter user name:")
649 self.sendPassword(self.password, "Enter Password:")
651 #self.ifThenSend("RPC-16>", "Status")
652 self.ifThenSend("RPC3-NC>", "Reboot %d" % node_port)
654 # Reboot Outlet N (Y/N)?
656 self.ifThenSend("(Y/N)?", "N")
658 self.ifThenSend("(Y/N)?", "Y")
659 self.ifThenSend("RPC3-NC>", "")
664 class BayTechGeorgeTown(PCUControl):
665 def run(self, node_port, dryrun):
666 self.open(self.host, self.username, None, "Enter user name:")
667 self.sendPassword(self.password, "Enter Password:")
669 #self.ifThenSend("RPC-16>", "Status")
671 self.ifThenSend("RPC-16>", "Reboot %d" % node_port)
673 # Reboot Outlet N (Y/N)?
675 self.ifThenSend("(Y/N)?", "N")
677 self.ifThenSend("(Y/N)?", "Y")
678 self.ifThenSend("RPC-16>", "")
683 class BayTechCtrlCUnibe(PCUControl):
685 For some reason, these units let you log in fine, but they hang
686 indefinitely, unless you send a Ctrl-C after the password. No idea
689 def run(self, node_port, dryrun):
690 print "BayTechCtrlC %s" % self.host
692 ssh_options="-o StrictHostKeyChecking=no -o PasswordAuthentication=yes -o PubkeyAuthentication=no"
694 if not s.login(self.host, self.username, self.password, ssh_options):
695 raise ExceptionPassword("Invalid Password")
696 # Otherwise, the login succeeded.
698 # Send a ctrl-c to the remote process.
699 print "sending ctrl-c"
702 # Control Outlets (5 ,1).........5
704 index = s.expect(["Enter Request :"])
709 index = s.expect(["DS-RPC>", "Enter user name:"])
711 s.send(self.username + "\r\n")
712 index = s.expect(["DS-RPC>"])
715 print "Reboot %d" % node_port
716 s.send("Reboot %d\r\n" % node_port)
718 index = s.expect(["(Y/N)?"])
727 #index = s.expect(["DS-RPC>"])
728 #print "got prompt back"
733 raise ExceptionPrompt("EOF before 'Enter Request' Prompt")
734 except pexpect.TIMEOUT:
735 raise ExceptionPrompt("Timeout before 'Enter Request' Prompt")
739 class BayTechCtrlC(PCUControl):
741 For some reason, these units let you log in fine, but they hang
742 indefinitely, unless you send a Ctrl-C after the password. No idea
745 def run(self, node_port, dryrun):
746 print "BayTechCtrlC %s" % self.host
748 ssh_options="-o StrictHostKeyChecking=no -o PasswordAuthentication=yes -o PubkeyAuthentication=no"
750 if not s.login(self.host, self.username, self.password, ssh_options):
751 raise ExceptionPassword("Invalid Password")
752 # Otherwise, the login succeeded.
754 # Send a ctrl-c to the remote process.
755 print "sending ctrl-c"
758 # Control Outlets (5 ,1).........5
760 index = s.expect(["Enter Request :"])
765 index = s.expect(["DS-RPC>", "Enter user name:"])
767 print "sending username"
768 s.send(self.username + "\r\n")
769 index = s.expect(["DS-RPC>"])
772 print "Reboot %d" % node_port
773 s.send("Reboot %d\r\n" % node_port)
775 index = s.expect(["(Y/N)?"])
784 index = s.expect(["DS-RPC>"])
785 #print "got prompt back"
790 raise ExceptionPrompt("EOF before 'Enter Request' Prompt")
791 except pexpect.TIMEOUT:
792 raise ExceptionPrompt("Timeout before Prompt")
796 class BayTech(PCUControl):
797 def run(self, node_port, dryrun):
798 self.open(self.host, self.username)
799 self.sendPassword(self.password)
801 # Control Outlets (5 ,1).........5
802 self.ifThenSend("Enter Request :", "5")
806 self.ifThenSend("DS-RPC>", "Reboot %d" % node_port, ExceptionNotFound)
807 except ExceptionNotFound, msg:
808 # one machine is configured to ask for a username,
809 # even after login...
810 print "msg: %s" % msg
811 self.transport.write(self.username + "\r\n")
812 self.ifThenSend("DS-RPC>", "Reboot %d" % node_port)
814 # Reboot Outlet N (Y/N)?
816 self.ifThenSend("(Y/N)?", "N")
818 self.ifThenSend("(Y/N)?", "Y")
819 self.ifThenSend("DS-RPC>", "")
824 class WTIIPS4(PCUControl):
825 def run(self, node_port, dryrun):
827 self.sendPassword(self.password, "Enter Password:")
829 self.ifThenSend("IPS> ", "/Boot %s" % node_port)
831 self.ifThenSend("Sure? (Y/N): ", "N")
833 self.ifThenSend("Sure? (Y/N): ", "Y")
835 self.ifThenSend("IPS> ", "")
840 class ePowerSwitchGood(PCUControl):
842 # The old code used Python's HTTPPasswordMgrWithDefaultRealm()
843 # For some reason this both doesn't work and in some cases, actually
844 # hangs the PCU. Definitely not what we want.
846 # The code below is much simpler. Just letting things fail first,
847 # and then, trying again with authentication string in the header.
849 def run(self, node_port, dryrun):
850 self.transport = None
851 self.url = "http://%s:%d/" % (self.host,80)
852 uri = "%s:%d" % (self.host,80)
854 req = urllib2.Request(self.url)
856 handle = urllib2.urlopen(req)
858 # NOTE: this is expected to fail initially
865 return "ERROR: not protected by HTTP authentication"
867 if not hasattr(e, 'code') or e.code != 401:
868 return "ERROR: failed for: %s" % str(e)
870 base64data = base64.encodestring("%s:%s" % (self.username, self.password))[:-1]
871 # NOTE: assuming basic realm authentication.
872 authheader = "Basic %s" % base64data
873 req.add_header("Authorization", authheader)
876 f = urllib2.urlopen(req)
878 # failing here means the User/passwd is wrong (hopefully)
879 raise ExceptionPassword("Incorrect username/password")
881 # TODO: after verifying that the user/password is correct, we should
882 # actually reboot the given node.
885 # add data to handler,
886 # fetch url one more time on cmd.html, econtrol.html or whatever.
889 if self.verbose: print f.read()
895 class ePowerSwitchOld(PCUControl):
896 def run(self, node_port, dryrun):
897 self.url = "http://%s:%d/" % (self.host,80)
898 uri = "%s:%d" % (self.host,80)
901 authinfo = urllib2.HTTPPasswordMgrWithDefaultRealm()
902 authinfo.add_password (None, uri, self.username, self.password)
903 authhandler = urllib2.HTTPBasicAuthHandler( authinfo )
905 # NOTE: it doesn't seem to matter whether this authinfo is here or not.
906 transport = urllib2.build_opener(authinfo)
907 f = transport.open(self.url)
908 if self.verbose: print f.read()
911 transport = urllib2.build_opener(authhandler)
912 f = transport.open(self.url + "cmd.html", "P%d=r" % node_port)
913 if self.verbose: print f.read()
918 class ePowerSwitch(PCUControl):
919 def run(self, node_port, dryrun):
920 self.url = "http://%s:%d/" % (self.host,80)
921 uri = "%s:%d" % (self.host,80)
923 # TODO: I'm still not sure what the deal is here.
924 # two independent calls appear to need to be made before the
925 # reboot will succeed. It doesn't seem to be possible to do
926 # this with a single call. I have no idea why.
929 authinfo = urllib2.HTTPPasswordMgrWithDefaultRealm()
930 authinfo.add_password (None, uri, self.username, self.password)
931 authhandler = urllib2.HTTPBasicAuthHandler( authinfo )
933 # NOTE: it doesn't seem to matter whether this authinfo is here or not.
934 transport = urllib2.build_opener()
935 f = transport.open(self.url + "elogin.html", "pwd=%s" % self.password)
936 if self.verbose: print f.read()
939 transport = urllib2.build_opener(authhandler)
940 f = transport.open(self.url + "econtrol.html", "P%d=r" % node_port)
941 if self.verbose: print f.read()
943 # data= "P%d=r" % node_port
944 #self.open(self.host, self.username, self.password)
945 #self.sendHTTP("elogin.html", "pwd=%s" % self.password)
946 #self.sendHTTP("econtrol.html", data)
947 #self.sendHTTP("cmd.html", data)
953 ### rebooting european BlackBox PSE boxes
954 # Thierry Parmentelat - May 11 2005
955 # tested on 4-ports models known as PSE505-FR
956 # uses http to POST a data 'P<port>=r'
957 # relies on basic authentication within http1.0
958 # first curl-based script was
959 # curl --http1.0 --basic --user <username>:<password> --data P<port>=r \
960 # http://<hostname>:<http_port>/cmd.html && echo OK
962 def bbpse_reboot (pcu_ip,username,password,port_in_pcu,http_port, dryrun):
966 url = "http://%s:%d/cmd.html" % (pcu_ip,http_port)
967 data= "P%d=r" % port_in_pcu
969 logger.debug("POSTing '%s' on %s" % (data,url))
971 authinfo = urllib2.HTTPPasswordMgrWithDefaultRealm()
972 uri = "%s:%d" % (pcu_ip,http_port)
973 authinfo.add_password (None, uri, username, password)
974 authhandler = urllib2.HTTPBasicAuthHandler( authinfo )
976 opener = urllib2.build_opener(authhandler)
977 urllib2.install_opener(opener)
983 f = urllib2.urlopen(url,data)
990 except urllib2.URLError,err:
991 logger.info('Could not open http connection', err)
994 ### rebooting x10toggle based systems addressed by port
995 # Marc E. Fiuczynski - May 31 2005
996 # tested on 4-ports models known as PSE505-FR
997 # uses ssh and password to login to an account
998 # that will cause the system to be powercycled.
1000 def x10toggle_reboot(ip, username, password, port, dryrun):
1005 ssh = pyssh.Ssh(username, ip)
1009 telnet_answer(ssh, "password:", password)
1013 telnet_answer(ssh, "x10toggle>", "A%d" % port)
1016 output = ssh.close()
1018 logger.debug(output)
1021 except Exception, err:
1025 output = ssh.close()
1027 logger.debug(output)
1028 return errno.ETIMEDOUT
1030 ### rebooting Dell systems via RAC card
1031 # Marc E. Fiuczynski - June 01 2005
1032 # tested with David Lowenthal's itchy/scratchy nodes at UGA
1035 def runcmd(command, args, username, password, timeout = None):
1038 result_ready = threading.Condition()
1042 result_ready.acquire()
1046 result_ready.notify()
1047 result_ready.release()
1049 def do_command(command, username, password):
1052 # Popen4 is a popen-type class that combines stdout and stderr
1053 p = popen2.Popen4(command)
1055 # read all output data
1056 p.tochild.write("%s\n" % username)
1057 p.tochild.write("%s\n" % password)
1059 data = p.fromchild.read()
1062 # might get interrupted by a signal in poll() or waitpid()
1065 set_result((retval, data))
1068 if ex.errno == errno.EINTR:
1071 except Exception, ex:
1075 command = " ".join([command] + args)
1077 worker = threading.Thread(target = do_command, args = (command, username, password, ))
1078 worker.setDaemon(True)
1079 result_ready.acquire()
1081 result_ready.wait(timeout)
1083 if result == [None]:
1084 raise Exception, "command timed-out: '%s'" % command
1086 result_ready.release()
1089 if isinstance(result, Exception):
1092 (retval, data) = result
1093 if os.WIFEXITED(retval) and os.WEXITSTATUS(retval) == 0:
1096 out = "system command ('%s') " % command
1097 if os.WIFEXITED(retval):
1098 out += "failed, rc = %d" % os.WEXITSTATUS(retval)
1100 out += "killed by signal %d" % os.WTERMSIG(retval)
1102 out += "; output follows:\n" + data
1103 raise Exception, out
1105 def racadm_reboot(host, username, password, port, dryrun):
1108 ip = socket.gethostbyname(host)
1110 cmd = "/usr/sbin/racadm"
1113 output = runcmd(cmd, ["-r %s -i serveraction powercycle" % ip],
1116 output = runcmd(cmd, ["-r %s -i getsysinfo" % ip],
1119 print "RUNCMD: %s" % output
1121 logger.debug(output)
1124 except Exception, err:
1125 logger.debug("runcmd raised exception %s" % err)
1131 if pcu['hostname'] is not None and pcu['hostname'] is not "":
1132 return pcu['hostname']
1133 elif pcu['ip'] is not None and pcu['ip'] is not "":
1139 fb =soltesz.dbLoad("findbadpcus")
1141 def get_pcu_values(pcu_id):
1142 # TODO: obviously, this shouldn't be loaded each time...
1145 values = fb['nodes']["id_%s" % pcu_id]['values']
1151 def reboot(nodename):
1152 return reboot_policy(nodename, True, False)
1154 def reboot_policy(nodename, continue_probe, dryrun):
1157 pcu = plc.getpcu(nodename)
1159 logger.debug("no pcu for %s" % hostname)
1160 print "no pcu for %s" % hostname
1161 return False # "%s has no pcu" % nodename
1163 values = get_pcu_values(pcu['pcu_id'])
1165 logger.debug("No values for pcu probe %s" % hostname)
1166 print "No values for pcu probe %s" % hostname
1167 return False #"no info for pcu_id %s" % pcu['pcu_id']
1170 logger.debug("Trying PCU %s %s" % (pcu['hostname'], pcu['model']))
1172 ret = reboot_test(nodename, values, continue_probe, verbose, dryrun)
1181 def reboot_test(nodename, values, continue_probe, verbose, dryrun):
1185 # DataProbe iPal (many sites)
1186 if continue_probe and values['model'].find("Dataprobe IP-41x/IP-81x") >= 0:
1187 ipal = IPAL(values, verbose, ['23', '80', '9100'])
1188 rb_ret = ipal.reboot(values[nodename], dryrun)
1190 # APC Masterswitch (Berkeley)
1191 elif continue_probe and values['model'].find("APC AP79xx/Masterswitch") >= 0:
1194 # TODO: make a more robust version of APC
1195 if values['pcu_id'] in [1163,1055,1111,1231,1113,1127,1128,1148]:
1196 apc = APCEurope(values, verbose, ['22', '23'])
1197 rb_ret = apc.reboot(values[nodename], dryrun)
1199 elif values['pcu_id'] in [1110,86]:
1200 apc = APCBrazil(values, verbose, ['22', '23'])
1201 rb_ret = apc.reboot(values[nodename], dryrun)
1203 elif values['pcu_id'] in [1221,1225]:
1204 apc = APCBerlin(values, verbose, ['22', '23'])
1205 rb_ret = apc.reboot(values[nodename], dryrun)
1207 elif values['pcu_id'] in [1173,1221,1220]:
1208 apc = APCFolsom(values, verbose, ['22', '23'])
1209 rb_ret = apc.reboot(values[nodename], dryrun)
1212 apc = APCMaster(values, verbose, ['22', '23'])
1213 rb_ret = apc.reboot(values[nodename], dryrun)
1216 elif continue_probe and values['model'].find("Baytech DS4-RPC") >= 0:
1217 if values['pcu_id'] in [1237,1052,1209,1002,1008,1041,1013,1022]:
1218 # These require a 'ctrl-c' to be sent...
1219 baytech = BayTechCtrlC(values, verbose, ['22', '23'])
1220 rb_ret = baytech.reboot(values[nodename], dryrun)
1222 elif values['pcu_id'] in [93]:
1223 baytech = BayTechAU(values, verbose, ['22', '23'])
1224 rb_ret = baytech.reboot(values[nodename], dryrun)
1226 elif values['pcu_id'] in [1057]:
1227 # These require a 'ctrl-c' to be sent...
1228 baytech = BayTechCtrlCUnibe(values, verbose, ['22', '23'])
1229 rb_ret = baytech.reboot(values[nodename], dryrun)
1231 elif values['pcu_id'] in [1012]:
1232 # This pcu sometimes doesn't present the 'Username' prompt,
1233 # unless you immediately try again...
1235 baytech = BayTechGeorgeTown(values, verbose, ['22', '23'])
1236 rb_ret = baytech.reboot(values[nodename], dryrun)
1238 baytech = BayTechGeorgeTown(values, verbose, ['22', '23'])
1239 rb_ret = baytech.reboot(values[nodename], dryrun)
1241 baytech = BayTech(values, verbose, ['22', '23'])
1242 rb_ret = baytech.reboot(values[nodename], dryrun)
1245 elif continue_probe and values['model'].find("HP iLO") >= 0:
1247 hpilo = HPiLO(values, verbose, ['22'])
1248 rb_ret = hpilo.reboot(0, dryrun)
1250 hpilo = HPiLOHttps(values, verbose, ['443'])
1251 rb_ret = hpilo.reboot(0, dryrun)
1253 hpilo = HPiLOHttps(values, verbose, ['443'])
1254 rb_ret = hpilo.reboot(0, dryrun)
1257 elif continue_probe and values['model'].find("Dell RAC") >= 0:
1258 # TODO: I don't think DRACRacAdm will throw an exception for the
1259 # default method to catch...
1261 drac = DRACRacAdm(values, verbose, ['443', '5869'])
1262 rb_ret = drac.reboot(0, dryrun)
1264 drac = DRAC(values, verbose, ['22'])
1265 rb_ret = drac.reboot(0, dryrun)
1267 elif continue_probe and values['model'].find("WTI IPS-4") >= 0:
1268 wti = WTIIPS4(values, verbose, ['23'])
1269 rb_ret = wti.reboot(values[nodename], dryrun)
1271 elif continue_probe and values['model'].find("Intel AMT") >= 0:
1272 amt = IntelAMT(values, verbose, ['16992'])
1273 rb_ret = amt.reboot(values[nodename], dryrun)
1275 # BlackBox PSExxx-xx (e.g. PSE505-FR)
1276 elif continue_probe and \
1277 (values['model'].find("BlackBox PS5xx") >= 0 or
1278 values['model'].find("ePowerSwitch 1/4/8x") >=0 ):
1280 # TODO: allow a different port than http 80.
1281 if values['pcu_id'] in [1089, 1071, 1046, 1035, 1118]:
1282 eps = ePowerSwitchGood(values, verbose, ['80'])
1283 elif values['pcu_id'] in [1003]:
1284 eps = ePowerSwitch(values, verbose, ['80'])
1286 eps = ePowerSwitchGood(values, verbose, ['80'])
1288 rb_ret = eps.reboot(values[nodename], dryrun)
1290 elif continue_probe:
1291 rb_ret = "Unsupported_PCU"
1293 elif continue_probe == False:
1294 if 'portstatus' in values:
1301 except ExceptionPort, err:
1306 #elif continue_probe and values['protocol'] == "racadm" and \
1307 # values['model'] == "RAC":
1308 # rb_ret = racadm_reboot(pcu_name(values),
1309 # values['username'],
1310 # values['password'],
1315 logger.setLevel(logging.DEBUG)
1316 ch = logging.StreamHandler()
1317 ch.setLevel(logging.DEBUG)
1318 formatter = logging.Formatter('LOGGER - %(message)s')
1319 ch.setFormatter(formatter)
1320 logger.addHandler(ch)
1323 if "test" in sys.argv:
1328 for node in sys.argv[1:]:
1329 if node == "test": continue
1331 print "Rebooting %s" % node
1332 if reboot_policy(node, True, dryrun):
1336 except Exception, err:
1337 import traceback; traceback.print_exc()
1340 if __name__ == '__main__':
1342 logger = logging.getLogger("monitor")