3 # Attempt to reboot a node in debug state.
12 from getsshkeys import SSHKnownHosts
20 import ssh.pxssh as pxssh
21 import ssh.fdpexpect as fdpexpect
22 import ssh.pexpect as pexpect
23 from unified_model import *
24 from emailTxt import mailtxt
25 from nodeconfig import network_config_to_str
30 class Sopen(subprocess.Popen):
31 def kill(self, signal = signal.SIGTERM):
32 os.kill(self.pid, signal)
34 #from Rpyc import SocketConnection, Async
35 from Rpyc import SocketConnection, Async
36 from Rpyc.Utils import *
42 fb = database.dbLoad("findbad")
43 fbnode = fb['nodes'][node]['values']
47 def __init__(self, connection, node, config):
52 def get_boot_state(self):
53 if self.c.modules.os.path.exists('/tmp/source'):
55 elif self.c.modules.os.path.exists('/vservers'):
61 self.c.modules.os.system("dmesg > /var/log/dmesg.bm.log")
62 download(self.c, "/var/log/dmesg.bm.log", "log/dmesg.%s.log" % self.node)
63 log = open("log/dmesg.%s.log" % self.node, 'r')
66 def get_bootmanager_log(self):
67 download(self.c, "/tmp/bm.log", "log/bm.%s.log.gz" % self.node)
68 os.system("zcat log/bm.%s.log.gz > log/bm.%s.log" % (self.node, self.node))
69 log = open("log/bm.%s.log" % self.node, 'r')
72 def dump_plconf_file(self):
74 self.c.modules.sys.path.append("/tmp/source/")
75 self.c.modules.os.chdir('/tmp/source')
77 log = c.modules.BootManager.log('/tmp/new.log')
78 bm = c.modules.BootManager.BootManager(log,'boot')
80 BootManagerException = c.modules.Exceptions.BootManagerException
81 InitializeBootManager = c.modules.BootManager.InitializeBootManager
82 ReadNodeConfiguration = c.modules.BootManager.ReadNodeConfiguration
85 InitializeBootManager.Run(bm.VARS, bm.LOG)
86 try: ReadNodeConfiguration.Run(bm.VARS, bm.LOG)
90 print " Possibly, unable to find valid configuration file"
92 if bm_continue and self.config and not self.config.quiet:
93 for key in bm.VARS.keys():
94 print key, " == ", bm.VARS[key]
96 if self.config and not self.config.quiet: print " Unable to read Node Configuration"
99 def compare_and_repair_nodekeys(self):
101 self.c.modules.sys.path.append("/tmp/source/")
102 self.c.modules.os.chdir('/tmp/source')
104 log = c.modules.BootManager.log('/tmp/new.log')
105 bm = c.modules.BootManager.BootManager(log,'boot')
107 BootManagerException = c.modules.Exceptions.BootManagerException
108 InitializeBootManager = c.modules.BootManager.InitializeBootManager
109 ReadNodeConfiguration = c.modules.BootManager.ReadNodeConfiguration
112 plcnode = api.GetNodes({'hostname': self.node}, None)[0]
114 InitializeBootManager.Run(bm.VARS, bm.LOG)
115 try: ReadNodeConfiguration.Run(bm.VARS, bm.LOG)
120 print " Possibly, unable to find valid configuration file"
123 print " NODE: %s" % bm.VARS['NODE_KEY']
124 print " PLC : %s" % plcnode['key']
126 if bm.VARS['NODE_KEY'] == plcnode['key']:
129 if api.UpdateNode(self.node, {'key': bm.VARS['NODE_KEY']}):
130 print " Successfully updated NODE_KEY with PLC"
135 #for key in bm.VARS.keys():
136 # print key, " == ", bm.VARS[key]
138 print " Unable to retrieve NODE_KEY"
140 def bootmanager_running(self):
141 if self.c.modules.os.path.exists('/tmp/BM_RUNNING'):
146 def set_nodestate(self, state='boot'):
147 return api.UpdateNode(self.node, {'boot_state' : state})
149 def restart_node(self, state='boot'):
150 api.UpdateNode(self.node, {'boot_state' : state})
152 pflags = PersistFlags(self.node, 1*60*60*24, db='restart_persistflags')
153 if not pflags.getRecentFlag('gentlekill'):
154 print " Killing all slice processes... : %s" % self.node
155 cmd_slicekill = "ls -d /proc/virtual/[0-9]* | awk -F '/' '{print $4}' | xargs -I{} /usr/sbin/vkill -s 9 --xid {} -- 0"
156 self.c.modules.os.system(cmd_slicekill)
157 cmd = """ shutdown -r +1 & """
158 print " Restarting %s : %s" % ( self.node, cmd)
159 self.c.modules.os.system(cmd)
161 pflags.setRecentFlag('gentlekill')
164 print " Restarting with sysrq 'sub' %s" % self.node
165 cmd = """ (sleep 5; echo 's' > /proc/sysrq-trigger; echo 'u' > /proc/sysrq-trigger; echo 'b' > /proc/sysrq-trigger ) & """
166 self.c.modules.os.system(cmd)
170 def restart_bootmanager(self, forceState):
172 self.c.modules.os.chdir('/tmp/source')
173 if self.c.modules.os.path.exists('/tmp/BM_RUNNING'):
174 print " BootManager is already running: try again soon..."
176 print " Starting 'BootManager.py %s' on %s " % (forceState, self.node)
177 cmd = "( touch /tmp/BM_RUNNING ; " + \
178 " python ./BootManager.py %s &> server.log < /dev/null ; " + \
179 " rm -f /tmp/BM_RUNNING " + \
181 cmd = cmd % forceState
182 self.c.modules.os.system(cmd)
188 class PlanetLabSession:
189 globalport = 22000 + int(random.random()*1000)
191 def __init__(self, node, nosetup, verbose):
192 self.verbose = verbose
195 self.nosetup = nosetup
199 def get_connection(self, config):
200 return NodeConnection(SocketConnection("localhost", self.port), self.node, config)
202 def setup_host(self):
203 self.port = PlanetLabSession.globalport
204 PlanetLabSession.globalport = PlanetLabSession.globalport + 1
207 args['port'] = self.port
208 args['user'] = 'root'
209 args['hostname'] = self.node
210 args['monitordir'] = config.MONITOR_SCRIPT_ROOT
214 print "Skipping setup"
217 # COPY Rpyc files to host
218 cmd = "rsync -qv -az -e ssh %(monitordir)s/Rpyc/ %(user)s@%(hostname)s:Rpyc 2> /dev/null" % args
219 if self.verbose: print cmd
222 localos = moncommands.CMD()
224 ret = localos.system(cmd, timeout)
227 print "\tUNKNOWN SSH KEY FOR %s; making an exception" % self.node
228 #print "MAKE EXPLICIT EXCEPTION FOR %s" % self.node
229 k = SSHKnownHosts(); k.updateDirect(self.node); k.write(); del k
230 ret = localos.system(cmd, timeout)
233 print "\tFAILED TWICE"
235 raise Exception("Failed twice trying to login with updated ssh host key")
238 # KILL any already running servers.
239 ssh = moncommands.SSH(args['user'], args['hostname'], ssh_port)
240 (ov,ev) = ssh.run_noexcept2("""<<\EOF
242 echo "kill server" >> out.log
243 ps ax | grep Rpyc | grep -v grep | awk '{print $1}' | xargs kill 2> /dev/null ;
244 echo "export" >> out.log
245 export PYTHONPATH=$HOME ;
246 echo "start server" >> out.log
247 python Rpyc/Servers/forking_server.py &> server.log &
248 echo "done" >> out.log
250 #cmd = """ssh %(user)s@%(hostname)s """ + \
251 # """'ps ax | grep Rpyc | grep -v grep | awk "{print \$1}" | xargs kill 2> /dev/null' """
253 #if self.verbose: print cmd
255 #print localos.system(cmd,timeout)
257 ## START a new rpyc server.
258 #cmd = """ssh -n %(user)s@%(hostname)s "export PYTHONPATH=\$HOME; """ + \
259 # """python Rpyc/Servers/forking_server.py &> server.log < /dev/null &" """
261 #if self.verbose: print cmd
262 #print localos.system(cmd,timeout)
266 # This was tricky to make synchronous. The combination of ssh-clients-4.7p1,
267 # and the following options seems to work well.
268 cmd = """ssh -o ExitOnForwardFailure=yes -o BatchMode=yes """ + \
269 """-o PermitLocalCommand=yes -o LocalCommand='echo "READY"' """ + \
270 """-o ConnectTimeout=120 """ + \
271 """-n -N -L %(port)s:localhost:18812 """ + \
272 """%(user)s@%(hostname)s"""
274 if self.verbose: print cmd
275 self.command = Sopen(cmd, shell=True, stdout=subprocess.PIPE)
276 # TODO: the read() here may block indefinitely. Need a better
277 # approach therefore, that includes a timeout.
278 #ret = self.command.stdout.read(5)
279 ret = moncommands.read_t(self.command.stdout, 5)
283 # NOTE: There is still a slight race for machines that are slow...
284 self.timeout = 2*(t2-t1)
285 print "Sleeping for %s sec" % self.timeout
286 time.sleep(self.timeout)
289 if self.command.returncode is not None:
290 print "Failed to establish tunnel!"
291 raise Exception("SSH Tunnel exception : %s %s" % (self.node, self.command.returncode))
293 raise Exception("Unknown SSH Tunnel Exception: still running, but did not report 'READY'")
297 if self.verbose: print "Killing SSH session %s" % self.port
301 def steps_to_list(steps):
303 for (id,label) in steps:
304 ret_list.append(label)
307 def index_to_id(steps,index):
308 if index < len(steps):
309 return steps[index][0]
313 def reboot(hostname, config=None, forced_action=None):
315 # NOTE: Nothing works if the bootcd is REALLY old.
316 # So, this is the first step.
317 fbnode = get_fbnode(hostname)
318 if fbnode['category'] == "OLDBOOTCD":
319 print "...NOTIFY OWNER TO UPDATE BOOTCD!!!"
321 args['hostname_list'] = " %s" % hostname
323 m = PersistMessage(hostname, "Please Update Boot Image for %s" % hostname,
324 mailtxt.newbootcd_one[1] % args, True, db='bootcd_persistmessages')
326 loginbase = plc.siteId(hostname)
327 emails = plc.getTechEmails(loginbase)
330 print "\tDisabling %s due to out-of-date BOOTCD" % hostname
331 api.UpdateNode(hostname, {'boot_state' : 'disable'})
335 print "Creating session for %s" % node
336 # update known_hosts file (in case the node has rebooted since last run)
337 if config and not config.quiet: print "...updating known_hosts ssh-rsa key for %s" % node
339 k = SSHKnownHosts(); k.update(node); k.write(); del k
341 from nodecommon import email_exception
343 print traceback.print_exc()
348 session = PlanetLabSession(node, False, True)
350 session = PlanetLabSession(node, config.nosetup, config.verbose)
352 msg = "ERROR setting up session for %s" % hostname
354 print traceback.print_exc()
355 from nodecommon import email_exception
361 conn = session.get_connection(config)
363 # NOTE: sometimes the wait in setup_host() is not long enough.
364 # So, here we try to wait a little longer before giving up entirely.
366 time.sleep(session.timeout*4)
367 conn = session.get_connection(config)
369 print traceback.print_exc()
370 from nodecommon import email_exception
374 if forced_action == "reboot":
375 conn.restart_node('rins')
378 boot_state = conn.get_boot_state()
379 if boot_state == "boot":
380 print "...Boot state of %s already completed : skipping..." % node
382 elif boot_state == "unknown":
383 print "...Unknown bootstate for %s : skipping..."% node
388 if conn.bootmanager_running():
389 print "...BootManager is currently running. Skipping host %s" % node
394 # conn.restart_bootmanager(config.force)
397 # Read persistent flags, tagged on one week intervals.
398 pflags = PersistFlags(hostname, 3*60*60*24, db='debug_persistflags')
401 if config and not config.quiet: print "...downloading dmesg from %s" % node
402 dmesg = conn.get_dmesg()
403 child = fdpexpect.fdspawn(dmesg)
408 ('scsierror' , 'SCSI error : <\d+ \d+ \d+ \d+> return code = 0x\d+'),
409 ('ioerror' , 'end_request: I/O error, dev sd\w+, sector \d+'),
410 ('ccisserror' , 'cciss: cmd \w+ has CHECK CONDITION byte \w+ = \w+'),
412 ('buffererror', 'Buffer I/O error on device dm-\d, logical block \d+'),
414 ('hdaseekerror', 'hda: dma_intr: status=0x\d+ { DriveReady SeekComplete Error }'),
415 ('hdacorrecterror', 'hda: dma_intr: error=0x\d+ { UncorrectableError }, LBAsect=\d+, sector=\d+'),
417 ('atareadyerror' , 'ata\d+: status=0x\d+ { DriveReady SeekComplete Error }'),
418 ('atacorrecterror' , 'ata\d+: error=0x\d+ { UncorrectableError }'),
420 ('sdXerror' , 'sd\w: Current: sense key: Medium Error'),
421 ('ext3error' , 'EXT3-fs error (device dm-\d+): ext3_find_entry: reading directory #\d+ offset \d+'),
423 ('floppytimeout','floppy0: floppy timeout called'),
424 ('floppyerror', 'end_request: I/O error, dev fd\w+, sector \d+'),
426 # hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
427 # hda: dma_intr: error=0x40 { UncorrectableError }, LBAsect=23331263, sector=23331263
429 # floppy0: floppy timeout called
430 # end_request: I/O error, dev fd0, sector 0
432 # Buffer I/O error on device dm-2, logical block 8888896
433 # ata1: status=0x51 { DriveReady SeekComplete Error }
434 # ata1: error=0x40 { UncorrectableError }
435 # SCSI error : <0 0 0 0> return code = 0x8000002
436 # sda: Current: sense key: Medium Error
437 # Additional sense: Unrecovered read error - auto reallocate failed
439 # SCSI error : <0 2 0 0> return code = 0x40001
440 # end_request: I/O error, dev sda, sector 572489600
442 id = index_to_id(steps, child.expect( steps_to_list(steps) + [ pexpect.EOF ]))
449 if config and not config.quiet: print "\tSET: ", s
452 print "...Potential drive errors on %s" % node
453 if len(s) == 2 and 'floppyerror' in s:
454 print "...Should investigate. Continuing with node."
456 print "...Should investigate. Skipping node."
457 # TODO: send message related to these errors.
459 args['hostname'] = hostname
460 args['log'] = conn.get_dmesg().read()
462 m = PersistMessage(hostname, mailtxt.baddisk[0] % args,
463 mailtxt.baddisk[1] % args, True, db='hardware_persistmessages')
465 loginbase = plc.siteId(hostname)
466 emails = plc.getTechEmails(loginbase)
468 conn.set_nodestate('disable')
471 print "...Downloading bm.log from %s" % node
472 log = conn.get_bootmanager_log()
473 child = fdpexpect.fdspawn(log)
476 if config.collect: return True
482 if config and not config.quiet: print "...Scanning bm.log for errors"
488 ('bminit' , 'Initializing the BootManager.'),
489 ('cfg' , 'Reading node configuration file.'),
490 ('auth' , 'Authenticating node with PLC.'),
491 ('getplc' , 'Retrieving details of node from PLC.'),
492 ('update' , 'Updating node boot state at PLC.'),
493 ('hardware' , 'Checking if hardware requirements met.'),
494 ('installinit' , 'Install: Initializing.'),
495 ('installdisk' , 'Install: partitioning disks.'),
496 ('installbootfs', 'Install: bootstrapfs tarball.'),
497 ('installcfg' , 'Install: Writing configuration files.'),
498 ('installstop' , 'Install: Shutting down installer.'),
499 ('update2' , 'Updating node boot state at PLC.'),
500 ('installinit2' , 'Install: Initializing.'),
501 ('validate' , 'Validating node installation.'),
502 ('rebuildinitrd', 'Rebuilding initrd'),
503 ('netcfg' , 'Install: Writing Network Configuration files.'),
504 ('update3' , 'Updating node configuration.'),
505 ('disk' , 'Checking for unused disks to add to LVM.'),
506 ('update4' , 'Sending hardware configuration to PLC.'),
507 ('debug' , 'Starting debug mode'),
508 ('bmexceptmount', 'BootManagerException during mount'),
509 ('bmexceptvgscan', 'BootManagerException during vgscan/vgchange'),
510 ('bmexceptrmfail', 'Unable to remove directory tree: /tmp/mnt'),
511 ('exception' , 'Exception'),
512 ('nocfg' , 'Found configuration file planet.cnf on floppy, but was unable to parse it.'),
513 ('protoerror' , 'XML RPC protocol error'),
514 ('nodehostname' , 'Configured node hostname does not resolve'),
515 ('implementerror', 'Implementation Error'),
516 ('readonlyfs' , '[Errno 30] Read-only file system'),
517 ('noinstall' , 'notinstalled'),
518 ('bziperror' , 'bzip2: Data integrity error when decompressing.'),
519 ('noblockdev' , "No block devices detected."),
520 ('dnserror' , 'Name or service not known'),
521 ('downloadfail' , 'Unable to download main tarball /boot/bootstrapfs-planetlab-i386.tar.bz2 from server.'),
522 ('disktoosmall' , 'The total usable disk size of all disks is insufficient to be usable as a PlanetLab node.'),
523 ('hardwarerequirefail' , 'Hardware requirements not met'),
524 ('mkfsfail' , 'while running: Running mkfs.ext2 -q -m 0 -j /dev/planetlab/vservers failed'),
525 ('nofilereference', "No such file or directory: '/tmp/mnt/sysimg//vservers/.vref/planetlab-f8-i386/etc/hosts'"),
526 ('kernelcopyfail', "cp: cannot stat `/tmp/mnt/sysimg/boot/kernel-boot': No such file or directory"),
527 ('chrootfail' , 'Running chroot /tmp/mnt/sysimg'),
528 ('modulefail' , 'Unable to get list of system modules'),
529 ('writeerror' , 'write error: No space left on device'),
530 ('nospace' , "No space left on device"),
531 ('nonode' , 'Failed to authenticate call: No such node'),
532 ('authfail' , 'Failed to authenticate call: Call could not be authenticated'),
533 ('bootcheckfail' , 'BootCheckAuthentication'),
534 ('bootupdatefail' , 'BootUpdateNode'),
536 list = steps_to_list(steps)
537 index = child.expect( list + [ pexpect.EOF ])
538 id = index_to_id(steps,index)
541 if id == "exception":
542 if config and not config.quiet: print "...Found An Exception!!!"
543 elif index == len(list):
547 s = "-".join(sequence)
548 print " FOUND SEQUENCE: ", s
550 # NOTE: We get or set the flag based on the current sequence identifier.
551 # By using the sequence identifier, we guarantee that there will be no
552 # frequent loops. I'm guessing there is a better way to track loops,
554 #if not config.force and pflags.getRecentFlag(s):
555 # pflags.setRecentFlag(s)
557 # print "... flag is set or it has already run recently. Skipping %s" % node
563 # restart_bootmanager_boot
564 for n in ["bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-netcfg-update3-disk-update4-done",
565 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-update3-disk-update4-update3-exception-protoerror-update-protoerror-debug-done",
566 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-disk-update4-update3-update3-implementerror-bootupdatefail-update-debug-done",
568 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-disk-update4-update3-update3-exception-protoerror-update-protoerror-debug-done",
570 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-update3-disk-update4-update3-exception-protoerror-update-debug-done",
571 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-disk-update4-update3-exception-chrootfail-update-debug-done",
572 "bminit-cfg-auth-getplc-update-debug-done",
573 "bminit-cfg-auth-getplc-exception-protoerror-update-protoerror-debug-done",
574 "bminit-cfg-auth-protoerror-exception-update-protoerror-debug-done",
575 "bminit-cfg-auth-protoerror-exception-update-bootupdatefail-authfail-debug-done",
576 "bminit-cfg-auth-protoerror-exception-update-debug-done",
577 "bminit-cfg-auth-getplc-exception-protoerror-update-debug-done",
578 "bminit-cfg-auth-getplc-implementerror-update-debug-done",
580 sequences.update({n : "restart_bootmanager_boot"})
582 # conn.restart_bootmanager('rins')
583 for n in [ "bminit-cfg-auth-getplc-installinit-validate-exception-modulefail-update-debug-done",
584 "bminit-cfg-auth-getplc-update-installinit-validate-exception-modulefail-update-debug-done",
585 "bminit-cfg-auth-getplc-installinit-validate-bmexceptmount-exception-noinstall-update-debug-done",
586 "bminit-cfg-auth-getplc-update-installinit-validate-bmexceptmount-exception-noinstall-update-debug-done",
587 "bminit-cfg-auth-getplc-installinit-validate-bmexceptvgscan-exception-noinstall-update-debug-done",
588 "bminit-cfg-auth-getplc-update-installinit-validate-exception-noinstall-update-debug-done",
589 "bminit-cfg-auth-getplc-hardware-installinit-installdisk-bziperror-exception-update-debug-done",
590 "bminit-cfg-auth-getplc-update-hardware-installinit-installdisk-installbootfs-exception-update-debug-done",
591 "bminit-cfg-auth-getplc-update-installinit-validate-bmexceptvgscan-exception-noinstall-update-debug-done",
592 "bminit-cfg-auth-getplc-hardware-installinit-installdisk-installbootfs-exception-update-debug-done",
593 "bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-netcfg-update3-implementerror-nofilereference-update-debug-done",
594 "bminit-cfg-auth-getplc-update-hardware-installinit-installdisk-exception-mkfsfail-update-debug-done",
595 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-exception-chrootfail-update-debug-done",
596 "bminit-cfg-auth-getplc-update-hardware-installinit-installdisk-installbootfs-installcfg-installstop-update-installinit-validate-rebuildinitrd-netcfg-disk-update4-update3-update3-kernelcopyfail-exception-update-debug-done",
597 "bminit-cfg-auth-getplc-hardware-installinit-installdisk-installbootfs-installcfg-installstop-update-installinit-validate-rebuildinitrd-netcfg-disk-update4-update3-update3-kernelcopyfail-exception-update-debug-done",
598 "bminit-cfg-auth-getplc-installinit-validate-exception-noinstall-update-debug-done",
599 # actual solution appears to involve removing the bad files, and
600 # continually trying to boot the node.
601 "bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-netcfg-disk-update4-update3-update3-implementerror-update-debug-done",
603 sequences.update({n : "restart_bootmanager_rins"})
606 sequences.update({"bminit-cfg-auth-bootcheckfail-authfail-exception-update-bootupdatefail-authfail-debug-done": "repair_node_keys"})
608 # conn.restart_node('rins')
609 for n in ["bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-exception-chrootfail-update-debug-done",
610 "bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-netcfg-update3-disk-update4-exception-chrootfail-update-debug-done",
611 "bminit-cfg-auth-getplc-hardware-installinit-installdisk-installbootfs-installcfg-exception-chrootfail-update-debug-done",
612 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-writeerror-exception-chrootfail-update-debug-done",
613 "bminit-cfg-auth-getplc-update-hardware-installinit-exception-bmexceptrmfail-update-debug-done",
614 "bminit-cfg-auth-getplc-hardware-installinit-exception-bmexceptrmfail-update-debug-done",
615 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-update3-disk-update4-update3-implementerror-bootupdatefail-update-debug-done",
616 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-update3-implementerror-readonlyfs-update-debug-done",
617 "bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-netcfg-update3-nospace-exception-update-debug-done",
618 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-update3-implementerror-nospace-update-debug-done",
619 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-update3-implementerror-update-debug-done",
620 "bminit-cfg-auth-getplc-update-hardware-installinit-installdisk-installbootfs-exception-downloadfail-update-debug-done",
622 sequences.update({n : "restart_node_rins"})
625 for n in ["bminit-cfg-auth-getplc-implementerror-bootupdatefail-update-debug-done",
626 "bminit-cfg-auth-implementerror-bootcheckfail-update-debug-done",
627 "bminit-cfg-auth-implementerror-bootcheckfail-update-implementerror-bootupdatefail-done",
628 "bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-netcfg-update3-implementerror-nospace-update-debug-done",
629 "bminit-cfg-auth-getplc-hardware-installinit-installdisk-installbootfs-exception-downloadfail-update-debug-done",
630 "bminit-cfg-auth-getplc-update-installinit-validate-implementerror-update-debug-done",
632 sequences.update({n: "restart_node_boot"})
634 # update_node_config_email
635 for n in ["bminit-cfg-exception-nocfg-update-bootupdatefail-nonode-debug-done",
636 "bminit-cfg-exception-update-bootupdatefail-nonode-debug-done",
637 "bminit-cfg-auth-bootcheckfail-nonode-exception-update-bootupdatefail-nonode-debug-done",
639 sequences.update({n : "update_node_config_email"})
641 for n in [ "bminit-cfg-exception-nodehostname-update-debug-done",
642 "bminit-cfg-update-exception-nodehostname-update-debug-done",
644 sequences.update({n : "nodenetwork_email"})
646 # update_bootcd_email
647 for n in ["bminit-cfg-auth-getplc-update-hardware-exception-noblockdev-hardwarerequirefail-update-debug-done",
648 "bminit-cfg-auth-getplc-hardware-exception-noblockdev-hardwarerequirefail-update-debug-done",
649 "bminit-cfg-auth-getplc-update-hardware-noblockdev-exception-hardwarerequirefail-update-debug-done",
650 "bminit-cfg-auth-getplc-hardware-noblockdev-exception-hardwarerequirefail-update-debug-done",
651 "bminit-cfg-auth-getplc-hardware-exception-hardwarerequirefail-update-debug-done",
653 sequences.update({n : "update_bootcd_email"})
655 for n in [ "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-update3-implementerror-nofilereference-update-debug-done",
657 sequences.update({n: "suspect_error_email"})
659 # update_hardware_email
660 sequences.update({"bminit-cfg-auth-getplc-hardware-exception-disktoosmall-hardwarerequirefail-update-debug-done" : "update_hardware_email"})
661 sequences.update({"bminit-cfg-auth-getplc-hardware-disktoosmall-exception-hardwarerequirefail-update-debug-done" : "update_hardware_email"})
663 # broken_hardware_email
664 sequences.update({"bminit-cfg-auth-getplc-update-hardware-exception-hardwarerequirefail-update-debug-done" : "broken_hardware_email"})
668 "bminit-cfg-update-implementerror-bootupdatefail-dnserror-update-implementerror-bootupdatefail-dnserror-done",
669 "bminit-cfg-auth-implementerror-bootcheckfail-dnserror-update-implementerror-bootupdatefail-dnserror-done",
671 sequences.update( { n : "bad_dns_email"})
676 if s not in sequences:
677 print " HOST %s" % hostname
678 print " UNKNOWN SEQUENCE: %s" % s
681 args['hostname'] = hostname
683 args['bmlog'] = conn.get_bootmanager_log().read()
684 m = PersistMessage(hostname, mailtxt.unknownsequence[0] % args,
685 mailtxt.unknownsequence[1] % args, False, db='unknown_persistmessages')
687 m.send([config.cc_email])
689 conn.restart_bootmanager('boot')
691 # NOTE: Do not set the pflags value for this sequence if it's unknown.
692 # This way, we can check it again after we've fixed it.
697 if sequences[s] == "restart_bootmanager_boot":
698 if config and not config.quiet: print "...Restarting BootManager.py on %s "% node
699 conn.restart_bootmanager('boot')
700 elif sequences[s] == "restart_bootmanager_rins":
701 if config and not config.quiet: print "...Restarting BootManager.py on %s "% node
702 conn.restart_bootmanager('rins')
703 elif sequences[s] == "restart_node_rins":
704 conn.restart_node('rins')
705 elif sequences[s] == "restart_node_boot":
706 conn.restart_node('boot')
707 elif sequences[s] == "repair_node_keys":
708 if conn.compare_and_repair_nodekeys():
709 # the keys either are in sync or were forced in sync.
710 # so try to reboot the node again.
711 conn.restart_bootmanager('rins')
714 # there was some failure to synchronize the keys.
715 print "...Unable to repair node keys on %s" % node
717 elif sequences[s] == "suspect_error_email":
719 args['hostname'] = hostname
721 args['bmlog'] = conn.get_bootmanager_log().read()
722 m = PersistMessage(hostname, "Suspicous error from BootManager on %s" % args,
723 mailtxt.unknownsequence[1] % args, False, db='suspect_persistmessages')
725 m.send([config.cc_email])
727 conn.restart_bootmanager('boot')
729 elif sequences[s] == "update_node_config_email":
730 print "...Sending message to UPDATE NODE CONFIG"
732 args['hostname'] = hostname
733 m = PersistMessage(hostname, mailtxt.plnode_cfg[0] % args, mailtxt.plnode_cfg[1] % args,
734 True, db='nodeid_persistmessages')
735 loginbase = plc.siteId(hostname)
736 emails = plc.getTechEmails(loginbase)
738 conn.dump_plconf_file()
739 conn.set_nodestate('disable')
741 elif sequences[s] == "nodenetwork_email":
742 print "...Sending message to LOOK AT NODE NETWORK"
744 args['hostname'] = hostname
745 args['bmlog'] = conn.get_bootmanager_log().read()
746 m = PersistMessage(hostname, mailtxt.plnode_cfg[0] % args, mailtxt.plnode_cfg[1] % args,
747 True, db='nodenet_persistmessages')
748 loginbase = plc.siteId(hostname)
749 emails = plc.getTechEmails(loginbase)
751 conn.dump_plconf_file()
752 conn.set_nodestate('disable')
754 elif sequences[s] == "update_bootcd_email":
755 print "...NOTIFY OWNER TO UPDATE BOOTCD!!!"
758 args.update(getconf.getconf(hostname)) # NOTE: Generates boot images for the user:
759 args['hostname_list'] = "%s" % hostname
761 m = PersistMessage(hostname, "Please Update Boot Image for %s" % hostname,
762 mailtxt.newalphacd_one[1] % args, True, db='bootcd_persistmessages')
764 loginbase = plc.siteId(hostname)
765 emails = plc.getTechEmails(loginbase)
768 print "\tDisabling %s due to out-of-date BOOTCD" % hostname
769 conn.set_nodestate('disable')
771 elif sequences[s] == "broken_hardware_email":
772 # MAKE An ACTION record that this host has failed hardware. May
773 # require either an exception "/minhw" or other manual intervention.
774 # Definitely need to send out some more EMAIL.
775 print "...NOTIFYING OWNERS OF BROKEN HARDWARE on %s!!!" % hostname
776 # TODO: email notice of broken hardware
778 args['hostname'] = hostname
779 args['log'] = conn.get_dmesg().read()
780 m = PersistMessage(hostname, mailtxt.baddisk[0] % args,
781 mailtxt.baddisk[1] % args, True, db='hardware_persistmessages')
783 loginbase = plc.siteId(hostname)
784 emails = plc.getTechEmails(loginbase)
786 conn.set_nodestate('disable')
788 elif sequences[s] == "update_hardware_email":
789 print "...NOTIFYING OWNERS OF MINIMAL HARDWARE FAILURE on %s!!!" % hostname
791 args['hostname'] = hostname
792 args['bmlog'] = conn.get_bootmanager_log().read()
793 m = PersistMessage(hostname, mailtxt.minimalhardware[0] % args,
794 mailtxt.minimalhardware[1] % args, True, db='minhardware_persistmessages')
796 loginbase = plc.siteId(hostname)
797 emails = plc.getTechEmails(loginbase)
799 conn.set_nodestate('disable')
801 elif sequences[s] == "bad_dns_email":
802 print "...NOTIFYING OWNERS OF DNS FAILURE on %s!!!" % hostname
805 node = api.GetNodes(hostname)[0]
806 net = api.GetNodeNetworks(node['nodenetwork_ids'])[0]
808 from nodecommon import email_exception
810 print traceback.print_exc()
811 # TODO: api error. skip email, b/c all info is not available,
812 # flag_set will not be recorded.
814 nodenet_str = network_config_to_str(net)
816 args['hostname'] = hostname
817 args['network_config'] = nodenet_str
818 args['nodenetwork_id'] = net['nodenetwork_id']
819 m = PersistMessage(hostname, mailtxt.baddns[0] % args,
820 mailtxt.baddns[1] % args, True, db='baddns_persistmessages')
822 loginbase = plc.siteId(hostname)
823 emails = plc.getTechEmails(loginbase)
825 conn.set_nodestate('disable')
828 pflags.setRecentFlag(s)
834 # MAIN -------------------------------------------------------------------
837 import parser as parsermodule
838 parser = parsermodule.getParser()
840 parser.set_defaults(child=False, collect=False, nosetup=False, verbose=False,
841 force=None, quiet=False)
842 parser.add_option("", "--child", dest="child", action="store_true",
843 help="This is the child mode of this process.")
844 parser.add_option("", "--force", dest="force", metavar="boot_state",
845 help="Force a boot state passed to BootManager.py.")
846 parser.add_option("", "--quiet", dest="quiet", action="store_true",
847 help="Extra quiet output messages.")
848 parser.add_option("", "--verbose", dest="verbose", action="store_true",
849 help="Extra debug output messages.")
850 parser.add_option("", "--nonet", dest="nonet", action="store_true",
851 help="Do not setup the network, use existing log files to re-run a test pass.")
852 parser.add_option("", "--collect", dest="collect", action="store_true",
853 help="No action, just collect dmesg, and bm.log")
854 parser.add_option("", "--nosetup", dest="nosetup", action="store_true",
855 help="Do not perform the orginary setup phase.")
857 parser = parsermodule.getParser(['nodesets', 'defaults'], parser)
858 config = parsermodule.parse_args(parser)
861 nodes = config.getListFromFile(config.nodelist)
863 nodes = [ config.node ]
871 if __name__ == "__main__":