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
29 class ExceptionDoubleSSHError(Exception): pass
32 class Sopen(subprocess.Popen):
33 def kill(self, signal = signal.SIGTERM):
34 os.kill(self.pid, signal)
36 #from Rpyc import SocketConnection, Async
37 from Rpyc import SocketConnection, Async
38 from Rpyc.Utils import *
44 fb = database.dbLoad("findbad")
45 fbnode = fb['nodes'][node]['values']
49 def __init__(self, connection, node, config):
54 def get_boot_state(self):
55 if self.c.modules.os.path.exists('/tmp/source'):
57 elif self.c.modules.os.path.exists('/vservers'):
63 t_stamp = time.strftime("%Y-%m-%d-%H:%M")
64 self.c.modules.os.system("dmesg > /var/log/dmesg.bm.log")
65 download(self.c, "/var/log/dmesg.bm.log", "log/history/%s-dmesg.%s.log" % (t_stamp, self.node))
66 os.system("cp log/history/%s-dmesg.%s.log log/dmesg.%s.log" % (t_stamp, self.node, self.node))
67 log = open("log/dmesg.%s.log" % self.node, 'r')
70 def get_bootmanager_log(self):
71 t_stamp = time.strftime("%Y-%m-%d-%H:%M")
72 download(self.c, "/tmp/bm.log", "log/history/%s-bm.%s.log" % (t_stamp, self.node))
73 #os.system("zcat log/bm.%s.log.gz > log/bm.%s.log" % (self.node, self.node))
74 os.system("cp log/history/%s-bm.%s.log log/bm.%s.log" % (t_stamp, self.node, self.node))
75 log = open("log/bm.%s.log" % self.node, 'r')
78 def dump_plconf_file(self):
80 self.c.modules.sys.path.append("/tmp/source/")
81 self.c.modules.os.chdir('/tmp/source')
83 log = c.modules.BootManager.log('/tmp/new.log')
84 bm = c.modules.BootManager.BootManager(log,'boot')
86 BootManagerException = c.modules.Exceptions.BootManagerException
87 InitializeBootManager = c.modules.BootManager.InitializeBootManager
88 ReadNodeConfiguration = c.modules.BootManager.ReadNodeConfiguration
91 InitializeBootManager.Run(bm.VARS, bm.LOG)
92 try: ReadNodeConfiguration.Run(bm.VARS, bm.LOG)
96 print " Possibly, unable to find valid configuration file"
98 if bm_continue and self.config and not self.config.quiet:
99 for key in bm.VARS.keys():
100 print key, " == ", bm.VARS[key]
102 if self.config and not self.config.quiet: print " Unable to read Node Configuration"
105 def compare_and_repair_nodekeys(self):
107 self.c.modules.sys.path.append("/tmp/source/")
108 self.c.modules.os.chdir('/tmp/source')
110 log = c.modules.BootManager.log('/tmp/new.log')
111 bm = c.modules.BootManager.BootManager(log,'boot')
113 BootManagerException = c.modules.Exceptions.BootManagerException
114 InitializeBootManager = c.modules.BootManager.InitializeBootManager
115 ReadNodeConfiguration = c.modules.BootManager.ReadNodeConfiguration
118 plcnode = api.GetNodes({'hostname': self.node}, None)[0]
120 InitializeBootManager.Run(bm.VARS, bm.LOG)
121 try: ReadNodeConfiguration.Run(bm.VARS, bm.LOG)
126 print " Possibly, unable to find valid configuration file"
129 print " NODE: %s" % bm.VARS['NODE_KEY']
130 print " PLC : %s" % plcnode['key']
132 if bm.VARS['NODE_KEY'] == plcnode['key']:
135 if api.UpdateNode(self.node, {'key': bm.VARS['NODE_KEY']}):
136 print " Successfully updated NODE_KEY with PLC"
141 #for key in bm.VARS.keys():
142 # print key, " == ", bm.VARS[key]
144 print " Unable to retrieve NODE_KEY"
146 def bootmanager_running(self):
147 if self.c.modules.os.path.exists('/tmp/BM_RUNNING'):
152 def set_nodestate(self, state='boot'):
153 return api.UpdateNode(self.node, {'boot_state' : state})
155 def restart_node(self, state='boot'):
156 api.UpdateNode(self.node, {'boot_state' : state})
158 pflags = PersistFlags(self.node, 1*60*60*24, db='restart_persistflags')
159 if not pflags.getRecentFlag('gentlekill'):
160 print " Killing all slice processes... : %s" % self.node
161 cmd_slicekill = "ls -d /proc/virtual/[0-9]* | awk -F '/' '{print $4}' | xargs -I{} /usr/sbin/vkill -s 9 --xid {} -- 0"
162 self.c.modules.os.system(cmd_slicekill)
163 cmd = """ shutdown -r +1 & """
164 print " Restarting %s : %s" % ( self.node, cmd)
165 self.c.modules.os.system(cmd)
167 pflags.setRecentFlag('gentlekill')
170 print " Restarting with sysrq 'sub' %s" % self.node
171 cmd = """ (sleep 5; echo 's' > /proc/sysrq-trigger; echo 'u' > /proc/sysrq-trigger; echo 'b' > /proc/sysrq-trigger ) & """
172 self.c.modules.os.system(cmd)
176 def restart_bootmanager(self, forceState):
178 self.c.modules.os.chdir('/tmp/source')
179 if self.c.modules.os.path.exists('/tmp/BM_RUNNING'):
180 print " BootManager is already running: try again soon..."
182 print " Starting 'BootManager.py %s' on %s " % (forceState, self.node)
183 cmd = "( touch /tmp/BM_RUNNING ; " + \
184 " python ./BootManager.py %s &> server.log < /dev/null ; " + \
185 " rm -f /tmp/BM_RUNNING " + \
187 cmd = cmd % forceState
188 self.c.modules.os.system(cmd)
194 class PlanetLabSession:
195 globalport = 22000 + int(random.random()*1000)
197 def __init__(self, node, nosetup, verbose):
198 self.verbose = verbose
201 self.nosetup = nosetup
205 def get_connection(self, config):
206 return NodeConnection(SocketConnection("localhost", self.port), self.node, config)
208 def setup_host(self):
209 self.port = PlanetLabSession.globalport
210 PlanetLabSession.globalport = PlanetLabSession.globalport + 1
213 args['port'] = self.port
214 args['user'] = 'root'
215 args['hostname'] = self.node
216 args['monitordir'] = config.MONITOR_SCRIPT_ROOT
220 print "Skipping setup"
223 # COPY Rpyc files to host
224 cmd = "rsync -qv -az -e ssh %(monitordir)s/Rpyc/ %(user)s@%(hostname)s:Rpyc 2> /dev/null" % args
225 if self.verbose: print cmd
228 localos = moncommands.CMD()
230 ret = localos.system(cmd, timeout)
233 print "\tUNKNOWN SSH KEY FOR %s; making an exception" % self.node
234 #print "MAKE EXPLICIT EXCEPTION FOR %s" % self.node
235 k = SSHKnownHosts(); k.updateDirect(self.node); k.write(); del k
236 ret = localos.system(cmd, timeout)
239 print "\tFAILED TWICE"
241 raise ExceptionDoubleSSHError("Failed twice trying to login with updated ssh host key")
244 # KILL any already running servers.
245 ssh = moncommands.SSH(args['user'], args['hostname'], ssh_port)
246 (ov,ev) = ssh.run_noexcept2("""<<\EOF
248 echo "kill server" >> out.log
249 ps ax | grep Rpyc | grep -v grep | awk '{print $1}' | xargs kill 2> /dev/null ;
250 echo "export" >> out.log
251 export PYTHONPATH=$HOME ;
252 echo "start server" >> out.log
253 python Rpyc/Servers/forking_server.py &> server.log &
254 echo "done" >> out.log
256 #cmd = """ssh %(user)s@%(hostname)s """ + \
257 # """'ps ax | grep Rpyc | grep -v grep | awk "{print \$1}" | xargs kill 2> /dev/null' """
259 #if self.verbose: print cmd
261 #print localos.system(cmd,timeout)
263 ## START a new rpyc server.
264 #cmd = """ssh -n %(user)s@%(hostname)s "export PYTHONPATH=\$HOME; """ + \
265 # """python Rpyc/Servers/forking_server.py &> server.log < /dev/null &" """
267 #if self.verbose: print cmd
268 #print localos.system(cmd,timeout)
272 # This was tricky to make synchronous. The combination of ssh-clients-4.7p1,
273 # and the following options seems to work well.
274 cmd = """ssh -o ExitOnForwardFailure=yes -o BatchMode=yes """ + \
275 """-o PermitLocalCommand=yes -o LocalCommand='echo "READY"' """ + \
276 """-o ConnectTimeout=120 """ + \
277 """-n -N -L %(port)s:localhost:18812 """ + \
278 """%(user)s@%(hostname)s"""
280 if self.verbose: print cmd
281 self.command = Sopen(cmd, shell=True, stdout=subprocess.PIPE)
282 # TODO: the read() here may block indefinitely. Need a better
283 # approach therefore, that includes a timeout.
284 #ret = self.command.stdout.read(5)
285 ret = moncommands.read_t(self.command.stdout, 5)
289 # NOTE: There is still a slight race for machines that are slow...
290 self.timeout = 2*(t2-t1)
291 print "Sleeping for %s sec" % self.timeout
292 time.sleep(self.timeout)
295 if self.command.returncode is not None:
296 print "Failed to establish tunnel!"
297 raise Exception("SSH Tunnel exception : %s %s" % (self.node, self.command.returncode))
299 raise Exception("Unknown SSH Tunnel Exception: still running, but did not report 'READY'")
303 if self.verbose: print "Killing SSH session %s" % self.port
307 def steps_to_list(steps):
309 for (id,label) in steps:
310 ret_list.append(label)
313 def index_to_id(steps,index):
314 if index < len(steps):
315 return steps[index][0]
319 def reboot(hostname, config=None, forced_action=None):
321 # NOTE: Nothing works if the bootcd is REALLY old.
322 # So, this is the first step.
323 fbnode = get_fbnode(hostname)
324 if fbnode['category'] == "OLDBOOTCD":
325 print "...NOTIFY OWNER TO UPDATE BOOTCD!!!"
327 args['hostname_list'] = " %s" % hostname
329 m = PersistMessage(hostname, "Please Update Boot Image for %s" % hostname,
330 mailtxt.newbootcd_one[1] % args, True, db='bootcd_persistmessages')
332 loginbase = plc.siteId(hostname)
333 emails = plc.getTechEmails(loginbase)
336 print "\tDisabling %s due to out-of-date BOOTCD" % hostname
337 api.UpdateNode(hostname, {'boot_state' : 'disable'})
341 print "Creating session for %s" % node
342 # update known_hosts file (in case the node has rebooted since last run)
343 if config and not config.quiet: print "...updating known_hosts ssh-rsa key for %s" % node
345 k = SSHKnownHosts(); k.update(node); k.write(); del k
347 from nodecommon import email_exception
349 print traceback.print_exc()
354 session = PlanetLabSession(node, False, True)
356 session = PlanetLabSession(node, config.nosetup, config.verbose)
357 except ExceptionDoubleSSHError, e:
358 msg = "ERROR setting up session for %s" % hostname
362 msg = "ERROR setting up session for %s" % hostname
364 print traceback.print_exc()
365 from nodecommon import email_exception
371 conn = session.get_connection(config)
373 # NOTE: sometimes the wait in setup_host() is not long enough.
374 # So, here we try to wait a little longer before giving up entirely.
376 time.sleep(session.timeout*4)
377 conn = session.get_connection(config)
379 # failed twice... no need to report this really, it's just in a
383 print traceback.print_exc()
384 from nodecommon import email_exception
385 email_exception(node)
388 if forced_action == "reboot":
389 conn.restart_node('rins')
392 boot_state = conn.get_boot_state()
393 if boot_state == "boot":
394 print "...Boot state of %s already completed : skipping..." % node
396 elif boot_state == "unknown":
397 print "...Unknown bootstate for %s : skipping..."% node
402 if conn.bootmanager_running():
403 print "...BootManager is currently running. Skipping host %s" % node
408 # conn.restart_bootmanager(config.force)
411 # Read persistent flags, tagged on one week intervals.
412 pflags = PersistFlags(hostname, 3*60*60*24, db='debug_persistflags')
415 if config and not config.quiet: print "...downloading dmesg from %s" % node
416 dmesg = conn.get_dmesg()
417 child = fdpexpect.fdspawn(dmesg)
422 ('scsierror' , 'SCSI error : <\d+ \d+ \d+ \d+> return code = 0x\d+'),
423 ('ioerror' , 'end_request: I/O error, dev sd\w+, sector \d+'),
424 ('ccisserror' , 'cciss: cmd \w+ has CHECK CONDITION byte \w+ = \w+'),
426 ('buffererror', 'Buffer I/O error on device dm-\d, logical block \d+'),
428 ('hdaseekerror', 'hda: dma_intr: status=0x\d+ { DriveReady SeekComplete Error }'),
429 ('hdacorrecterror', 'hda: dma_intr: error=0x\d+ { UncorrectableError }, LBAsect=\d+, sector=\d+'),
431 ('atareadyerror' , 'ata\d+: status=0x\d+ { DriveReady SeekComplete Error }'),
432 ('atacorrecterror' , 'ata\d+: error=0x\d+ { UncorrectableError }'),
434 ('sdXerror' , 'sd\w: Current: sense key: Medium Error'),
435 ('ext3error' , 'EXT3-fs error (device dm-\d+): ext3_find_entry: reading directory #\d+ offset \d+'),
437 ('floppytimeout','floppy0: floppy timeout called'),
438 ('floppyerror', 'end_request: I/O error, dev fd\w+, sector \d+'),
440 # hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
441 # hda: dma_intr: error=0x40 { UncorrectableError }, LBAsect=23331263, sector=23331263
443 # floppy0: floppy timeout called
444 # end_request: I/O error, dev fd0, sector 0
446 # Buffer I/O error on device dm-2, logical block 8888896
447 # ata1: status=0x51 { DriveReady SeekComplete Error }
448 # ata1: error=0x40 { UncorrectableError }
449 # SCSI error : <0 0 0 0> return code = 0x8000002
450 # sda: Current: sense key: Medium Error
451 # Additional sense: Unrecovered read error - auto reallocate failed
453 # SCSI error : <0 2 0 0> return code = 0x40001
454 # end_request: I/O error, dev sda, sector 572489600
456 id = index_to_id(steps, child.expect( steps_to_list(steps) + [ pexpect.EOF ]))
463 if config and not config.quiet: print "\tSET: ", s
466 print "...Potential drive errors on %s" % node
467 if len(s) == 2 and 'floppyerror' in s:
468 print "...Should investigate. Continuing with node."
470 print "...Should investigate. Skipping node."
471 # TODO: send message related to these errors.
473 args['hostname'] = hostname
474 args['log'] = conn.get_dmesg().read()
476 m = PersistMessage(hostname, mailtxt.baddisk[0] % args,
477 mailtxt.baddisk[1] % args, True, db='hardware_persistmessages')
479 loginbase = plc.siteId(hostname)
480 emails = plc.getTechEmails(loginbase)
482 conn.set_nodestate('disable')
485 print "...Downloading bm.log from %s" % node
486 log = conn.get_bootmanager_log()
487 child = fdpexpect.fdspawn(log)
490 if config.collect: return True
496 if config and not config.quiet: print "...Scanning bm.log for errors"
502 ('bminit' , 'Initializing the BootManager.'),
503 ('cfg' , 'Reading node configuration file.'),
504 ('auth' , 'Authenticating node with PLC.'),
505 ('getplc' , 'Retrieving details of node from PLC.'),
506 ('update' , 'Updating node boot state at PLC.'),
507 ('hardware' , 'Checking if hardware requirements met.'),
508 ('installinit' , 'Install: Initializing.'),
509 ('installdisk' , 'Install: partitioning disks.'),
510 ('installbootfs', 'Install: bootstrapfs tarball.'),
511 ('installcfg' , 'Install: Writing configuration files.'),
512 ('installstop' , 'Install: Shutting down installer.'),
513 ('update2' , 'Updating node boot state at PLC.'),
514 ('installinit2' , 'Install: Initializing.'),
515 ('validate' , 'Validating node installation.'),
516 ('rebuildinitrd', 'Rebuilding initrd'),
517 ('netcfg' , 'Install: Writing Network Configuration files.'),
518 ('update3' , 'Updating node configuration.'),
519 ('disk' , 'Checking for unused disks to add to LVM.'),
520 ('update4' , 'Sending hardware configuration to PLC.'),
521 ('debug' , 'Starting debug mode'),
522 ('bmexceptmount', 'BootManagerException during mount'),
523 ('bmexceptvgscan', 'BootManagerException during vgscan/vgchange'),
524 ('bmexceptrmfail', 'Unable to remove directory tree: /tmp/mnt'),
525 ('exception' , 'Exception'),
526 ('nocfg' , 'Found configuration file planet.cnf on floppy, but was unable to parse it.'),
527 ('protoerror' , 'XML RPC protocol error'),
528 ('nodehostname' , 'Configured node hostname does not resolve'),
529 ('implementerror', 'Implementation Error'),
530 ('readonlyfs' , '[Errno 30] Read-only file system'),
531 ('noinstall' , 'notinstalled'),
532 ('bziperror' , 'bzip2: Data integrity error when decompressing.'),
533 ('noblockdev' , "No block devices detected."),
534 ('dnserror' , 'Name or service not known'),
535 ('downloadfail' , 'Unable to download main tarball /boot/bootstrapfs-planetlab-i386.tar.bz2 from server.'),
536 ('disktoosmall' , 'The total usable disk size of all disks is insufficient to be usable as a PlanetLab node.'),
537 ('hardwarerequirefail' , 'Hardware requirements not met'),
538 ('mkfsfail' , 'while running: Running mkfs.ext2 -q -m 0 -j /dev/planetlab/vservers failed'),
539 ('nofilereference', "No such file or directory: '/tmp/mnt/sysimg//vservers/.vref/planetlab-f8-i386/etc/hosts'"),
540 ('kernelcopyfail', "cp: cannot stat `/tmp/mnt/sysimg/boot/kernel-boot': No such file or directory"),
541 ('chrootfail' , 'Running chroot /tmp/mnt/sysimg'),
542 ('modulefail' , 'Unable to get list of system modules'),
543 ('writeerror' , 'write error: No space left on device'),
544 ('nospace' , "No space left on device"),
545 ('nonode' , 'Failed to authenticate call: No such node'),
546 ('authfail' , 'Failed to authenticate call: Call could not be authenticated'),
547 ('bootcheckfail' , 'BootCheckAuthentication'),
548 ('bootupdatefail' , 'BootUpdateNode'),
550 list = steps_to_list(steps)
551 index = child.expect( list + [ pexpect.EOF ])
552 id = index_to_id(steps,index)
555 if id == "exception":
556 if config and not config.quiet: print "...Found An Exception!!!"
557 elif index == len(list):
561 s = "-".join(sequence)
562 print " FOUND SEQUENCE: ", s
564 # NOTE: We get or set the flag based on the current sequence identifier.
565 # By using the sequence identifier, we guarantee that there will be no
566 # frequent loops. I'm guessing there is a better way to track loops,
568 #if not config.force and pflags.getRecentFlag(s):
569 # pflags.setRecentFlag(s)
571 # print "... flag is set or it has already run recently. Skipping %s" % node
577 # restart_bootmanager_boot
578 for n in ["bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-netcfg-update3-disk-update4-done",
579 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-update3-disk-update4-update3-exception-protoerror-update-protoerror-debug-done",
580 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-disk-update4-update3-update3-implementerror-bootupdatefail-update-debug-done",
582 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-disk-update4-update3-update3-exception-protoerror-update-protoerror-debug-done",
584 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-update3-disk-update4-update3-exception-protoerror-update-debug-done",
585 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-disk-update4-update3-exception-chrootfail-update-debug-done",
586 "bminit-cfg-auth-getplc-update-debug-done",
587 "bminit-cfg-auth-getplc-exception-protoerror-update-protoerror-debug-done",
588 "bminit-cfg-auth-protoerror-exception-update-protoerror-debug-done",
589 "bminit-cfg-auth-protoerror-exception-update-bootupdatefail-authfail-debug-done",
590 "bminit-cfg-auth-protoerror-exception-update-debug-done",
591 "bminit-cfg-auth-getplc-exception-protoerror-update-debug-done",
592 "bminit-cfg-auth-getplc-implementerror-update-debug-done",
594 sequences.update({n : "restart_bootmanager_boot"})
596 # conn.restart_bootmanager('rins')
597 for n in [ "bminit-cfg-auth-getplc-installinit-validate-exception-modulefail-update-debug-done",
598 "bminit-cfg-auth-getplc-update-installinit-validate-exception-modulefail-update-debug-done",
599 "bminit-cfg-auth-getplc-installinit-validate-bmexceptmount-exception-noinstall-update-debug-done",
600 "bminit-cfg-auth-getplc-update-installinit-validate-bmexceptmount-exception-noinstall-update-debug-done",
601 "bminit-cfg-auth-getplc-installinit-validate-bmexceptvgscan-exception-noinstall-update-debug-done",
602 "bminit-cfg-auth-getplc-update-installinit-validate-exception-noinstall-update-debug-done",
603 "bminit-cfg-auth-getplc-hardware-installinit-installdisk-bziperror-exception-update-debug-done",
604 "bminit-cfg-auth-getplc-update-hardware-installinit-installdisk-installbootfs-exception-update-debug-done",
605 "bminit-cfg-auth-getplc-update-installinit-validate-bmexceptvgscan-exception-noinstall-update-debug-done",
606 "bminit-cfg-auth-getplc-hardware-installinit-installdisk-installbootfs-exception-update-debug-done",
607 "bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-netcfg-update3-implementerror-nofilereference-update-debug-done",
608 "bminit-cfg-auth-getplc-update-hardware-installinit-installdisk-exception-mkfsfail-update-debug-done",
609 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-exception-chrootfail-update-debug-done",
610 "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",
611 "bminit-cfg-auth-getplc-hardware-installinit-installdisk-installbootfs-installcfg-installstop-update-installinit-validate-rebuildinitrd-netcfg-disk-update4-update3-update3-kernelcopyfail-exception-update-debug-done",
612 "bminit-cfg-auth-getplc-installinit-validate-exception-noinstall-update-debug-done",
613 # actual solution appears to involve removing the bad files, and
614 # continually trying to boot the node.
615 "bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-netcfg-disk-update4-update3-update3-implementerror-update-debug-done",
616 "bminit-cfg-auth-getplc-installinit-validate-exception-bmexceptmount-exception-noinstall-update-debug-done",
617 "bminit-cfg-auth-getplc-update-installinit-validate-exception-bmexceptmount-exception-noinstall-update-debug-done",
619 sequences.update({n : "restart_bootmanager_rins"})
622 sequences.update({"bminit-cfg-auth-bootcheckfail-authfail-exception-update-bootupdatefail-authfail-debug-done": "repair_node_keys"})
624 # conn.restart_node('rins')
625 for n in ["bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-exception-chrootfail-update-debug-done",
626 "bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-netcfg-update3-disk-update4-exception-chrootfail-update-debug-done",
627 "bminit-cfg-auth-getplc-hardware-installinit-installdisk-installbootfs-installcfg-exception-chrootfail-update-debug-done",
628 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-writeerror-exception-chrootfail-update-debug-done",
629 "bminit-cfg-auth-getplc-update-hardware-installinit-exception-bmexceptrmfail-update-debug-done",
630 "bminit-cfg-auth-getplc-hardware-installinit-exception-bmexceptrmfail-update-debug-done",
631 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-update3-disk-update4-update3-implementerror-bootupdatefail-update-debug-done",
632 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-update3-implementerror-readonlyfs-update-debug-done",
633 "bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-netcfg-update3-nospace-exception-update-debug-done",
634 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-update3-implementerror-nospace-update-debug-done",
635 "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-update3-implementerror-update-debug-done",
636 "bminit-cfg-auth-getplc-update-hardware-installinit-installdisk-installbootfs-exception-downloadfail-update-debug-done",
638 sequences.update({n : "restart_node_rins"})
641 for n in ["bminit-cfg-auth-getplc-implementerror-bootupdatefail-update-debug-done",
642 "bminit-cfg-auth-implementerror-bootcheckfail-update-debug-done",
643 "bminit-cfg-auth-implementerror-bootcheckfail-update-implementerror-bootupdatefail-done",
644 "bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-netcfg-update3-implementerror-nospace-update-debug-done",
645 "bminit-cfg-auth-getplc-hardware-installinit-installdisk-installbootfs-exception-downloadfail-update-debug-done",
646 "bminit-cfg-auth-getplc-update-installinit-validate-implementerror-update-debug-done",
648 sequences.update({n: "restart_node_boot"})
650 # update_node_config_email
651 for n in ["bminit-cfg-exception-nocfg-update-bootupdatefail-nonode-debug-done",
652 "bminit-cfg-exception-update-bootupdatefail-nonode-debug-done",
653 "bminit-cfg-auth-bootcheckfail-nonode-exception-update-bootupdatefail-nonode-debug-done",
655 sequences.update({n : "update_node_config_email"})
657 for n in [ "bminit-cfg-exception-nodehostname-update-debug-done",
658 "bminit-cfg-update-exception-nodehostname-update-debug-done",
660 sequences.update({n : "nodenetwork_email"})
662 # update_bootcd_email
663 for n in ["bminit-cfg-auth-getplc-update-hardware-exception-noblockdev-hardwarerequirefail-update-debug-done",
664 "bminit-cfg-auth-getplc-hardware-exception-noblockdev-hardwarerequirefail-update-debug-done",
665 "bminit-cfg-auth-getplc-update-hardware-noblockdev-exception-hardwarerequirefail-update-debug-done",
666 "bminit-cfg-auth-getplc-hardware-noblockdev-exception-hardwarerequirefail-update-debug-done",
667 "bminit-cfg-auth-getplc-hardware-exception-hardwarerequirefail-update-debug-done",
669 sequences.update({n : "update_bootcd_email"})
671 for n in [ "bminit-cfg-auth-getplc-installinit-validate-rebuildinitrd-netcfg-update3-implementerror-nofilereference-update-debug-done",
673 sequences.update({n: "suspect_error_email"})
675 # update_hardware_email
676 sequences.update({"bminit-cfg-auth-getplc-hardware-exception-disktoosmall-hardwarerequirefail-update-debug-done" : "update_hardware_email"})
677 sequences.update({"bminit-cfg-auth-getplc-hardware-disktoosmall-exception-hardwarerequirefail-update-debug-done" : "update_hardware_email"})
679 # broken_hardware_email
680 sequences.update({"bminit-cfg-auth-getplc-update-hardware-exception-hardwarerequirefail-update-debug-done" : "broken_hardware_email"})
684 "bminit-cfg-update-implementerror-bootupdatefail-dnserror-update-implementerror-bootupdatefail-dnserror-done",
685 "bminit-cfg-auth-implementerror-bootcheckfail-dnserror-update-implementerror-bootupdatefail-dnserror-done",
687 sequences.update( { n : "bad_dns_email"})
692 if s not in sequences:
693 print " HOST %s" % hostname
694 print " UNKNOWN SEQUENCE: %s" % s
697 args['hostname'] = hostname
699 args['bmlog'] = conn.get_bootmanager_log().read()
700 m = PersistMessage(hostname, mailtxt.unknownsequence[0] % args,
701 mailtxt.unknownsequence[1] % args, False, db='unknown_persistmessages')
703 m.send([config.cc_email])
705 conn.restart_bootmanager('boot')
707 # NOTE: Do not set the pflags value for this sequence if it's unknown.
708 # This way, we can check it again after we've fixed it.
713 if sequences[s] == "restart_bootmanager_boot":
714 if config and not config.quiet: print "...Restarting BootManager.py on %s "% node
715 conn.restart_bootmanager('boot')
716 elif sequences[s] == "restart_bootmanager_rins":
717 if config and not config.quiet: print "...Restarting BootManager.py on %s "% node
718 conn.restart_bootmanager('rins')
719 elif sequences[s] == "restart_node_rins":
720 conn.restart_node('rins')
721 elif sequences[s] == "restart_node_boot":
722 conn.restart_node('boot')
723 elif sequences[s] == "repair_node_keys":
724 if conn.compare_and_repair_nodekeys():
725 # the keys either are in sync or were forced in sync.
726 # so try to reboot the node again.
727 conn.restart_bootmanager('rins')
730 # there was some failure to synchronize the keys.
731 print "...Unable to repair node keys on %s" % node
733 elif sequences[s] == "suspect_error_email":
735 args['hostname'] = hostname
737 args['bmlog'] = conn.get_bootmanager_log().read()
738 m = PersistMessage(hostname, "Suspicous error from BootManager on %s" % args,
739 mailtxt.unknownsequence[1] % args, False, db='suspect_persistmessages')
741 m.send([config.cc_email])
743 conn.restart_bootmanager('boot')
745 elif sequences[s] == "update_node_config_email":
746 print "...Sending message to UPDATE NODE CONFIG"
748 args['hostname'] = hostname
749 m = PersistMessage(hostname, mailtxt.plnode_cfg[0] % args, mailtxt.plnode_cfg[1] % args,
750 True, db='nodeid_persistmessages')
751 loginbase = plc.siteId(hostname)
752 emails = plc.getTechEmails(loginbase)
754 conn.dump_plconf_file()
755 conn.set_nodestate('disable')
757 elif sequences[s] == "nodenetwork_email":
758 print "...Sending message to LOOK AT NODE NETWORK"
760 args['hostname'] = hostname
761 args['bmlog'] = conn.get_bootmanager_log().read()
762 m = PersistMessage(hostname, mailtxt.plnode_cfg[0] % args, mailtxt.plnode_cfg[1] % args,
763 True, db='nodenet_persistmessages')
764 loginbase = plc.siteId(hostname)
765 emails = plc.getTechEmails(loginbase)
767 conn.dump_plconf_file()
768 conn.set_nodestate('disable')
770 elif sequences[s] == "update_bootcd_email":
771 print "...NOTIFY OWNER TO UPDATE BOOTCD!!!"
774 args.update(getconf.getconf(hostname)) # NOTE: Generates boot images for the user:
775 args['hostname_list'] = "%s" % hostname
777 m = PersistMessage(hostname, "Please Update Boot Image for %s" % hostname,
778 mailtxt.newalphacd_one[1] % args, True, db='bootcd_persistmessages')
780 loginbase = plc.siteId(hostname)
781 emails = plc.getTechEmails(loginbase)
784 print "\tDisabling %s due to out-of-date BOOTCD" % hostname
785 conn.set_nodestate('disable')
787 elif sequences[s] == "broken_hardware_email":
788 # MAKE An ACTION record that this host has failed hardware. May
789 # require either an exception "/minhw" or other manual intervention.
790 # Definitely need to send out some more EMAIL.
791 print "...NOTIFYING OWNERS OF BROKEN HARDWARE on %s!!!" % hostname
792 # TODO: email notice of broken hardware
794 args['hostname'] = hostname
795 args['log'] = conn.get_dmesg().read()
796 m = PersistMessage(hostname, mailtxt.baddisk[0] % args,
797 mailtxt.baddisk[1] % args, True, db='hardware_persistmessages')
799 loginbase = plc.siteId(hostname)
800 emails = plc.getTechEmails(loginbase)
802 conn.set_nodestate('disable')
804 elif sequences[s] == "update_hardware_email":
805 print "...NOTIFYING OWNERS OF MINIMAL HARDWARE FAILURE on %s!!!" % hostname
807 args['hostname'] = hostname
808 args['bmlog'] = conn.get_bootmanager_log().read()
809 m = PersistMessage(hostname, mailtxt.minimalhardware[0] % args,
810 mailtxt.minimalhardware[1] % args, True, db='minhardware_persistmessages')
812 loginbase = plc.siteId(hostname)
813 emails = plc.getTechEmails(loginbase)
815 conn.set_nodestate('disable')
817 elif sequences[s] == "bad_dns_email":
818 print "...NOTIFYING OWNERS OF DNS FAILURE on %s!!!" % hostname
821 node = api.GetNodes(hostname)[0]
822 net = api.GetNodeNetworks(node['nodenetwork_ids'])[0]
824 from nodecommon import email_exception
826 print traceback.print_exc()
827 # TODO: api error. skip email, b/c all info is not available,
828 # flag_set will not be recorded.
830 nodenet_str = network_config_to_str(net)
832 args['hostname'] = hostname
833 args['network_config'] = nodenet_str
834 args['nodenetwork_id'] = net['nodenetwork_id']
835 m = PersistMessage(hostname, mailtxt.baddns[0] % args,
836 mailtxt.baddns[1] % args, True, db='baddns_persistmessages')
838 loginbase = plc.siteId(hostname)
839 emails = plc.getTechEmails(loginbase)
841 conn.set_nodestate('disable')
844 pflags.setRecentFlag(s)
850 # MAIN -------------------------------------------------------------------
853 import parser as parsermodule
854 parser = parsermodule.getParser()
856 parser.set_defaults(child=False, collect=False, nosetup=False, verbose=False,
857 force=None, quiet=False)
858 parser.add_option("", "--child", dest="child", action="store_true",
859 help="This is the child mode of this process.")
860 parser.add_option("", "--force", dest="force", metavar="boot_state",
861 help="Force a boot state passed to BootManager.py.")
862 parser.add_option("", "--quiet", dest="quiet", action="store_true",
863 help="Extra quiet output messages.")
864 parser.add_option("", "--verbose", dest="verbose", action="store_true",
865 help="Extra debug output messages.")
866 parser.add_option("", "--nonet", dest="nonet", action="store_true",
867 help="Do not setup the network, use existing log files to re-run a test pass.")
868 parser.add_option("", "--collect", dest="collect", action="store_true",
869 help="No action, just collect dmesg, and bm.log")
870 parser.add_option("", "--nosetup", dest="nosetup", action="store_true",
871 help="Do not perform the orginary setup phase.")
873 parser = parsermodule.getParser(['nodesets', 'defaults'], parser)
874 config = parsermodule.parse_args(parser)
877 nodes = config.getListFromFile(config.nodelist)
879 nodes = [ config.node ]
887 if __name__ == "__main__":