4 # Copyright (C) 2006, 2007, 2008, 2010, 2011 Google Inc.
6 # This program is free software; you can redistribute it and/or modify
7 # it under the terms of the GNU General Public License as published by
8 # the Free Software Foundation; either version 2 of the License, or
9 # (at your option) any later version.
11 # This program is distributed in the hope that it will be useful, but
12 # WITHOUT ANY WARRANTY; without even the implied warranty of
13 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 # General Public License for more details.
16 # You should have received a copy of the GNU General Public License
17 # along with this program; if not, write to the Free Software
18 # Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
22 """Functions to bootstrap a new cluster.
32 from ganeti import rpc
33 from ganeti import ssh
34 from ganeti import utils
35 from ganeti import errors
36 from ganeti import config
37 from ganeti import constants
38 from ganeti import objects
39 from ganeti import ssconf
40 from ganeti import serializer
41 from ganeti import hypervisor
42 from ganeti import bdev
43 from ganeti import netutils
44 from ganeti import backend
45 from ganeti import luxi
46 from ganeti import jstore
49 # ec_id for InitConfig's temporary reservation manager
50 _INITCONF_ECID = "initconfig-ecid"
52 #: After how many seconds daemon must be responsive
53 _DAEMON_READY_TIMEOUT = 10.0
57 """Setup the SSH configuration for the cluster.
59 This generates a dsa keypair for root, adds the pub key to the
60 permitted hosts and adds the hostkey to its own known hosts.
63 priv_key, pub_key, auth_keys = ssh.GetUserFiles(constants.GANETI_RUNAS)
65 for name in priv_key, pub_key:
66 if os.path.exists(name):
67 utils.CreateBackup(name)
68 utils.RemoveFile(name)
70 result = utils.RunCmd(["ssh-keygen", "-t", "dsa",
74 raise errors.OpExecError("Could not generate ssh keypair, error %s" %
77 utils.AddAuthorizedKey(auth_keys, utils.ReadFile(pub_key))
80 def GenerateHmacKey(file_name):
81 """Writes a new HMAC key.
84 @param file_name: Path to output file
87 utils.WriteFile(file_name, data="%s\n" % utils.GenerateSecret(), mode=0400,
91 def GenerateClusterCrypto(new_cluster_cert, new_rapi_cert, new_confd_hmac_key,
92 new_cds, rapi_cert_pem=None, cds=None,
93 nodecert_file=constants.NODED_CERT_FILE,
94 rapicert_file=constants.RAPI_CERT_FILE,
95 hmackey_file=constants.CONFD_HMAC_KEY,
96 cds_file=constants.CLUSTER_DOMAIN_SECRET_FILE):
97 """Updates the cluster certificates, keys and secrets.
99 @type new_cluster_cert: bool
100 @param new_cluster_cert: Whether to generate a new cluster certificate
101 @type new_rapi_cert: bool
102 @param new_rapi_cert: Whether to generate a new RAPI certificate
103 @type new_confd_hmac_key: bool
104 @param new_confd_hmac_key: Whether to generate a new HMAC key
106 @param new_cds: Whether to generate a new cluster domain secret
107 @type rapi_cert_pem: string
108 @param rapi_cert_pem: New RAPI certificate in PEM format
110 @param cds: New cluster domain secret
111 @type nodecert_file: string
112 @param nodecert_file: optional override of the node cert file path
113 @type rapicert_file: string
114 @param rapicert_file: optional override of the rapi cert file path
115 @type hmackey_file: string
116 @param hmackey_file: optional override of the hmac key file path
119 # noded SSL certificate
120 cluster_cert_exists = os.path.exists(nodecert_file)
121 if new_cluster_cert or not cluster_cert_exists:
122 if cluster_cert_exists:
123 utils.CreateBackup(nodecert_file)
125 logging.debug("Generating new cluster certificate at %s", nodecert_file)
126 utils.GenerateSelfSignedSslCert(nodecert_file)
129 if new_confd_hmac_key or not os.path.exists(hmackey_file):
130 logging.debug("Writing new confd HMAC key to %s", hmackey_file)
131 GenerateHmacKey(hmackey_file)
134 rapi_cert_exists = os.path.exists(rapicert_file)
137 # Assume rapi_pem contains a valid PEM-formatted certificate and key
138 logging.debug("Writing RAPI certificate at %s", rapicert_file)
139 utils.WriteFile(rapicert_file, data=rapi_cert_pem, backup=True)
141 elif new_rapi_cert or not rapi_cert_exists:
143 utils.CreateBackup(rapicert_file)
145 logging.debug("Generating new RAPI certificate at %s", rapicert_file)
146 utils.GenerateSelfSignedSslCert(rapicert_file)
148 # Cluster domain secret
150 logging.debug("Writing cluster domain secret to %s", cds_file)
151 utils.WriteFile(cds_file, data=cds, backup=True)
153 elif new_cds or not os.path.exists(cds_file):
154 logging.debug("Generating new cluster domain secret at %s", cds_file)
155 GenerateHmacKey(cds_file)
158 def _InitGanetiServerSetup(master_name):
159 """Setup the necessary configuration for the initial node daemon.
161 This creates the nodepass file containing the shared password for
162 the cluster, generates the SSL certificate and starts the node daemon.
164 @type master_name: str
165 @param master_name: Name of the master node
168 # Generate cluster secrets
169 GenerateClusterCrypto(True, False, False, False)
171 result = utils.RunCmd([constants.DAEMON_UTIL, "start", constants.NODED])
173 raise errors.OpExecError("Could not start the node daemon, command %s"
174 " had exitcode %s and error %s" %
175 (result.cmd, result.exit_code, result.output))
177 _WaitForNodeDaemon(master_name)
180 def _WaitForNodeDaemon(node_name):
181 """Wait for node daemon to become responsive.
184 def _CheckNodeDaemon():
185 result = rpc.RpcRunner.call_version([node_name])[node_name]
187 raise utils.RetryAgain()
190 utils.Retry(_CheckNodeDaemon, 1.0, _DAEMON_READY_TIMEOUT)
191 except utils.RetryTimeout:
192 raise errors.OpExecError("Node daemon on %s didn't answer queries within"
193 " %s seconds" % (node_name, _DAEMON_READY_TIMEOUT))
196 def _WaitForMasterDaemon():
197 """Wait for master daemon to become responsive.
200 def _CheckMasterDaemon():
203 (cluster_name, ) = cl.QueryConfigValues(["cluster_name"])
205 raise utils.RetryAgain()
207 logging.debug("Received cluster name %s from master", cluster_name)
210 utils.Retry(_CheckMasterDaemon, 1.0, _DAEMON_READY_TIMEOUT)
211 except utils.RetryTimeout:
212 raise errors.OpExecError("Master daemon didn't answer queries within"
213 " %s seconds" % _DAEMON_READY_TIMEOUT)
216 def _InitFileStorage(file_storage_dir):
217 """Initialize if needed the file storage.
219 @param file_storage_dir: the user-supplied value
220 @return: either empty string (if file storage was disabled at build
221 time) or the normalized path to the storage directory
224 file_storage_dir = os.path.normpath(file_storage_dir)
226 if not os.path.isabs(file_storage_dir):
227 raise errors.OpPrereqError("File storage directory '%s' is not an absolute"
228 " path" % file_storage_dir, errors.ECODE_INVAL)
230 if not os.path.exists(file_storage_dir):
232 os.makedirs(file_storage_dir, 0750)
234 raise errors.OpPrereqError("Cannot create file storage directory"
235 " '%s': %s" % (file_storage_dir, err),
236 errors.ECODE_ENVIRON)
238 if not os.path.isdir(file_storage_dir):
239 raise errors.OpPrereqError("The file storage directory '%s' is not"
240 " a directory." % file_storage_dir,
241 errors.ECODE_ENVIRON)
242 return file_storage_dir
245 def InitCluster(cluster_name, mac_prefix, # pylint: disable-msg=R0913
246 master_netdev, file_storage_dir, shared_file_storage_dir,
247 candidate_pool_size, secondary_ip=None, vg_name=None,
248 beparams=None, nicparams=None, ndparams=None, hvparams=None,
249 enabled_hypervisors=None, modify_etc_hosts=True,
250 modify_ssh_setup=True, maintain_node_health=False,
251 drbd_helper=None, uid_pool=None, default_iallocator=None,
252 primary_ip_version=None, prealloc_wipe_disks=False):
253 """Initialise the cluster.
255 @type candidate_pool_size: int
256 @param candidate_pool_size: master candidate pool size
259 # TODO: complete the docstring
260 if config.ConfigWriter.IsCluster():
261 raise errors.OpPrereqError("Cluster is already initialised",
264 if not enabled_hypervisors:
265 raise errors.OpPrereqError("Enabled hypervisors list must contain at"
266 " least one member", errors.ECODE_INVAL)
267 invalid_hvs = set(enabled_hypervisors) - constants.HYPER_TYPES
269 raise errors.OpPrereqError("Enabled hypervisors contains invalid"
270 " entries: %s" % invalid_hvs,
275 if primary_ip_version == constants.IP4_VERSION:
276 ipcls = netutils.IP4Address
277 elif primary_ip_version == constants.IP6_VERSION:
278 ipcls = netutils.IP6Address
280 raise errors.OpPrereqError("Invalid primary ip version: %d." %
283 hostname = netutils.GetHostname(family=ipcls.family)
284 if not ipcls.IsValid(hostname.ip):
285 raise errors.OpPrereqError("This host's IP (%s) is not a valid IPv%d"
286 " address." % (hostname.ip, primary_ip_version))
288 if ipcls.IsLoopback(hostname.ip):
289 raise errors.OpPrereqError("This host's IP (%s) resolves to a loopback"
290 " address. Please fix DNS or %s." %
291 (hostname.ip, constants.ETC_HOSTS),
292 errors.ECODE_ENVIRON)
294 if not ipcls.Own(hostname.ip):
295 raise errors.OpPrereqError("Inconsistency: this host's name resolves"
296 " to %s,\nbut this ip address does not"
297 " belong to this host" %
298 hostname.ip, errors.ECODE_ENVIRON)
300 clustername = netutils.GetHostname(name=cluster_name, family=ipcls.family)
302 if netutils.TcpPing(clustername.ip, constants.DEFAULT_NODED_PORT, timeout=5):
303 raise errors.OpPrereqError("Cluster IP already active",
304 errors.ECODE_NOTUNIQUE)
307 if primary_ip_version == constants.IP6_VERSION:
308 raise errors.OpPrereqError("When using a IPv6 primary address, a valid"
309 " IPv4 address must be given as secondary",
311 secondary_ip = hostname.ip
313 if not netutils.IP4Address.IsValid(secondary_ip):
314 raise errors.OpPrereqError("Secondary IP address (%s) has to be a valid"
315 " IPv4 address." % secondary_ip,
318 if not netutils.IP4Address.Own(secondary_ip):
319 raise errors.OpPrereqError("You gave %s as secondary IP,"
320 " but it does not belong to this host." %
321 secondary_ip, errors.ECODE_ENVIRON)
323 if vg_name is not None:
324 # Check if volume group is valid
325 vgstatus = utils.CheckVolumeGroupSize(utils.ListVolumeGroups(), vg_name,
326 constants.MIN_VG_SIZE)
328 raise errors.OpPrereqError("Error: %s\nspecify --no-lvm-storage if"
329 " you are not using lvm" % vgstatus,
332 if drbd_helper is not None:
334 curr_helper = bdev.BaseDRBD.GetUsermodeHelper()
335 except errors.BlockDeviceError, err:
336 raise errors.OpPrereqError("Error while checking drbd helper"
337 " (specify --no-drbd-storage if you are not"
338 " using drbd): %s" % str(err),
339 errors.ECODE_ENVIRON)
340 if drbd_helper != curr_helper:
341 raise errors.OpPrereqError("Error: requiring %s as drbd helper but %s"
342 " is the current helper" % (drbd_helper,
346 if constants.ENABLE_FILE_STORAGE:
347 file_storage_dir = _InitFileStorage(file_storage_dir)
349 file_storage_dir = ""
351 if constants.ENABLE_SHARED_FILE_STORAGE:
352 shared_file_storage_dir = _InitFileStorage(shared_file_storage_dir)
354 shared_file_storage_dir = ""
356 if not re.match("^[0-9a-z]{2}:[0-9a-z]{2}:[0-9a-z]{2}$", mac_prefix):
357 raise errors.OpPrereqError("Invalid mac prefix given '%s'" % mac_prefix,
360 result = utils.RunCmd(["ip", "link", "show", "dev", master_netdev])
362 raise errors.OpPrereqError("Invalid master netdev given (%s): '%s'" %
364 result.output.strip()), errors.ECODE_INVAL)
366 dirs = [(constants.RUN_GANETI_DIR, constants.RUN_DIRS_MODE)]
367 utils.EnsureDirs(dirs)
369 utils.ForceDictType(beparams, constants.BES_PARAMETER_TYPES)
370 utils.ForceDictType(nicparams, constants.NICS_PARAMETER_TYPES)
371 objects.NIC.CheckParameterSyntax(nicparams)
373 if ndparams is not None:
374 utils.ForceDictType(ndparams, constants.NDS_PARAMETER_TYPES)
376 ndparams = dict(constants.NDC_DEFAULTS)
378 # hvparams is a mapping of hypervisor->hvparams dict
379 for hv_name, hv_params in hvparams.iteritems():
380 utils.ForceDictType(hv_params, constants.HVS_PARAMETER_TYPES)
381 hv_class = hypervisor.GetHypervisor(hv_name)
382 hv_class.CheckParameterSyntax(hv_params)
384 # set up ssh config and /etc/hosts
385 sshline = utils.ReadFile(constants.SSH_HOST_RSA_PUB)
386 sshkey = sshline.split(" ")[1]
389 utils.AddHostToEtcHosts(hostname.name, hostname.ip)
394 if default_iallocator is not None:
395 alloc_script = utils.FindFile(default_iallocator,
396 constants.IALLOCATOR_SEARCH_PATH,
398 if alloc_script is None:
399 raise errors.OpPrereqError("Invalid default iallocator script '%s'"
400 " specified" % default_iallocator,
402 elif constants.HTOOLS:
403 # htools was enabled at build-time, we default to it
404 if utils.FindFile(constants.IALLOC_HAIL,
405 constants.IALLOCATOR_SEARCH_PATH,
407 default_iallocator = constants.IALLOC_HAIL
411 # init of cluster config file
412 cluster_config = objects.Cluster(
414 rsahostkeypub=sshkey,
415 highest_used_port=(constants.FIRST_DRBD_PORT - 1),
416 mac_prefix=mac_prefix,
417 volume_group_name=vg_name,
418 tcpudp_port_pool=set(),
419 master_node=hostname.name,
420 master_ip=clustername.ip,
421 master_netdev=master_netdev,
422 cluster_name=clustername.name,
423 file_storage_dir=file_storage_dir,
424 shared_file_storage_dir=shared_file_storage_dir,
425 enabled_hypervisors=enabled_hypervisors,
426 beparams={constants.PP_DEFAULT: beparams},
427 nicparams={constants.PP_DEFAULT: nicparams},
430 candidate_pool_size=candidate_pool_size,
431 modify_etc_hosts=modify_etc_hosts,
432 modify_ssh_setup=modify_ssh_setup,
436 maintain_node_health=maintain_node_health,
437 drbd_usermode_helper=drbd_helper,
438 default_iallocator=default_iallocator,
439 primary_ip_family=ipcls.family,
440 prealloc_wipe_disks=prealloc_wipe_disks,
442 master_node_config = objects.Node(name=hostname.name,
443 primary_ip=hostname.ip,
444 secondary_ip=secondary_ip,
446 master_candidate=True,
447 offline=False, drained=False,
448 ctime=now, mtime=now,
450 InitConfig(constants.CONFIG_VERSION, cluster_config, master_node_config)
451 cfg = config.ConfigWriter(offline=True)
452 ssh.WriteKnownHostsFile(cfg, constants.SSH_KNOWN_HOSTS_FILE)
453 cfg.Update(cfg.GetClusterInfo(), logging.error)
454 backend.WriteSsconfFiles(cfg.GetSsconfValues())
456 # set up the inter-node password and certificate
457 _InitGanetiServerSetup(hostname.name)
459 logging.debug("Starting daemons")
460 result = utils.RunCmd([constants.DAEMON_UTIL, "start-all"])
462 raise errors.OpExecError("Could not start daemons, command %s"
463 " had exitcode %s and error %s" %
464 (result.cmd, result.exit_code, result.output))
466 _WaitForMasterDaemon()
469 def InitConfig(version, cluster_config, master_node_config,
470 cfg_file=constants.CLUSTER_CONF_FILE):
471 """Create the initial cluster configuration.
473 It will contain the current node, which will also be the master
474 node, and no instances.
477 @param version: configuration version
478 @type cluster_config: L{objects.Cluster}
479 @param cluster_config: cluster configuration
480 @type master_node_config: L{objects.Node}
481 @param master_node_config: master node configuration
482 @type cfg_file: string
483 @param cfg_file: configuration file path
486 uuid_generator = config.TemporaryReservationManager()
487 cluster_config.uuid = uuid_generator.Generate([], utils.NewUUID,
489 master_node_config.uuid = uuid_generator.Generate([], utils.NewUUID,
492 master_node_config.name: master_node_config,
494 default_nodegroup = objects.NodeGroup(
495 uuid=uuid_generator.Generate([], utils.NewUUID, _INITCONF_ECID),
496 name=constants.INITIAL_NODE_GROUP_NAME,
497 members=[master_node_config.name],
500 default_nodegroup.uuid: default_nodegroup,
503 config_data = objects.ConfigData(version=version,
504 cluster=cluster_config,
505 nodegroups=nodegroups,
509 ctime=now, mtime=now)
510 utils.WriteFile(cfg_file,
511 data=serializer.Dump(config_data.ToDict()),
515 def FinalizeClusterDestroy(master):
516 """Execute the last steps of cluster destroy
518 This function shuts down all the daemons, completing the destroy
519 begun in cmdlib.LUDestroyOpcode.
522 cfg = config.ConfigWriter()
523 modify_ssh_setup = cfg.GetClusterInfo().modify_ssh_setup
524 result = rpc.RpcRunner.call_node_stop_master(master, True)
525 msg = result.fail_msg
527 logging.warning("Could not disable the master role: %s", msg)
528 result = rpc.RpcRunner.call_node_leave_cluster(master, modify_ssh_setup)
529 msg = result.fail_msg
531 logging.warning("Could not shutdown the node daemon and cleanup"
532 " the node: %s", msg)
535 def SetupNodeDaemon(cluster_name, node, ssh_key_check):
536 """Add a node to the cluster.
538 This function must be called before the actual opcode, and will ssh
539 to the remote node, copy the needed files, and start ganeti-noded,
540 allowing the master to do the rest via normal rpc calls.
542 @param cluster_name: the cluster name
543 @param node: the name of the new node
544 @param ssh_key_check: whether to do a strict key check
547 family = ssconf.SimpleStore().GetPrimaryIPFamily()
548 sshrunner = ssh.SshRunner(cluster_name,
549 ipv6=(family == netutils.IP6Address.family))
551 bind_address = constants.IP4_ADDRESS_ANY
552 if family == netutils.IP6Address.family:
553 bind_address = constants.IP6_ADDRESS_ANY
555 # set up inter-node password and certificate and restarts the node daemon
556 # and then connect with ssh to set password and start ganeti-noded
557 # note that all the below variables are sanitized at this point,
558 # either by being constants or by the checks above
559 sshrunner.CopyFileToNode(node, constants.NODED_CERT_FILE)
560 sshrunner.CopyFileToNode(node, constants.RAPI_CERT_FILE)
561 sshrunner.CopyFileToNode(node, constants.CONFD_HMAC_KEY)
562 mycommand = ("%s stop-all; %s start %s -b %s" %
563 (constants.DAEMON_UTIL, constants.DAEMON_UTIL, constants.NODED,
564 utils.ShellQuote(bind_address)))
566 result = sshrunner.Run(node, 'root', mycommand, batch=False,
567 ask_key=ssh_key_check,
568 use_cluster_key=True,
569 strict_host_check=ssh_key_check)
571 raise errors.OpExecError("Remote command on node %s, error: %s,"
573 (node, result.fail_reason, result.output))
575 _WaitForNodeDaemon(node)
578 def MasterFailover(no_voting=False):
579 """Failover the master node.
581 This checks that we are not already the master, and will cause the
582 current master to cease being master, and the non-master to become
585 @type no_voting: boolean
586 @param no_voting: force the operation without remote nodes agreement
590 sstore = ssconf.SimpleStore()
592 old_master, new_master = ssconf.GetMasterAndMyself(sstore)
593 node_list = sstore.GetNodeList()
594 mc_list = sstore.GetMasterCandidates()
596 if old_master == new_master:
597 raise errors.OpPrereqError("This commands must be run on the node"
598 " where you want the new master to be."
599 " %s is already the master" %
600 old_master, errors.ECODE_INVAL)
602 if new_master not in mc_list:
603 mc_no_master = [name for name in mc_list if name != old_master]
604 raise errors.OpPrereqError("This node is not among the nodes marked"
605 " as master candidates. Only these nodes"
606 " can become masters. Current list of"
607 " master candidates is:\n"
608 "%s" % ('\n'.join(mc_no_master)),
612 vote_list = GatherMasterVotes(node_list)
615 voted_master = vote_list[0][0]
616 if voted_master is None:
617 raise errors.OpPrereqError("Cluster is inconsistent, most nodes did"
618 " not respond.", errors.ECODE_ENVIRON)
619 elif voted_master != old_master:
620 raise errors.OpPrereqError("I have a wrong configuration, I believe"
621 " the master is %s but the other nodes"
622 " voted %s. Please resync the configuration"
624 (old_master, voted_master),
630 logging.info("Setting master to %s, old master: %s", new_master, old_master)
633 # instantiate a real config writer, as we now know we have the
635 cfg = config.ConfigWriter(accept_foreign=True)
637 cluster_info = cfg.GetClusterInfo()
638 cluster_info.master_node = new_master
639 # this will also regenerate the ssconf files, since we updated the
641 cfg.Update(cluster_info, logging.error)
642 except errors.ConfigurationError, err:
643 logging.error("Error while trying to set the new master: %s",
647 # if cfg.Update worked, then it means the old master daemon won't be
648 # able now to write its own config file (we rely on locking in both
649 # backend.UploadFile() and ConfigWriter._Write(); hence the next
650 # step is to kill the old master
652 logging.info("Stopping the master daemon on node %s", old_master)
654 result = rpc.RpcRunner.call_node_stop_master(old_master, True)
655 msg = result.fail_msg
657 logging.error("Could not disable the master role on the old master"
658 " %s, please disable manually: %s", old_master, msg)
660 logging.info("Checking master IP non-reachability...")
662 master_ip = sstore.GetMasterIP()
664 # Here we have a phase where no master should be running
666 if netutils.TcpPing(master_ip, constants.DEFAULT_NODED_PORT):
667 raise utils.RetryAgain()
670 utils.Retry(_check_ip, (1, 1.5, 5), total_timeout)
671 except utils.RetryTimeout:
672 logging.warning("The master IP is still reachable after %s seconds,"
673 " continuing but activating the master on the current"
674 " node will probably fail", total_timeout)
676 if jstore.CheckDrainFlag():
677 logging.info("Undraining job queue")
678 jstore.SetDrainFlag(False)
680 logging.info("Starting the master daemons on the new master")
682 result = rpc.RpcRunner.call_node_start_master(new_master, True, no_voting)
683 msg = result.fail_msg
685 logging.error("Could not start the master role on the new master"
686 " %s, please check: %s", new_master, msg)
689 logging.info("Master failed over from %s to %s", old_master, new_master)
694 """Returns the current master node.
696 This is a separate function in bootstrap since it's needed by
697 gnt-cluster, and instead of importing directly ssconf, it's better
698 to abstract it in bootstrap, where we do use ssconf in other
702 sstore = ssconf.SimpleStore()
704 old_master, _ = ssconf.GetMasterAndMyself(sstore)
709 def GatherMasterVotes(node_list):
710 """Check the agreement on who is the master.
712 This function will return a list of (node, number of votes), ordered
713 by the number of votes. Errors will be denoted by the key 'None'.
715 Note that the sum of votes is the number of nodes this machine
716 knows, whereas the number of entries in the list could be different
717 (if some nodes vote for another master).
719 We remove ourselves from the list since we know that (bugs aside)
720 since we use the same source for configuration information for both
721 backend and boostrap, we'll always vote for ourselves.
723 @type node_list: list
724 @param node_list: the list of nodes to query for master info; the current
725 node will be removed if it is in the list
727 @return: list of (node, votes)
730 myself = netutils.Hostname.GetSysName()
732 node_list.remove(myself)
736 # no nodes left (eventually after removing myself)
738 results = rpc.RpcRunner.call_master_info(node_list)
739 if not isinstance(results, dict):
740 # this should not happen (unless internal error in rpc)
741 logging.critical("Can't complete rpc call, aborting master startup")
742 return [(None, len(node_list))]
750 logging.warning("Error contacting node %s: %s", node, msg)
752 # for now we accept both length 3 and 4 (data[3] is primary ip version)
753 elif not isinstance(data, (tuple, list)) or len(data) < 3:
754 logging.warning("Invalid data received from node %s: %s", node, data)
757 if None not in votes:
761 master_node = data[2]
762 if master_node not in votes:
763 votes[master_node] = 0
764 votes[master_node] += 1
766 vote_list = [v for v in votes.items()]
767 # sort first on number of votes then on name, since we want None
768 # sorted later if we have the half of the nodes not responding, and
769 # half voting all for the same master
770 vote_list.sort(key=lambda x: (x[1], x[0]), reverse=True)