Synnefo » snf-ganeti

#

#

# Copyright (C) 2006, 2007, 2008 Google Inc.

#

# This program is free software; you can redistribute it and/or modify

# it under the terms of the GNU General Public License as published by

# the Free Software Foundation; either version 2 of the License, or

# (at your option) any later version.

#

# This program is distributed in the hope that it will be useful, but

# WITHOUT ANY WARRANTY; without even the implied warranty of

# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU

# General Public License for more details.

#

# You should have received a copy of the GNU General Public License

# along with this program; if not, write to the Free Software

# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA

# 02110-1301, USA.

"""Functions to bootstrap a new cluster.

"""

import os

import os.path

import sha

import re

import logging

from ganeti import rpc

from ganeti import ssh

from ganeti import utils

from ganeti import errors

from ganeti import config

from ganeti import constants

from ganeti import objects

from ganeti import ssconf

from ganeti.rpc import RpcRunner

def _InitSSHSetup(node):

  """Setup the SSH configuration for the cluster.

  This generates a dsa keypair for root, adds the pub key to the

  permitted hosts and adds the hostkey to its own known hosts.

  Args:

    node: the name of this host as a fqdn

  """

  priv_key, pub_key, auth_keys = ssh.GetUserFiles(constants.GANETI_RUNAS)

  for name in priv_key, pub_key:

    if os.path.exists(name):

      utils.CreateBackup(name)

    utils.RemoveFile(name)

  result = utils.RunCmd(["ssh-keygen", "-t", "dsa",

                         "-f", priv_key,

                         "-q", "-N", ""])

  if result.failed:

    raise errors.OpExecError("Could not generate ssh keypair, error %s" %

                             result.output)

  f = open(pub_key, 'r')

  try:

    utils.AddAuthorizedKey(auth_keys, f.read(8192))

  finally:

    f.close()

def _InitGanetiServerSetup():

  """Setup the necessary configuration for the initial node daemon.

  This creates the nodepass file containing the shared password for

  the cluster and also generates the SSL certificate.

  """

  # Create pseudo random password

  randpass = utils.GenerateSecret()

  # and write it into the config file

  utils.WriteFile(constants.CLUSTER_PASSWORD_FILE,

                  data="%s\n" % randpass, mode=0400)

  result = utils.RunCmd(["openssl", "req", "-new", "-newkey", "rsa:1024",

                         "-days", str(365*5), "-nodes", "-x509",

                         "-keyout", constants.SSL_CERT_FILE,

                         "-out", constants.SSL_CERT_FILE, "-batch"])

  if result.failed:

    raise errors.OpExecError("could not generate server ssl cert, command"

                             " %s had exitcode %s and error message %s" %

                             (result.cmd, result.exit_code, result.output))

  os.chmod(constants.SSL_CERT_FILE, 0400)

  result = utils.RunCmd([constants.NODE_INITD_SCRIPT, "restart"])

  if result.failed:

    raise errors.OpExecError("Could not start the node daemon, command %s"

                             " had exitcode %s and error %s" %

                             (result.cmd, result.exit_code, result.output))

def InitCluster(cluster_name, hypervisor_type, mac_prefix, def_bridge,

                master_netdev, file_storage_dir,

                secondary_ip=None,

                vg_name=None, beparams=None, hvparams=None,

                enabled_hypervisors=None):

  """Initialise the cluster.

  """

  if config.ConfigWriter.IsCluster():

    raise errors.OpPrereqError("Cluster is already initialised")

  if hypervisor_type == constants.HT_XEN_HVM:

    if not os.path.exists(constants.VNC_PASSWORD_FILE):

      raise errors.OpPrereqError("Please prepare the cluster VNC"

                                 "password file %s" %

                                 constants.VNC_PASSWORD_FILE)

  hostname = utils.HostInfo()

  if hostname.ip.startswith("127."):

    raise errors.OpPrereqError("This host's IP resolves to the private"

                               " range (%s). Please fix DNS or %s." %

                               (hostname.ip, constants.ETC_HOSTS))

  if not utils.OwnIpAddress(hostname.ip):

    raise errors.OpPrereqError("Inconsistency: this host's name resolves"

                               " to %s,\nbut this ip address does not"

                               " belong to this host."

                               " Aborting." % hostname.ip)

  clustername = utils.HostInfo(cluster_name)

  if utils.TcpPing(clustername.ip, constants.DEFAULT_NODED_PORT,

                   timeout=5):

    raise errors.OpPrereqError("Cluster IP already active. Aborting.")

  if secondary_ip:

    if not utils.IsValidIP(secondary_ip):

      raise errors.OpPrereqError("Invalid secondary ip given")

    if (secondary_ip != hostname.ip and

        not utils.OwnIpAddress(secondary_ip)):

      raise errors.OpPrereqError("You gave %s as secondary IP,"

                                 " but it does not belong to this host." %

                                 secondary_ip)

  else:

    secondary_ip = hostname.ip

  if vg_name is not None:

    # Check if volume group is valid

    vgstatus = utils.CheckVolumeGroupSize(utils.ListVolumeGroups(), vg_name,

                                          constants.MIN_VG_SIZE)

    if vgstatus:

      raise errors.OpPrereqError("Error: %s\nspecify --no-lvm-storage if"

                                 " you are not using lvm" % vgstatus)

  file_storage_dir = os.path.normpath(file_storage_dir)

  if not os.path.isabs(file_storage_dir):

    raise errors.OpPrereqError("The file storage directory you passed is"

                               " not an absolute path.")

  if not os.path.exists(file_storage_dir):

    try:

      os.makedirs(file_storage_dir, 0750)

    except OSError, err:

      raise errors.OpPrereqError("Cannot create file storage directory"

                                 " '%s': %s" %

                                 (file_storage_dir, err))

  if not os.path.isdir(file_storage_dir):

    raise errors.OpPrereqError("The file storage directory '%s' is not"

                               " a directory." % file_storage_dir)

  if not re.match("^[0-9a-z]{2}:[0-9a-z]{2}:[0-9a-z]{2}$", mac_prefix):

    raise errors.OpPrereqError("Invalid mac prefix given '%s'" % mac_prefix)

  if hypervisor_type not in constants.HYPER_TYPES:

    raise errors.OpPrereqError("Invalid hypervisor type given '%s'" %

                               hypervisor_type)

  result = utils.RunCmd(["ip", "link", "show", "dev", master_netdev])

  if result.failed:

    raise errors.OpPrereqError("Invalid master netdev given (%s): '%s'" %

                               (master_netdev,

                                result.output.strip()))

  if not (os.path.isfile(constants.NODE_INITD_SCRIPT) and

          os.access(constants.NODE_INITD_SCRIPT, os.X_OK)):

    raise errors.OpPrereqError("Init.d script '%s' missing or not"

                               " executable." % constants.NODE_INITD_SCRIPT)

  # set up the inter-node password and certificate

  _InitGanetiServerSetup()

  # set up ssh config and /etc/hosts

  f = open(constants.SSH_HOST_RSA_PUB, 'r')

  try:

    sshline = f.read()

  finally:

    f.close()

  sshkey = sshline.split(" ")[1]

  utils.AddHostToEtcHosts(hostname.name)

  _InitSSHSetup(hostname.name)

  # init of cluster config file

  cluster_config = objects.Cluster(

    serial_no=1,

    rsahostkeypub=sshkey,

    highest_used_port=(constants.FIRST_DRBD_PORT - 1),

    mac_prefix=mac_prefix,

    volume_group_name=vg_name,

    default_bridge=def_bridge,

    tcpudp_port_pool=set(),

    hypervisor=hypervisor_type,

    master_node=hostname.name,

    master_ip=clustername.ip,

    master_netdev=master_netdev,

    cluster_name=clustername.name,

    file_storage_dir=file_storage_dir,

    enabled_hypervisors=enabled_hypervisors,

    beparams={constants.BEGR_DEFAULT: beparams},

    hvparams=hvparams,

    )

  master_node_config = objects.Node(name=hostname.name,

                                    primary_ip=hostname.ip,

                                    secondary_ip=secondary_ip)

  cfg = InitConfig(constants.CONFIG_VERSION,

                   cluster_config, master_node_config)

  ssh.WriteKnownHostsFile(cfg, constants.SSH_KNOWN_HOSTS_FILE)

  # start the master ip

  # TODO: Review rpc call from bootstrap

  RpcRunner.call_node_start_master(hostname.name, True)

def InitConfig(version, cluster_config, master_node_config,

               cfg_file=constants.CLUSTER_CONF_FILE):

  """Create the initial cluster configuration.

  It will contain the current node, which will also be the master

  node, and no instances.

  @type version: int

  @param version: Configuration version

  @type cluster_config: objects.Cluster

  @param cluster_config: Cluster configuration

  @type master_node_config: objects.Node

  @param master_node_config: Master node configuration

  @type file_name: string

  @param file_name: Configuration file path

  @rtype: ssconf.SimpleConfigWriter

  @returns: Initialized config instance

  """

  nodes = {

    master_node_config.name: master_node_config,

    }

  config_data = objects.ConfigData(version=version,

                                   cluster=cluster_config,

                                   nodes=nodes,

                                   instances={},

                                   serial_no=1)

  cfg = ssconf.SimpleConfigWriter.FromDict(config_data.ToDict(), cfg_file)

  cfg.Save()

  return cfg

def FinalizeClusterDestroy(master):

  """Execute the last steps of cluster destroy

  This function shuts down all the daemons, completing the destroy

  begun in cmdlib.LUDestroyOpcode.

  """

  if not RpcRunner.call_node_stop_master(master, True):

    logging.warning("Could not disable the master role")

  if not RpcRunner.call_node_leave_cluster(master):

    logging.warning("Could not shutdown the node daemon and cleanup the node")

def SetupNodeDaemon(node, ssh_key_check):

  """Add a node to the cluster.

  This function must be called before the actual opcode, and will ssh

  to the remote node, copy the needed files, and start ganeti-noded,

  allowing the master to do the rest via normal rpc calls.

  Args:

    node: fully qualified domain name for the new node

  """

  cfg = ssconf.SimpleConfigReader()

  sshrunner = ssh.SshRunner(cfg.GetClusterName())

  gntpass = utils.GetNodeDaemonPassword()

  if not re.match('^[a-zA-Z0-9.]{1,64}$', gntpass):

    raise errors.OpExecError("ganeti password corruption detected")

  f = open(constants.SSL_CERT_FILE)

  try:

    gntpem = f.read(8192)

  finally:

    f.close()

  # in the base64 pem encoding, neither '!' nor '.' are valid chars,

  # so we use this to detect an invalid certificate; as long as the

  # cert doesn't contain this, the here-document will be correctly

  # parsed by the shell sequence below

  if re.search('^!EOF\.', gntpem, re.MULTILINE):

    raise errors.OpExecError("invalid PEM encoding in the SSL certificate")

  if not gntpem.endswith("\n"):

    raise errors.OpExecError("PEM must end with newline")

  # set up inter-node password and certificate and restarts the node daemon

  # and then connect with ssh to set password and start ganeti-noded

  # note that all the below variables are sanitized at this point,

  # either by being constants or by the checks above

  mycommand = ("umask 077 && "

               "echo '%s' > '%s' && "

               "cat > '%s' << '!EOF.' && \n"

               "%s!EOF.\n%s restart" %

               (gntpass, constants.CLUSTER_PASSWORD_FILE,

                constants.SSL_CERT_FILE, gntpem,

                constants.NODE_INITD_SCRIPT))

  result = sshrunner.Run(node, 'root', mycommand, batch=False,

                         ask_key=ssh_key_check,

                         use_cluster_key=False,

                         strict_host_check=ssh_key_check)

  if result.failed:

    raise errors.OpExecError("Remote command on node %s, error: %s,"

                             " output: %s" %

                             (node, result.fail_reason, result.output))

  return 0

def MasterFailover():

  """Failover the master node.

  This checks that we are not already the master, and will cause the

  current master to cease being master, and the non-master to become

  new master.

  """

  cfg = ssconf.SimpleConfigWriter()

  new_master = utils.HostInfo().name

  old_master = cfg.GetMasterNode()

  node_list = cfg.GetNodeList()

  if old_master == new_master:

    raise errors.OpPrereqError("This commands must be run on the node"

                               " where you want the new master to be."

                               " %s is already the master" %

                               old_master)

  vote_list = GatherMasterVotes(node_list)

  if vote_list:

    voted_master = vote_list[0][0]

    if voted_master is None:

      raise errors.OpPrereqError("Cluster is inconsistent, most nodes did not"

                                 " respond.")

    elif voted_master != old_master:

      raise errors.OpPrereqError("I have wrong configuration, I believe the"

                                 " master is %s but the other nodes voted for"

                                 " %s. Please resync the configuration of"

                                 " this node." % (old_master, voted_master))

  # end checks

  rcode = 0

  logging.info("Setting master to %s, old master: %s", new_master, old_master)

  if not RpcRunner.call_node_stop_master(old_master, True):

    logging.error("Could not disable the master role on the old master"

                 " %s, please disable manually", old_master)

  cfg.SetMasterNode(new_master)

  cfg.Save()

  # Here we have a phase where no master should be running

  if not RpcRunner.call_upload_file(cfg.GetNodeList(),

                                    constants.CLUSTER_CONF_FILE):

    logging.error("Could not distribute the new configuration"

                  " to the other nodes, please check.")

  if not RpcRunner.call_node_start_master(new_master, True):

    logging.error("Could not start the master role on the new master"

                  " %s, please check", new_master)

    rcode = 1

  return rcode

def GatherMasterVotes(node_list):

  """Check the agreement on who is the master.

  This function will return a list of (node, number of votes), ordered

  by the number of votes. Errors will be denoted by the key 'None'.

  Note that the sum of votes is the number of nodes this machine

  knows, whereas the number of entries in the list could be different

  (if some nodes vote for another master).

  We remove ourselves from the list since we know that (bugs aside)

  since we use the same source for configuration information for both

  backend and boostrap, we'll always vote for ourselves.

  @type node_list: list

  @param node_list: the list of nodes to query for master info; the current

      node wil be removed if it is in the list

  @rtype: list

  @return: list of (node, votes)

  """

  myself = utils.HostInfo().name

  try:

    node_list.remove(myself)

  except ValueError:

    pass

  if not node_list:

    # no nodes left (eventually after removing myself)

    return []

  results = rpc.RpcRunner.call_master_info(node_list)

  if not isinstance(results, dict):

    # this should not happen (unless internal error in rpc)

    logging.critical("Can't complete rpc call, aborting master startup")

    return [(None, len(node_list))]

  positive = negative = 0

  other_masters = {}

  votes = {}

  for node in results:

    if not isinstance(results[node], (tuple, list)) or len(results[node]) < 3:

      # here the rpc layer should have already logged errors

      if None not in votes:

        votes[None] = 0

      votes[None] += 1

      continue

    master_node = results[node][2]

    if master_node not in votes:

      votes[master_node] = 0

    votes[master_node] += 1

  vote_list = [v for v in votes.items()]

  # sort first on number of votes then on name, since we want None

  # sorted later if we have the half of the nodes not responding, and

  # half voting all for the same master

  vote_list.sort(key=lambda x: (x[1], x[0]), reverse=True)

  return vote_list