Source code for cfme.utils.appliance

import json
import logging
import socket
import traceback
from copy import copy
from datetime import datetime
from tempfile import NamedTemporaryFile
from time import sleep, time

import attr
import dateutil.parser
import fauxfactory
import os
import re
import requests
import sentaku
import six
import warnings
import yaml
from cached_property import cached_property
from debtcollector import removals
from manageiq_client.api import APIException, ManageIQClient as VanillaMiqApi
from six.moves.urllib.parse import urlparse
from wrapanapi import VmState
from wrapanapi.exceptions import VMInstanceNotFound
from werkzeug.local import LocalStack, LocalProxy

from cfme.fixtures import ui_coverage
from cfme.fixtures.pytest_store import store
from cfme.utils import clear_property_cache
from cfme.utils import conf, ssh, ports
from cfme.utils.blockers import BZ
from cfme.utils.datafile import load_data_file
from cfme.utils.log import logger, create_sublogger, logger_wrap
from import net_check
from cfme.utils.path import data_path, patches_path, scripts_path, conf_path
from cfme.utils.ssh import SSHTail
from cfme.utils.version import Version, get_stream, pick
from cfme.utils.wait import wait_for, TimedOutError
from .db import ApplianceDB
from import ViaREST
from .implementations.ssui import ViaSSUI
from .implementations.ui import ViaUI
from .services import SystemdService

RUNNING_UNDER_SPROUT = os.environ.get("RUNNING_UNDER_SPROUT", "false") != "false"
# EMS types recognized by IP or credentials
    "InfraManager", "ContainerManager", "Openstack::CloudManager"
RECOGNIZED_BY_CREDS = ["CloudManager", "Nuage::NetworkManager"]

# A helper for the IDs
SEQ_FACT = 1e12

def _current_miqqe_version():
    """Parses MiqQE JS patch version from the patch file

    Returns: Version as int
    with patches_path.join('miq_application.js.diff').open("r") as f:
        match ="MiqQE_version = (\d+);",, flags=0)
    version = int(
    return version

current_miqqe_version = _current_miqqe_version()

[docs]class MiqApi(VanillaMiqApi):
[docs] def get_entity_by_href(self, href): """Parses the collections""" parsed = urlparse(href) # TODO: Check the netloc, scheme path = [step for step in parsed.path.split('/') if step] # Drop the /api path = path[1:] collection = getattr(self.collections, path.pop(0)) entity = collection(int(path.pop(0))) if path: raise ValueError('Subcollections not supported! ({})'.format(parsed.path)) return entity
[docs]class ApplianceException(Exception): pass
[docs]class ApplianceConsole(object): """ApplianceConsole is used for navigating and running appliance_console commands against an appliance.""" def __init__(self, appliance): self.appliance = appliance
[docs] def timezone_check(self, timezone): channel = self.appliance.ssh_client.invoke_shell() channel.settimeout(20) channel.send("ap") result = '' try: while True: result += channel.recv(1) if ("{}".format(timezone[0])) in result: break except socket.timeout: pass logger.debug(result)
[docs] def run_commands(self, commands, autoreturn=True, timeout=10, channel=None): if not channel: channel = self.appliance.ssh_client.invoke_shell() self.commands = commands for command in commands: if isinstance(command, six.string_types): command_string, timeout = command, timeout else: command_string, timeout = command channel.settimeout(timeout) if autoreturn: command_string = (command_string + '\n') channel.send("{}".format(command_string)) result = '' try: while True: result += channel.recv(1) if 'Press any key to continue' in result: break except socket.timeout: pass logger.debug(result)
[docs]class ApplianceConsoleCli(object): def __init__(self, appliance): self.appliance = appliance def _run(self, appliance_console_cli_command): return self.appliance.ssh_client.run_command( "appliance_console_cli {}".format(appliance_console_cli_command))
[docs] def set_hostname(self, hostname): return self._run("--host {host}".format(host=hostname))
[docs] def configure_appliance_external_join(self, dbhostname, username, password, dbname, fetch_key, sshlogin, sshpass): self._run("--hostname {dbhostname} --username {username} --password {password}" " --dbname {dbname} --verbose --fetch-key {fetch_key} --sshlogin {sshlogin}" " --sshpassword {sshpass}".format(dbhostname=dbhostname, username=username, password=password, dbname=dbname, fetch_key=fetch_key, sshlogin=sshlogin, sshpass=sshpass))
[docs] def configure_appliance_external_create(self, region, dbhostname, username, password, dbname, fetch_key, sshlogin, sshpass): self._run("--region {region} --hostname {dbhostname} --username {username}" " --password {password} --dbname {dbname} --verbose --fetch-key {fetch_key}" " --sshlogin {sshlogin} --sshpassword {sshpass}".format( region=region, dbhostname=dbhostname, username=username, password=password, dbname=dbname, fetch_key=fetch_key, sshlogin=sshlogin, sshpass=sshpass))
[docs] def configure_appliance_internal(self, region, dbhostname, username, password, dbname, dbdisk): self._run("--region {region} --internal --hostname {dbhostname} --username {username}" " --password {password} --dbname {dbname} --verbose --dbdisk {dbdisk}".format( region=region, dbhostname=dbhostname, username=username, password=password, dbname=dbname, dbdisk=dbdisk))
[docs] def configure_appliance_internal_fetch_key(self, region, dbhostname, username, password, dbname, dbdisk, fetch_key, sshlogin, sshpass): self._run("--region {region} --internal --hostname {dbhostname} --username {username}" " --password {password} --dbname {dbname} --verbose --dbdisk {dbdisk} --fetch-key" " {fetch_key} --sshlogin {sshlogin} --sshpassword {sshpass}".format( region=region, dbhostname=dbhostname, username=username, password=password, dbname=dbname, dbdisk=dbdisk, fetch_key=fetch_key, sshlogin=sshlogin, sshpass=sshpass))
[docs] def configure_appliance_dedicated_db(self, username, password, dbname, dbdisk): self._run("--internal --username {username} --password {password}" " --dbname {dbname} --verbose --dbdisk {dbdisk} --key --standalone".format( username=username, password=password, dbname=dbname, dbdisk=dbdisk))
[docs] def configure_ipa(self, ipaserver, ipaprincipal, ipapassword, ipadomain=None, iparealm=None): cmd_result = self._run( '--ipaserver {s} --ipaprincipal {u} --ipapassword {p} {d} {r}' .format(s=ipaserver, u=ipaprincipal, p=ipapassword, d='--ipadomain {}'.format(ipadomain) if ipadomain else '', r='--iparealm {}'.format(iparealm) if iparealm else '')) logger.debug('IPA configuration output: %s', str(cmd_result)) assert cmd_result.success assert 'ipa-client-install exit code: 1' not in cmd_result.output self.appliance.sssd.wait_for_running() assert self.appliance.ssh_client.run_command("cat /etc/ipa/default.conf " "| grep 'enable_ra = True'")
[docs] def configure_appliance_dedicated_ha_primary( self, username, password, reptype, primhost, node, dbname): self._run("--username {username} --password {password} --replication {reptype}" " --primary-host {primhost} --cluster-node-number {node} --auto-failover --verbose" " --dbname {dbname}".format( username=username, password=password, reptype=reptype, primhost=primhost, node=node, dbname=dbname))
[docs] def configure_appliance_dedicated_ha_standby( self, username, password, reptype, primhost, standhost, node, dbname, dbdisk): self._run("--internal --username {username} --password {password} --replication {reptype}" " --primary-host {primhost} --standby-host {standhost} --cluster-node-number {node}" " --auto-failover --dbname {dbname} --verbose --dbdisk {dbdisk}" " --standalone".format(username=username, password=password, reptype=reptype, primhost=primhost, standhost=standhost, node=node, dbname=dbname, dbdisk=dbdisk))
[docs] def uninstall_ipa_client(self): assert self._run("--uninstall-ipa") assert not self.appliance.ssh_client.run_command("cat /etc/ipa/default.conf")
[docs]class IPAppliance(object): """IPAppliance represents an already provisioned cfme appliance whos provider is unknown but who has an IP address. This has a lot of core functionality that Appliance uses, since it knows both the provider, vm_name and can there for derive the IP address. Args: hostname: The IP address or host name of the provider ui_protocol: The protocol used in the URL ui_port: The port where the UI runs. browser_steal: If True then then current browser is killed and the new appliance is used to generate a new session. container: If the appliance is running as a container or as a pod, specifies its name. project: openshift's project where the appliance is deployed openshift_creds: If the appliance runs as a project on openshift, provides credentials for the openshift host so the framework can interact with the project. db_host: If the database is located somewhere else than on the appliance itself, specify the host here. db_port: Database port. ssh_port: SSH port. """ _nav_steps = {} evmserverd = SystemdService.declare(unit_name='evmserverd') httpd = SystemdService.declare(unit_name='httpd') sssd = SystemdService.declare(unit_name='sssd') db = ApplianceDB.declare() CONFIG_MAPPING = { 'hostname': 'hostname', 'ui_protocol': 'ui_protocol', 'ui_port': 'ui_port', 'browser_steal': 'browser_steal', 'container': 'container', 'pod': 'container', 'openshift_creds': 'openshift_creds', 'is_dev': 'is_dev', 'db_host': 'db_host', 'db_port': 'db_port', 'ssh_port': 'ssh_port', 'project': 'project', } CONFIG_NONGLOBAL = {'hostname'} PROTOCOL_PORT_MAPPING = {'http': 80, 'https': 443} CONF_FILES = { 'upstream_templates': '/var/www/miq/system/TEMPLATE', 'downstream_templates': '/opt/rh/cfme-appliance/TEMPLATE', 'pam_httpd_auth': '/etc/pam.d/httpd-auth', 'httpd_remote_user': '/etc/httpd/conf.d/manageiq-remote-user.conf', 'httpd_ext_auth': '/etc/httpd/conf.d/manageiq-external-auth.conf', 'openldap': '/etc/openldap/ldap.conf', 'sssd': '/etc/sssd/sssd.conf' } @property def as_json(self): """Dumps the arguments that can create this appliance as a JSON. None values are ignored.""" return json.dumps({ k: getattr(self, k) for k in set(self.CONFIG_MAPPING.values())}) @classmethod
[docs] def from_json(cls, json_string): return cls(**json.loads(json_string))
def __init__( self, hostname, ui_protocol='https', ui_port=None, browser_steal=False, project=None, container=None, openshift_creds=None, db_host=None, db_port=None, ssh_port=None, is_dev=False ): if not isinstance(hostname, six.string_types): raise TypeError('Appliance\'s hostname must be a string!') self.hostname = hostname if ui_protocol not in self.PROTOCOL_PORT_MAPPING: raise TypeError( 'Wrong protocol {!r} passed, expected {!r}'.format( ui_protocol, list(self.PROTOCOL_PORT_MAPPING.keys()))) self.ui_protocol = ui_protocol self.ui_port = ui_port or self.PROTOCOL_PORT_MAPPING[ui_protocol] self.ssh_port = ssh_port or ports.SSH self.db_port = db_port or ports.DB self.db_host = db_host self.browser = ViaUI(owner=self) self.ssui = ViaSSUI(owner=self) self.rest_context = ViaREST(owner=self) self.rest_context.strict_calls = False self.context = MiqImplementationContext.from_instances( [self.browser, self.ssui, self.rest_context]) from cfme.modeling.base import EntityCollections self.collections = EntityCollections.for_appliance(self) self.browser_steal = browser_steal self.container = container self.project = project self.openshift_creds = openshift_creds or {} self.is_dev = is_dev self._user = None self.appliance_console = ApplianceConsole(self) self.appliance_console_cli = ApplianceConsoleCli(self) if self.openshift_creds: self.is_pod = True else: self.is_pod = False
[docs] def unregister(self): """ unregisters appliance from RHSM/SAT6 """ self.ssh_client.run_command('subscription-manager remove --all') self.ssh_client.run_command('subscription-manager unregister') self.ssh_client.run_command('subscription-manager clean') self.ssh_client.run_command('mv -f /etc/rhsm/rhsm.conf.kat-backup /etc/rhsm/rhsm.conf') self.ssh_client.run_command('rpm -qa | grep katello-ca-consumer | xargs rpm -e')
[docs] def is_registration_complete(self, used_repo_or_channel): """ Checks if an appliance has the correct repos enabled with RHSM or SAT6 """ result = self.ssh_client.run_command('yum repolist enabled') # Check that the specified (or default) repo (can be multiple, separated by a space) # is enabled and that there are packages available for repo in used_repo_or_channel.split(' '): if (repo not in result.output) or (not'repolist: [^0]', result.output)): return False return True
@property def default_zone(self): return @property def server(self): return self.collections.servers.get_master() @property def user(self): from cfme.base.credential import Credential if self._user is None: # Admin by default username = conf.credentials['default']['username'] password = conf.credentials['default']['password'] '%r.user was set to None before, therefore generating an admin user: %s/%s', self, username, password) cred = Credential(principal=username, secret=password) user = self.collections.users.instantiate( credential=cred, name='Administrator' ) self._user = user return self._user @user.setter def user(self, user_object): if user_object is None:'%r.user set to None, will be set to admin on next access', self) self._user = user_object @property def appliance(self): return self def __repr__(self): # TODO: Put something better here. This solves the purpose temporarily. return '{}.from_json({!r})'.format(type(self).__name__, self.as_json)
[docs] def __call__(self, **kwargs): """Syntactic sugar for overriding certain instance variables for context managers. Currently possible variables are: * `browser_steal` """ self.browser_steal = kwargs.get("browser_steal", self.browser_steal) return self
[docs] def __enter__(self): """ This method will replace the current appliance in the store """ stack.push(self) return self
def _screenshot_capture_at_context_leave(self, exc_type, exc_val, exc_tb): try: from cfme.fixtures.artifactor_plugin import fire_art_hook from pytest import config from cfme.fixtures.pytest_store import store except ImportError:'Not inside pytest run, ignoring') return if ( exc_type is not None and not RUNNING_UNDER_SPROUT): from cfme.utils.browser import take_screenshot"Before we pop this appliance, a screenshot and a traceback will be taken.") ss, ss_error = take_screenshot() full_tb = "".join(traceback.format_tb(exc_tb)) short_tb = "{}: {}".format(exc_type.__name__, str(exc_val)) full_tb = "{}\n{}".format(full_tb, short_tb) g_id = "appliance-cm-screenshot-{}".format(fauxfactory.gen_alpha(length=6)) fire_art_hook( config, 'filedump', slaveid=store.slaveid, description="Appliance CM error traceback", contents=full_tb, file_type="traceback", display_type="danger", display_glyph="align-justify", group_id=g_id) if ss: fire_art_hook( config, 'filedump', slaveid=store.slaveid, description="Appliance CM error screenshot", file_type="screenshot", mode="wb", contents_base64=True, contents=ss, display_glyph="camera", group_id=g_id) if ss_error: fire_art_hook( config, 'filedump', slaveid=store.slaveid, description="Appliance CM error screenshot failure", mode="w", contents_base64=False, contents=ss_error, display_type="danger", group_id=g_id) elif exc_type is not None:"Error happened but we are not inside a test run so no screenshot now.") def __exit__(self, exc_type, exc_val, exc_tb): try: self._screenshot_capture_at_context_leave(exc_type, exc_val, exc_tb) except Exception: # repr is used in order to avoid having the appliance object in the log record logger.exception("taking a screenshot for %s failed", repr(self)) finally: assert stack.pop() is self, 'appliance stack inconsistent' def __eq__(self, other): return isinstance(other, IPAppliance) and self.hostname == other.hostname def __ne__(self, other): return not self.__eq__(other) def __hash__(self): return hash(self.hostname) @cached_property def rest_logger(self): return create_sublogger('rest-api') # Configuration methods @logger_wrap("Configure IPAppliance: {}") def configure(self, log_callback=None, **kwargs): """Configures appliance - database setup, rename, ntp sync Utility method to make things easier. Note: db_address, name_to_set are not used currently. Args: db_address: Address of external database if set, internal database if ``None`` (default ``None``) name_to_set: Name to set the appliance name to if not ``None`` (default ``None``) region: Number to assign to region (default ``0``) fix_ntp_clock: Fixes appliance time if ``True`` (default ``True``) loosen_pgssl: Loosens postgres connections if ``True`` (default ``True``) key_address: Fetch encryption key from this address if set, generate a new key if ``None`` (default ``None``) on_openstack: If appliance is running on Openstack provider (default ``False``) on_gce: If appliance is running on GCE provider (default ``False``) """ log_callback("Configuring appliance {}".format(self.hostname)) loosen_pgssl = kwargs.pop('loosen_pgssl', True) fix_ntp_clock = kwargs.pop('fix_ntp_clock', True) region = kwargs.pop('region', 0) key_address = kwargs.pop('key_address', None) db_address = kwargs.pop('db_address', None) on_gce = kwargs.pop('on_gce', False) on_openstack = kwargs.pop('on_openstack', False) with self as ipapp: ipapp.wait_for_ssh() # Debugging - ifcfg-eth0 overwritten by unknown process # Rules are permanent and will be reloade after machine reboot self.ssh_client.run_command( "cp -pr /etc/sysconfig/network-scripts/ifcfg-eth0 /var/tmp", ensure_host=True) self.ssh_client.run_command( "echo '-w /etc/sysconfig/network-scripts/ifcfg-eth0 -p wa' >> " "/etc/audit/rules.d/audit.rules", ensure_host=True) self.ssh_client.run_command("systemctl daemon-reload", ensure_host=True) self.ssh_client.run_command("service auditd restart", ensure_host=True) ipapp.wait_for_ssh() self.deploy_merkyl(start=True, log_callback=log_callback) if fix_ntp_clock and not self.is_pod: self.fix_ntp_clock(log_callback=log_callback) # TODO: Handle external DB setup # This is workaround for appliances to use only one disk for the VMDB # If they have been provisioned with a second disk in the infra, # 'self.unpartitioned_disks' should exist and therefore this won't run. if self.is_downstream and not self.unpartitioned_disks: self.db.create_db_lvm() if on_openstack: self.set_resolvable_hostname(log_callback=log_callback) self.db.setup(region=region, key_address=key_address, db_address=db_address, is_pod=self.is_pod) if on_gce: # evm serverd does not auto start on GCE instance.. self.start_evm_service(log_callback=log_callback) self.wait_for_evm_service(timeout=1200, log_callback=log_callback) # Some conditionally ran items require the evm service be # restarted: restart_evm = False if loosen_pgssl: self.db.loosen_pgssl() restart_evm = True if self.version >= '5.8': self.configure_vm_console_cert(log_callback=log_callback) restart_evm = True if restart_evm: self.restart_evm_service(log_callback=log_callback) self.wait_for_web_ui(timeout=1800, log_callback=log_callback)
[docs] def configure_gce(self, log_callback=None): # Force use of IPAppliance's configure method return IPAppliance.configure(self, on_gce=True)
[docs] def seal_for_templatizing(self): """Prepares the VM to be "generalized" for saving as a template.""" with self.ssh_client as ssh_client: # Seals the VM in order to work when spawned again. ssh_client.run_command("rm -rf /etc/ssh/ssh_host_*", ensure_host=True) if ssh_client.run_command( "grep '^HOSTNAME' /etc/sysconfig/network", ensure_host=True).success: # Replace it ssh_client.run_command( "sed -i -r -e 's/^HOSTNAME=.*$/HOSTNAME=localhost.localdomain/' " "/etc/sysconfig/network", ensure_host=True) else: # Set it ssh_client.run_command( "echo HOSTNAME=localhost.localdomain >> /etc/sysconfig/network", ensure_host=True) # clear any set hostname from /etc/hosts self.remove_resolvable_hostname() ssh_client.run_command( "sed -i -r -e '/^HWADDR/d' /etc/sysconfig/network-scripts/ifcfg-eth0", ensure_host=True) ssh_client.run_command( "sed -i -r -e '/^UUID/d' /etc/sysconfig/network-scripts/ifcfg-eth0", ensure_host=True) ssh_client.run_command("rm -f /etc/udev/rules.d/70-*", ensure_host=True) # Fix SELinux things ssh_client.run_command("restorecon -R /etc/sysconfig/network-scripts", ensure_host=True) ssh_client.run_command("restorecon /etc/sysconfig/network", ensure_host=True) # Stop the evmserverd and move the logs somewhere ssh_client.run_command("systemctl stop evmserverd", ensure_host=True) ssh_client.run_command("mkdir -p /var/www/miq/vmdb/log/preconfigure-logs", ensure_host=True) ssh_client.run_command( "mv /var/www/miq/vmdb/log/*.log /var/www/miq/vmdb/log/preconfigure-logs/", ensure_host=True) ssh_client.run_command( "mv /var/www/miq/vmdb/log/*.gz /var/www/miq/vmdb/log/preconfigure-logs/", ensure_host=True) # Reduce swapping, because it can do nasty things to our providers ssh_client.run_command('echo "vm.swappiness = 1" >> /etc/sysctl.conf', ensure_host=True)
def _encrypt_string(self, string): try: # Let's not log passwords logging.disable(logging.CRITICAL) result = self.ssh_client.run_rails_command( "\"puts MiqPassword.encrypt('{}')\"".format(string)) return result.output.strip() finally: logging.disable(logging.NOTSET) @property def managed_provider_names(self): """Returns a list of names for all providers configured on the appliance Note: Unlike ``managed_known_providers``, this will also return names of providers that were not recognized, but are present. """ known_ems_list = [] for ems in self.rest_api.collections.providers: if not any( p_type in ems['type'] for p_type in RECOGNIZED_BY_IP + RECOGNIZED_BY_CREDS): continue known_ems_list.append(ems['name']) return known_ems_list @property def managed_known_providers(self): """Returns a set of provider crud objects of known providers managed by this appliance Note: Recognized by name only. """ from cfme.utils.providers import list_providers prov_cruds = list_providers(use_global_filters=False) found_cruds = set() unrecognized_ems_names = set() for ems_name in self.managed_provider_names: for prov in prov_cruds: # Name check is authoritative and the only proper way to recognize a known provider if ems_name == found_cruds.add(prov) break else: unrecognized_ems_names.add(ems_name) if unrecognized_ems_names: self.log.warning( "Unrecognized managed providers: {}".format(', '.join(unrecognized_ems_names))) return list(found_cruds) @classmethod
[docs] def from_url(cls, url, **kwargs): """Create an appliance instance from a URL. Supported format using a simple regexp expression: ``(https?://)?hostname_or_ip(:port)?/?`` Args: url: URL to be parsed from **kwargs: For setting and overriding the params parsed from the URL Returns: A :py:class:`IPAppliance` instance. """ if not isinstance(url, six.string_types): raise TypeError('url for .from_url must be a string') parsed = urlparse(url) new_kwargs = {} if parsed.netloc: host_part = parsed.netloc elif parsed.path and not parsed.netloc: # If you only pass the hostname (+ port possibly) without scheme or anything else host_part = parsed.path else: raise ValueError('Unsupported url specification: {}'.format(url)) if ':' in host_part: hostname, port = host_part.rsplit(':', 1) port = int(port) else: hostname = host_part if parsed.scheme: port = cls.PROTOCOL_PORT_MAPPING[parsed.scheme] else: port = None new_kwargs['hostname'] = hostname if port is not None: new_kwargs['ui_port'] = port if parsed.scheme: new_kwargs['ui_protocol'] = parsed.scheme new_kwargs.update(kwargs) return cls(**new_kwargs)
[docs] def new_rest_api_instance( self, entry_point=None, auth=None, logger="default", verify_ssl=False): """Returns new REST API instance.""" return MiqApi( entry_point=entry_point or self.url_path('/api'), auth=auth or (conf.credentials["default"]["username"], conf.credentials["default"]["password"]), logger=self.rest_logger if logger == "default" else logger, verify_ssl=verify_ssl)
@cached_property def rest_api(self): return self.new_rest_api_instance() @cached_property def miqqe_version(self): """Returns version of applied JS patch or None if not present""" result = self.ssh_client.run_command('grep "[0-9]\\+" /var/www/miq/vmdb/.miqqe_version') if result.success: return int(result.output) return None @property def url(self): """Returns a proper URL of the appliance. If the ports do not correspond the protocols' default port numbers, then the ports are explicitly specified as well. """ show_port = self.PROTOCOL_PORT_MAPPING[self.ui_protocol] != self.ui_port if show_port: return '{}://{}:{}/'.format(self.ui_protocol, self.hostname, self.ui_port) else: return '{}://{}/'.format(self.ui_protocol, self.hostname)
[docs] def url_path(self, path): """generates URL with an additional path. Useful for generating REST or SSUI URLs.""" return '{}/{}'.format(self.url.rstrip('/'), path.lstrip('/'))
@property def disks_and_partitions(self): """Returns list of all disks and partitions""" disks_and_partitions = self.ssh_client.run_command( "ls -1 /dev/ | egrep '^[sv]d[a-z][0-9]?'").output.strip() disks_and_partitions = re.split(r'\s+', disks_and_partitions) return sorted('/dev/{}'.format(disk) for disk in disks_and_partitions) @property def disks(self): """Returns list of disks only, excludes their partitions""" disk_regexp = re.compile('^/dev/[sv]d[a-z]$') return [ disk for disk in self.disks_and_partitions if disk_regexp.match(disk) ] @property def unpartitioned_disks(self): """Returns list of any disks that have no partitions""" partition_regexp = re.compile('^/dev/[sv]d[a-z][0-9]$') unpartitioned_disks = set() for disk in self.disks: add = True for dp in self.disks_and_partitions: if dp.startswith(disk) and partition_regexp.match(dp) is not None: add = False if add: unpartitioned_disks.add(disk) return sorted(disk for disk in unpartitioned_disks) @cached_property def product_name(self): try: return self.rest_api.product_info['name'] except (AttributeError, KeyError, IOError): self.log.exception( 'appliance.product_name could not be retrieved from REST, falling back') try: # TODO: Review this section. Does not work unconfigured # # We need to print to a file here because the deprecation warnings make it hard # # to get robust output and they do not seem to go to stderr # result = self.ssh_client.run_rails_command( # '"\'/tmp/product_name.txt\', \'w\') ' # '{|f| f.write(I18n.t(\'\')) }"') # result = self.ssh_client.run_command('cat /tmp/product_name.txt') # return result.output res = self.ssh_client.run_command('cat /etc/redhat-release') if res.failed: raise RuntimeError('Unable to retrieve /etc/redhat-release') version_string = res.output.strip() if 'CentOS' in version_string: return 'ManageIQ' else: return 'CFME' except Exception: logger.exception( "Couldn't fetch the product name from appliance, using ManageIQ as default") return 'ManageIQ' @cached_property def is_downstream(self): return self.product_name == 'CFME' @cached_property def version(self): try: return Version(self.rest_api.server_info['version']) except (AttributeError, KeyError, IOError, APIException): self.log.exception('appliance.version could not be retrieved from REST, falling back') return self.ssh_client.vmdb_version @cached_property def build(self): if not self.is_downstream: return 'master' try: return self.rest_api.server_info['build'] except (AttributeError, KeyError, IOError): self.log.exception(' could not be retrieved from REST, falling back') res = self.ssh_client.run_command('cat /var/www/miq/vmdb/BUILD') if res.failed: raise RuntimeError('Unable to retrieve appliance VMDB version') return res.output.strip("\n") @cached_property def os_version(self): # Currently parses the os version out of redhat release file to allow for # rhel and centos appliances res = self.ssh_client.run_command( r"cat /etc/redhat-release | sed 's/.* release \(.*\) (.*/\1/' #)") if res.failed: raise RuntimeError('Unable to retrieve appliance OS version') return Version(res.output) @cached_property def log(self): return create_sublogger(self.hostname) @cached_property def coverage(self): return ui_coverage.CoverageManager(self)
[docs] def ssh_client_with_privatekey(self): with open(conf_path.join('appliance_private_key').strpath, 'w') as key: key.write(conf.credentials['ssh']['private_key']) connect_kwargs = { 'hostname': self.hostname, 'username': conf.credentials['ssh']['ssh-user'], 'key_filename': conf_path.join('appliance_private_key').strpath, } ssh_client = ssh.SSHClient(**connect_kwargs) # FIXME: properly store ssh clients we made store.ssh_clients_to_close.append(ssh_client) return ssh_client
@cached_property def ssh_client(self): """Creates an ssh client connected to this appliance Returns: A configured :py:class:``utils.ssh.SSHClient`` instance. Usage: with appliance.ssh_client as ssh: result = ssh.run_command('...') Note: The credentials default to those found under ``ssh`` key in ``credentials.yaml``. """ if not self.is_ssh_running: raise Exception('SSH is unavailable') # IPAppliance.ssh_client only connects to its address if self.openshift_creds: connect_kwargs = { 'hostname': self.openshift_creds['hostname'], 'username': self.openshift_creds['ssh']['username'], 'password': self.openshift_creds['ssh']['password'], 'oc_username': self.openshift_creds['username'], 'oc_password': self.openshift_creds['password'], 'container': self.container, 'is_pod': self.is_pod, 'port': self.ssh_port, 'project': self.project } else: connect_kwargs = { 'hostname': self.hostname, 'username': conf.credentials['ssh']['username'], 'password': conf.credentials['ssh']['password'], 'container': self.container, 'is_pod': self.is_pod, 'port': self.ssh_port, } if self.is_dev: connect_kwargs.update({'is_dev': True}) ssh_client = ssh.SSHClient(**connect_kwargs) try: ssh_client.get_transport().is_active()'default appliance ssh credentials are valid') except Exception as e: if self.is_dev: raise Exception('SSH access on a dev alliance, (unsupported)') logger.error(e) logger.error('default appliance ssh credentials failed, trying establish ssh connection' ' using ssh private key') ssh_client = self.ssh_client_with_privatekey() # FIXME: properly store ssh clients we made store.ssh_clients_to_close.append(ssh_client) return ssh_client @property def swap(self): """Retrieves the value of swap for the appliance. Might raise an exception if SSH fails. Return: An integer value of swap in the VM in megabytes. If ``None`` is returned, it means it was not possible to parse the command output. Raises: :py:class:`paramiko.ssh_exception.SSHException` or :py:class:`socket.error` """ try: server = self.rest_api.get_entity_by_href(self.rest_api.server_info['server_href']) return server.system_swap_used / 1024 / 1024 except (AttributeError, KeyError, IOError): self.log.exception('appliance.swap could not be retrieved from REST, falling back') value = self.ssh_client.run_command( 'free -m | tr -s " " " " | cut -f 3 -d " " | tail -n 1', reraise=True, timeout=15) try: value = int(value.output.strip()) except (TypeError, ValueError): value = None return value
[docs] def event_listener(self): """Returns an instance of the event listening class pointed to this appliance.""" # There is no REST API for event streams on versions < 5.9 if self.version <= '5.9': from cfme.utils.events_db import DbEventListener return DbEventListener(self) else: from import RestEventListener return RestEventListener(self)
[docs] def diagnose_evm_failure(self): """Go through various EVM processes, trying to figure out what fails Returns: A string describing the error, or None if no errors occurred. This is intended to be run after an appliance is configured but failed for some reason, such as in the template tester. """'Diagnosing EVM failures, this can take a while...') if not self.hostname: return 'appliance has no IP Address; provisioning failed or networking is broken''Checking appliance SSH Connection') if not self.is_ssh_running: return 'SSH is not running on the appliance' # Now for the DB'Checking appliance database') if not # postgres isn't running, try to start it cmd = 'systemctl restart {}-postgresql'.format(self.db.postgres_version) result = self.db.ssh_client.run_command(cmd) if result.failed: return 'postgres failed to start:\n{}'.format(result.output) else: return 'postgres was not running for unknown reasons' if not self.db.has_database: return 'vmdb_production database does not exist' if not self.db.has_tables: return 'vmdb_production has no tables' # try to start EVM'Checking appliance evmserverd service') try: self.restart_evm_service() except ApplianceException as ex: return 'evmserverd failed to start:\n{}'.format(ex.args[0])
# This should be pretty comprehensive, but we might add some net_checks for # 3000, 4000, and 80 at this point, and waiting a reasonable amount of time # before exploding if any of them don't appear in time after evm restarts. @logger_wrap("Fix NTP Clock: {}") def fix_ntp_clock(self, log_callback=None): """Fixes appliance time using ntpdate on appliance""" log_callback('Fixing appliance clock') client = self.ssh_client # checking whether chrony is installed check_cmd = 'yum list installed chrony' if client.run_command(check_cmd).failed: raise ApplianceException("Chrony isn't installed") # # checking whether it is enabled and enable it is_enabled_cmd = 'systemctl is-enabled chronyd' if client.run_command(is_enabled_cmd).failed: logger.debug("chrony will start on system startup") client.run_command('systemctl enable chronyd') client.run_command('systemctl daemon-reload') # Retrieve time servers from yamls server_template = 'server {srv} iburst' time_servers = set() try: logger.debug('obtaining clock servers from config file') clock_servers = conf.cfme_data.get('clock_servers') for clock_server in clock_servers: time_servers.add(server_template.format(srv=clock_server)) except TypeError: msg = 'No clock servers configured in cfme_data.yaml' log_callback(msg) raise ApplianceException(msg) filename = '/etc/chrony.conf' chrony_conf = set(client.run_command("cat {f}".format(f=filename)).output.strip() .split('\n')) modified_chrony_conf = chrony_conf.union(time_servers) if modified_chrony_conf != chrony_conf: modified_chrony_conf = "\n".join(list(modified_chrony_conf)) client.run_command('echo "{txt}" > {f}'.format(txt=modified_chrony_conf, f=filename))"chrony's config file updated") conf_file_updated = True else:"chrony's config file hasn't been changed") conf_file_updated = False if conf_file_updated or client.run_command('systemctl status chronyd').failed: logger.debug('restarting chronyd') client.run_command('systemctl restart chronyd') # check that chrony is running correctly now result = client.run_command('chronyc tracking') if result.success:'chronyc is running correctly') else: raise ApplianceException("chrony doesn't work. " "Error message: {e}".format(e=result.output)) @property def is_miqqe_patch_candidate(self): return self.version < "5.6.3" @property def miqqe_patch_applied(self): return self.miqqe_version == current_miqqe_version @logger_wrap("Patch appliance with MiqQE js: {}") def patch_with_miqqe(self, log_callback=None): # (local_path, remote_path, md5/None) trio autofocus_patch = pick({ '5.5': 'autofocus.js.diff', '5.7': 'autofocus_57.js.diff' }) patch_args = ( (str(patches_path.join('miq_application.js.diff')), '/var/www/miq/vmdb/app/assets/javascripts/miq_application.js', None), (str(patches_path.join(autofocus_patch)), '/var/www/miq/vmdb/app/assets/javascripts/directives/autofocus.js', None), ) for local_path, remote_path, md5 in patch_args: self.ssh_client.patch_file(local_path, remote_path, md5) self.precompile_assets() self.restart_evm_service()"Waiting for Web UI to start") wait_for( func=self.is_web_ui_running, message='appliance.is_web_ui_running', delay=20, timeout=300)"Web UI is up and running") self.ssh_client.run_command( "echo '{}' > /var/www/miq/vmdb/.miqqe_version".format(current_miqqe_version)) # Invalidate cached version del self.miqqe_version @logger_wrap("Work around missing Gem file: {}") def workaround_missing_gemfile(self, log_callback=None): """Fix Gemfile issue. Early 5.4 builds have issues with Gemfile not present (BUG 1191496). This circumvents the issue with pointing the env variable that Bundler uses to get the Gemfile to the Gemfile in vmdb which *should* be correct. When this issue is resolved, this method will do nothing. """ client = self.ssh_client result = client.run_command("ls /opt/rh/cfme-gemset") if result.failed: return # Not needed log_callback('Fixing Gemfile issue') # Check if the error is there result = client.run_rails_command("puts 1") if result.success: return # All OK! client.run_command('echo "export BUNDLE_GEMFILE=/var/www/miq/vmdb/Gemfile" >> /etc/bashrc') # To be 100% sure self.reboot(wait_for_web_ui=False, log_callback=log_callback) @logger_wrap("Precompile assets: {}") def precompile_assets(self, log_callback=None): """Precompile the static assets (images, css, etc) on an appliance """ log_callback('Precompiling assets') client = self.ssh_client store.terminalreporter.write_line('Precompiling assets') store.terminalreporter.write_line( 'THIS IS NOT STUCK. Just wait until it\'s done, it will be only done once', red=True) store.terminalreporter.write_line('Phase 1 of 2: rake assets:clobber') result = client.run_rake_command("assets:clobber") if result.failed: msg = 'Appliance {} failed to nuke old assets'.format(self.hostname) log_callback(msg) raise ApplianceException(msg) store.terminalreporter.write_line('Phase 2 of 2: rake assets:precompile') result = client.run_rake_command("assets:precompile") if result.failed: msg = 'Appliance {} failed to precompile assets'.format(self.hostname) log_callback(msg) raise ApplianceException(msg) store.terminalreporter.write_line('Asset precompilation done') return result.rc @logger_wrap("Clone automate domain: {}") def clone_domain(self, source="ManageIQ", dest="Default", log_callback=None): """Clones Automate domain Args: src: Source domain name. dst: Destination domain name. """ client = self.ssh_client # Make sure the database is ready log_callback('Waiting for database') self.db.wait_for() # Make sure the working dir exists client.run_command('mkdir -p /tmp/{}'.format(source)) export_opts = 'DOMAIN={} EXPORT_DIR=/tmp/{} PREVIEW=false OVERWRITE=true'.format(source, source) export_cmd = 'evm:automate:export {}'.format(export_opts) log_callback('Exporting domain ({}) ...'.format(export_cmd)) result = client.run_rake_command(export_cmd) if result.failed: msg = 'Failed to export {} domain'.format(source) log_callback(msg) raise ApplianceException(msg) ro_fix_cmd = ("sed -i 's/system: true/system: false/g' " "/tmp/{}/{}/__domain__.yaml".format(source, source)) result = client.run_command(ro_fix_cmd) if result.failed: msg = 'Setting {} domain to read/write failed'.format(dest) log_callback(msg) raise ApplianceException(msg) import_opts = 'DOMAIN={} IMPORT_DIR=/tmp/{} PREVIEW=false'.format(source, source) import_opts += ' OVERWRITE=true IMPORT_AS={} ENABLED=true'.format(dest) import_cmd = 'evm:automate:import {}'.format(import_opts) log_callback('Importing domain ({}) ...'.format(import_cmd)) result = client.run_rake_command(import_cmd) if result.failed: msg = 'Failed to import {} domain'.format(dest) log_callback(msg) raise ApplianceException(msg) return result.rc, result.output @logger_wrap("Deploying Merkyl: {}") def deploy_merkyl(self, start=False, log_callback=None): """Deploys the Merkyl log relay service to the appliance""" client = self.ssh_client client.run_command('mkdir -p /root/merkyl') for filename in ['', 'merkyl.tpl', ('', ''), 'allowed.files']: try: src, dest = filename except (TypeError, ValueError): # object is not iterable or too many values to unpack src = dest = filename log_callback('Sending {} to appliance'.format(src)) client.put_file(data_path.join( 'bundles', 'merkyl', src).strpath, os.path.join('/root/merkyl', dest)) client.put_file(data_path.join( 'bundles', 'merkyl', 'merkyl').strpath, os.path.join('/etc/init.d/merkyl')) client.run_command('chmod 775 /etc/init.d/merkyl') client.run_command( '/bin/bash -c \'if ! [[ $(iptables -L -n | grep "state NEW tcp dpt:8192") ]]; then ' 'iptables -I INPUT 6 -m state --state NEW -m tcp -p tcp --dport 8192 -j ACCEPT; fi\'') if start: log_callback("Starting ...") client.run_command('systemctl restart merkyl') log_callback("Setting it to start after reboot") client.run_command("chkconfig merkyl on")
[docs] def get_repofile_list(self): """Returns list of repofiles present at the appliance. Ignores certain files, like redhat.repo. """ repofiles = self.ssh_client.run_command('ls /etc/yum.repos.d').output.strip().split('\n') return [f for f in repofiles if f not in {"redhat.repo"} and f.endswith(".repo")]
[docs] def read_repos(self): """Reads repofiles so it gives you mapping of id and url.""" result = {} name_regexp = re.compile(r"^\[update-([^\]]+)\]") baseurl_regexp = re.compile(r"baseurl\s*=\s*([^\s]+)") for repofile in self.get_repofile_list(): result = self.ssh_client.run_command("cat /etc/yum.repos.d/{}".format(repofile)) if result.failed: # Something happened meanwhile? continue out = result.output.strip() name_match = if name_match is None: continue baseurl_match = if baseurl_match is None: continue result[name_match.groups()[0]] = baseurl_match.groups()[0] return result
# Regexp that looks for product type and version in the update URL product_url_regexp = re.compile( r"/((?:[A-Z]+|CloudForms|rhel|RHEL_Guest))(?:-|/|/server/)(\d+[^/]*)/")
[docs] def find_product_repos(self): """Returns a dictionary of products, where the keys are names of product (repos) and values are dictionaries where keys are the versions and values the names of the repositories. """ products = {} for repo_name, repo_url in self.read_repos().items(): match = if match is None: continue product, ver = match.groups() if product not in products: products[product] = {} products[product][ver] = repo_name return products
[docs] def write_repofile(self, repo_id, repo_url, **kwargs): """Wrapper around writing a repofile. You can specify conf options in kwargs.""" if "gpgcheck" not in kwargs: kwargs["gpgcheck"] = 0 if "enabled" not in kwargs: kwargs["enabled"] = 1 filename = "/etc/yum.repos.d/{}.repo".format(repo_id)"Writing a new repofile %s %s", repo_id, repo_url) self.ssh_client.run_command('echo "[update-{}]" > {}'.format(repo_id, filename)) self.ssh_client.run_command('echo "name=update-url-{}" >> {}'.format(repo_id, filename)) self.ssh_client.run_command('echo "baseurl={}" >> {}'.format(repo_url, filename)) for k, v in kwargs.items(): self.ssh_client.run_command('echo "{}={}" >> {}'.format(k, v, filename)) return repo_id
[docs] def add_product_repo(self, repo_url, **kwargs): """This method ensures that when we add a new repo URL, there will be no other version of such product present in the yum.repos.d. You can specify conf options in kwargs. They will be applied only to newly created repo file. Returns: The repo id. """ match = if match is None: raise ValueError( "The URL {} does not contain information about product and version.".format( repo_url)) for repo_id, url in self.read_repos().items(): if url == repo_url: # It is already there, so just enable it self.enable_disable_repo(repo_id, True) return repo_id product, ver = match.groups() repos = self.find_product_repos() if product in repos: for v, i in repos[product].items():"Deleting %s repo with version %s (%s)", product, v, i) self.ssh_client.run_command("rm -f /etc/yum.repos.d/{}.repo".format(i)) return self.write_repofile(fauxfactory.gen_alpha(), repo_url, **kwargs)
[docs] def enable_disable_repo(self, repo_id, enable):"%s repository %s", "Enabling" if enable else "Disabling", repo_id) return self.ssh_client.run_command( "sed -i 's/^enabled=./enabled={}/' /etc/yum.repos.d/{}.repo".format( 1 if enable else 0, repo_id)).success
@logger_wrap("Update RHEL: {}") def update_rhel(self, *urls, **kwargs): """Update RHEL on appliance Will pull URLs from the 'updates_urls' environment variable (whitespace-separated URLs), or cfme_data. If the env var is not set, URLs will be pulled from cfme_data. If the env var is set, it is the only source for update URLs. Generic rhel update URLs cfme_data.get('basic_info', {})['rhel_updates_urls'] (yaml list) On downstream builds, an additional RH SCL updates url can be inserted at cfme_data.get('basic_info', {})['rhscl_updates_urls']. If the ``skip_broken`` kwarg is passed, and evaluated as True, broken packages will be ignored in the yum update. """ urls = list(urls) log_callback = kwargs.pop("log_callback") skip_broken = kwargs.pop("skip_broken", False) reboot = kwargs.pop("reboot", True) streaming = kwargs.pop("streaming", False) cleanup = kwargs.pop('cleanup', False) log_callback('updating appliance') if not urls: basic_info = conf.cfme_data.get('basic_info', {}) if os.environ.get('updates_urls'): # try to pull URLs from env if var is non-empty urls.extend(os.environ['update_urls'].split()) else: # fall back to cfme_data updates_url = basic_info.get('rhel7_updates_url') if updates_url: urls.append(updates_url) if streaming: client = self.ssh_client(stream_output=True) else: client = self.ssh_client if cleanup: client.run_command( "cd /etc/yum.repos.d && find . -not -name 'redhat.repo' " "-not -name 'rhel-source.repo' -not -name . -exec rm {} \\;") for url in urls: self.add_product_repo(url) # update log_callback('Running rhel updates on appliance') # clean yum beforehand to clear metadata from earlier update repos, if any try: skip = '--skip-broken' if skip_broken else '' result = client.run_command('yum update -y --nogpgcheck {}'.format(skip), timeout=3600) except socket.timeout: msg = 'SSH timed out while updating appliance, exiting' log_callback(msg) # failure to update is fatal, kill this process raise KeyboardInterrupt(msg) self.log.error(result.output) if result.failed: self.log.error('appliance update failed') msg = 'Appliance {} failed to update RHEL, error in logs'.format(self.hostname) log_callback(msg) raise ApplianceException(msg) if reboot: self.reboot(wait_for_web_ui=False, log_callback=log_callback) return result
[docs] def utc_time(self): client = self.ssh_client result = client.run_command('date --iso-8601=seconds -u') if result.success: return dateutil.parser.parse(result.output) else: raise Exception("Couldn't get datetime: {}".format(result.output))
def _check_appliance_ui_wait_fn(self): # Get the URL, don't verify ssl cert try: response = requests.get(self.url, timeout=15, verify=False) if response.status_code == 200:"Appliance online") return True else: self.log.debug('Appliance online, status code %s', response.status_code) except requests.exceptions.Timeout: self.log.debug('Appliance offline, connection timed out') except ValueError: # requests exposes invalid URLs as ValueErrors, which is excellent raise except Exception as ex: self.log.debug('Appliance online, but connection failed: %s', str(ex)) return False
[docs] def is_web_ui_running(self, unsure=False): """Triple checks if web UI is up and running Args: unsure: Variable to return when not sure if web UI is running or not (default ``False``) """ num_of_tries = 3 was_running_count = 0 for try_num in range(num_of_tries): if self._check_appliance_ui_wait_fn(): was_running_count += 1 sleep(3) if was_running_count == 0: return False elif was_running_count == num_of_tries: return True else: return unsure
def _evm_service_command(self, command, log_callback, expected_exit_code=None): """Runs given systemctl command against the ``evmserverd`` service Args: command: Command to run, e.g. "start" expected_exit_code: If the exit codes don't match, ApplianceException is raised """ log_callback("Running command '{}' against the evmserverd service".format(command)) with self.ssh_client as ssh: result = ssh.run_command('systemctl {} evmserverd'.format(command)) if expected_exit_code is not None and result.rc != expected_exit_code: msg = ('Failed to {} evmserverd on {}\nError: {}' .format(command, self.hostname, result.output)) log_callback(msg) raise ApplianceException(msg) return result.rc @logger_wrap("Status of EVM service: {}") def is_evm_service_running(self, log_callback=None): """Checks the ``evmserverd`` service status on this appliance """ return self._evm_service_command("status", log_callback=log_callback) == 0 @logger_wrap("Start EVM Service: {}") def start_evm_service(self, log_callback=None): """Starts the ``evmserverd`` service on this appliance """ self._evm_service_command('start', expected_exit_code=0, log_callback=log_callback) @logger_wrap("Stop EVM Service: {}") def stop_evm_service(self, log_callback=None): """Stops the ``evmserverd`` service on this appliance """ self._evm_service_command('stop', expected_exit_code=0, log_callback=log_callback) @logger_wrap("Restart EVM Service: {}") def restart_evm_service(self, rude=False, log_callback=None): """Restarts the ``evmserverd`` service on this appliance """ store.terminalreporter.write_line('evmserverd is being restarted, be patient please') with self.ssh_client as ssh: if rude: self.evmserverd.stop() log_callback('Waiting for evm service to stop') try: wait_for( self.is_evm_service_running, num_sec=120, fail_condition=True, delay=10, message='evm service to stop') except TimedOutError: # Don't care if it's still running pass log_callback('killing any remaining processes and restarting postgres') ssh.run_command( 'killall -9 ruby; systemctl restart {}-postgresql' .format(self.db.postgres_version)) log_callback('Waiting for database to be available') wait_for( lambda: self.db.is_online, num_sec=90, delay=10, fail_condition=False, message="database to be available") self.evmserverd.start() else: self.evmserverd.restart() @logger_wrap("Waiting for EVM service: {}") def wait_for_evm_service(self, timeout=900, log_callback=None): """Waits for the evemserverd service to be running Args: timeout: Number of seconds to wait until timeout (default ``900``) """ log_callback('Waiting for evmserverd to be running') result, wait = wait_for(self.is_evm_service_running, num_sec=timeout, fail_condition=False, delay=10) return result @logger_wrap("Rebooting Appliance: {}") def reboot(self, wait_for_web_ui=True, log_callback=None): log_callback('Rebooting appliance') client = self.ssh_client old_uptime = client.uptime() client.run_command('reboot') wait_for(lambda: client.uptime() < old_uptime, handle_exception=True, num_sec=600, message='appliance to reboot', delay=10) if wait_for_web_ui: self.wait_for_web_ui() @logger_wrap("Waiting for web_ui: {}") def wait_for_web_ui(self, timeout=900, running=True, log_callback=None): """Waits for the web UI to be running / to not be running Args: timeout: Number of seconds to wait until timeout (default ``600``) running: Specifies if we wait for web UI to start or stop (default ``True``) ``True`` == start, ``False`` == stop """ prefix = "" if running else "dis" (log_callback or'Waiting for web UI to ' + prefix + 'appear') result, wait = wait_for(self._check_appliance_ui_wait_fn, num_sec=timeout, fail_condition=not running, delay=10) return result @logger_wrap("Install VDDK: {}") def install_vddk(self, force=False, vddk_url=None, log_callback=None): """Install the vddk on a appliance""" def log_raise(exception_class, message): log_callback(message) raise exception_class(message) if vddk_url is None: # fallback to VDDK 5.5 vddk_url = conf.cfme_data.get("basic_info", {}).get("vddk_url", {}).get("v5_5") if vddk_url is None: raise Exception("vddk_url not specified!") with self.ssh_client as client: is_already_installed = False if client.run_command('test -d /usr/lib/vmware-vix-disklib/lib64').success: is_already_installed = True if not is_already_installed or force: # start filename = vddk_url.split('/')[-1] # download log_callback('Downloading VDDK') result = client.run_command('curl {} -o {}'.format(vddk_url, filename)) if result.failed: log_raise(Exception, "Could not download VDDK") # install log_callback('Installing vddk') result = client.run_command( 'yum -y install {}'.format(filename)) if result.failed: log_raise( Exception, 'VDDK installation failure (rc: {})\n{}'.format(result.rc, result.output) ) # verify log_callback('Verifying vddk') result = client.run_command('ldconfig -p | grep vix') if len(result.output) < 2: log_raise( Exception, "Potential installation issue, libraries not detected\n{}" .format(result.output) ) @logger_wrap("Uninstall VDDK: {}") def uninstall_vddk(self, log_callback=None): """Uninstall the vddk from an appliance""" with self.ssh_client as client: is_installed = client.run_command('test -d /usr/lib/vmware-vix-disklib/lib64').success if is_installed: result = client.run_command('yum -y remove vmware-vix-disklib') if result.failed: log_callback('VDDK removing failure (rc: {})\n{}' .format(result.rc, result.output)) raise Exception('VDDK removing failure (rc: {})\n{}' .format(result.rc, result.output)) else: log_callback('VDDK has been successfully removed.') else: log_callback('VDDK is not installed.') @logger_wrap("Install Netapp SDK: {}") def install_netapp_sdk(self, sdk_url=None, reboot=False, log_callback=None): """Installs the Netapp SDK. Args: sdk_url: Where the SDK zip file is located? (optional) reboot: Whether to reboot the appliance afterwards? (Default False but reboot is needed) """ def log_raise(exception_class, message): log_callback(message) raise exception_class(message) if sdk_url is None: try: sdk_url = conf.cfme_data['basic_info']['netapp_sdk_url'] except KeyError: raise Exception("cfme_data.yaml/basic_info/netapp_sdk_url is not present!") filename = sdk_url.split('/')[-1] foldername = os.path.splitext(filename)[0] with self.ssh_client as ssh: log_callback('Downloading SDK from {}'.format(sdk_url)) result = ssh.run_command( 'wget {url} -O {file} > /root/unzip.out 2>&1'.format( url=sdk_url, file=filename)) if result.failed: log_raise(Exception, 'Could not download Netapp SDK: {}'.format(result.output)) log_callback('Extracting SDK ({})'.format(filename)) result = ssh.run_command( 'unzip -o -d /var/www/miq/vmdb/lib/ {}'.format(filename)) if result.failed: log_raise(Exception, 'Could not extract Netapp SDK: {}'.format(result.output)) path = '/var/www/miq/vmdb/lib/{}/lib/linux-64'.format(foldername) # Check if we haven't already added this line if ssh.run_command("grep -F '{}' /etc/default/evm".format(path)).failed: log_callback('Installing SDK ({})'.format(foldername)) result = ssh.run_command( 'echo "export LD_LIBRARY_PATH=\\$LD_LIBRARY_PATH:{}"' '>> /etc/default/evm'.format( path)) if result.failed: log_raise(Exception, 'SDK installation failure ($?={}): {}' .format(result.rc, result.output)) else: log_callback("Not needed to install, already done") log_callback('ldconfig') ssh.run_command('ldconfig') log_callback('Modifying YAML configuration') c_yaml = {'product': {'storage': True}} self.update_advanced_settings(c_yaml) # To mark that we installed netapp ssh.run_command("touch /var/www/miq/vmdb/HAS_NETAPP") if reboot: self.reboot(log_callback=log_callback) else: log_callback( 'Appliance must be restarted before the netapp functionality can be used.') clear_property_cache(self, 'is_storage_enabled') @logger_wrap('Updating appliance UUID: {}') def update_guid(self, log_callback=None): guid_gen = 'uuidgen |tee /var/www/miq/vmdb/GUID' log_callback('Running {} to generate UUID'.format(guid_gen)) with self.ssh_client as ssh: result = ssh.run_command(guid_gen) assert result.success, 'Failed to generate UUID' log_callback('Updated UUID: {}'.format(str(result))) try: del self.__dict__['guid'] # invalidate cached_property except KeyError: logger.exception('Exception clearing cached_property "guid"') return str(result).rstrip('\n') # should return UUID from stdout
[docs] def wait_for_ssh(self, timeout=600): """Waits for appliance SSH connection to be ready Args: timeout: Number of seconds to wait until timeout (default ``600``) """ wait_for(func=lambda: self.is_ssh_running, message='appliance.is_ssh_running', delay=5, num_sec=timeout)
@property def _ansible_pod_name(self): if self.is_pod: if self.version >= '5.9': get_ansible_name = ("basename $(oc get pods -lname=ansible " "-o name --namespace={n})".format(n=self.project)) return str(self.ssh_client.run_command(get_ansible_name, ensure_host=True)).strip() else: # ansible stuff lives in the same container with main app in 5.8 return self.container else: return None @property def is_supervisord_running(self): output = self.ssh_client.run_command("systemctl status supervisord", container=self._ansible_pod_name) return output.success @property def is_nginx_running(self): output = self.ssh_client.run_command("systemctl status nginx", container=self._ansible_pod_name) return output.success @property def is_rabbitmq_running(self): output = self.ssh_client.run_command("systemctl status rabbitmq-server", container=self._ansible_pod_name) return output.success @property def is_embedded_ansible_role_enabled(self): return self.server_roles.get("embedded_ansible", False) @property def is_embedded_ansible_running(self): return self.is_embedded_ansible_role_enabled and self.is_supervisord_running
[docs] def wait_for_embedded_ansible(self, timeout=900): """Waits for embedded ansible to be ready Args: timeout: Number of seconds to wait until timeout (default ``900``) """ if self.is_pod and BZ(1576744, forced_streams=['5.9']).blocks: # TODO: to remove this when BZ is fixed # openshift's ansible pod gets ready very long first time. # it even gets restarted once or twice timeout *= 3 wait_for( func=lambda: self.is_embedded_ansible_running, message='appliance.is_embedded_ansible_running', delay=60, num_sec=timeout )
@cached_property def get_host_address(self): try: server = self.advanced_settings.get('server') if server: return server.get('host') except Exception as e: logger.exception(e) self.log.error('Exception occured while fetching host address')
[docs] def wait_for_host_address(self): try: wait_for(func=lambda: getattr(self, 'get_host_address'), fail_condition=None, delay=5, num_sec=120) return self.get_host_address except Exception as e: logger.exception(e) self.log.error('waiting for host address from yaml_config timedout')
@property def is_ssh_running(self): if self.openshift_creds and 'hostname' in self.openshift_creds: hostname = self.openshift_creds['hostname'] else: hostname = self.hostname return net_check(ports.SSH, hostname, force=True) @property def has_cli(self): return self.ssh_client.run_command('hash appliance_console_cli').success @property def is_idle(self): """Return appliance idle state measured by last production.log activity. It runs one liner script, which first gathers current date on appliance and then gathers date of last entry in production.log(which has to be parsed) with /api calls filtered (These calls occur every minute.) Then it deducts that last time in log from current date and if it is lower than idle_time it returns False else True. Args: Returns: True if appliance is idling for longer or equal to idle_time seconds. False if appliance is not idling for longer or equal to idle_time seconds. """ idle_time = 3600 ssh_output = self.ssh_client.run_command('if [ $((`date "+%s"` - `date -d "$(egrep -v ' r'"(Processing by Api::ApiController\#index as JSON|Started GET "/api" for ' '|Completed 200 OK in)" /var/www/miq/vmdb/log/production.log | tail -1 |cut ' '-d"[" -f3 | cut -d"]" -f1 | cut -d" " -f1)\" \"+%s\"`)) -lt {} ];' 'then echo "False";' 'else echo "True";' 'fi;'.format(idle_time)) return True if 'True' in ssh_output else False @cached_property def build_datetime(self): build_datetime_string ='_', 1)[0] return datetime.strptime(build_datetime_string, '%Y%m%d%H%M%S') @cached_property def build_date(self): return
[docs] def has_netapp(self): return self.ssh_client.appliance_has_netapp()
@cached_property def guid(self): try: server = self.rest_api.get_entity_by_href(self.rest_api.server_info['server_href']) return server.guid except (AttributeError, KeyError, IOError): self.log.exception('appliance.guid could not be retrieved from REST, falling back') result = self.ssh_client.run_command('cat /var/www/miq/vmdb/GUID') return result.output @cached_property def evm_id(self): try: server = self.rest_api.get_entity_by_href(self.rest_api.server_info['server_href']) return except (AttributeError, KeyError, IOError): self.log.exception('appliance.evm_id could not be retrieved from REST, falling back') miq_servers = self.db.client['miq_servers'] return self.db.client.session.query( == self.guid)[0][0] @property def fqdn(self): """fqdn from appliance_console This should likely be 'hostname' as that is what its called on the appliance Currently hostname attribute holds IP addr """ return self.rest_api.get_entity_by_href(self.rest_api.server_info['server_href']).hostname
[docs] def get_disabled_regions(self, provider=None): """Fetch appliance advanced config, get disabled regions for given provider's type Only relevant for cloud providers azure and ec2 at the moment Args: provider: A BaseProvider object with settings_key attribute Returns: Default: Dict of ems_<provider> keys and values of disabled_regions map when provider given: disabled_regions list from config when no matching config found: None """ ems_config = self.advanced_settings.get('ems') if provider and ems_config: try: prov_config = ems_config.get(getattr(provider, 'settings_key', None), {}) # safe regions = prov_config['disabled_regions'] # KeyError except KeyError: regions = [] elif ems_config: regions = {ems_key: yaml['disabled_regions'] for ems_key, yaml in ems_config.items() if 'disabled_regions' in yaml} else: # 'ems' was NOT in advanced_settings regions = {} return regions
[docs] def set_disabled_regions(self, provider, *regions): """Modify config to set disabled regions to given regions for the given provider's type Only relevant for cloud providers azure and ec2 at the moment Does NOT APPEND to the list of disabled regions, SETS it Args: provider: A BaseProvider object with settings_key attribute *regions: none, one or many region names, on None enables all regions for provider type Raises: AssertionError - when the disabled regions don't match after setting ApplianceException - when there's a KeyError modifying the yaml """ try: yaml_conf = { 'ems': {getattr(provider, 'settings_key', None): {'disabled_regions': regions}} } except KeyError: # catches not-found settings_key or 'None' when the provider doesn't have it raise ApplianceException('Provider %s settings_key attribute not set ' 'or not found in config %s' .format(provider, yaml_conf['ems'])) self.update_advanced_settings(yaml_conf) assert self.get_disabled_regions(provider) == list(regions) # its a tuple if empty
@property def server_roles(self): """Return a dictionary of server roles from database""" asr = self.db.client['assigned_server_roles'] sr = self.db.client['server_roles'] all_role_names = {row[0] for row in self.db.client.session.query(} # Query all active server roles assigned to this server query = self.db.client.session\ .query(\ .join(asr, asr.server_role_id ==\ .filter(asr.miq_server_id == self.evm_id)\ .filter( == True) # noqa active_roles = {row[0] for row in query} roles = {role_name: role_name in active_roles for role_name in all_role_names} dead_keys = ['database_owner', 'vdi_inventory'] for key in roles: if not self.is_storage_enabled: if key.startswith('storage'): dead_keys.append(key) if key == 'vmdb_storage_bridge': dead_keys.append(key) for key in dead_keys: try: del roles[key] except KeyError: pass return roles @server_roles.setter def server_roles(self, roles): """Sets the server roles. Requires a dictionary full of the role keys with bool values.""" if self.server_roles == roles: self.log.debug(' Roles already match, returning...') return ansible_old = self.server_roles.get('embedded_ansible', False) ansible_new = roles.get('embedded_ansible', False) enabling_ansible = ansible_old is False and ansible_new is True server_data = self.advanced_settings.get('server', {}) server_data['role'] = ','.join([role for role, boolean in roles.items() if boolean]) self.update_advanced_settings({'server': server_data}) timeout = 600 if enabling_ansible else 300 wait_for(lambda: self.server_roles == roles, num_sec=timeout, delay=15) if enabling_ansible: self.wait_for_embedded_ansible()
[docs] def enable_embedded_ansible_role(self): """Enables embbeded ansible role This is necessary because server_roles does not wait long enough""" roles = self.server_roles roles['embedded_ansible'] = True try: self.server_roles = roles except TimedOutError: wait_for(lambda: self.server_roles == roles, num_sec=600, delay=15) self.wait_for_embedded_ansible()
[docs] def disable_embedded_ansible_role(self): """disables embbeded ansible role""" roles = self.server_roles roles['embedded_ansible'] = False self.server_roles = roles
[docs] def update_server_roles(self, changed_roles): server_roles = self.server_roles.copy() server_roles.update(changed_roles) self.server_roles = server_roles return server_roles == self.server_roles
[docs] def server_id(self): try: return self.server.sid except IndexError: return None
[docs] def server_region_string(self): r = return "{} Region: Region {} [{}]".format( self.product_name, r, r)
@cached_property def company_name(self): return self.advanced_settings["server"]["company"]
[docs] def host_id(self, hostname): hosts = list( self.db.client.session.query(self.db.client["hosts"]).filter( self.db.client["hosts"].name == hostname ) ) if hosts: return str(hosts[0].id) else: return None
@cached_property def is_storage_enabled(self): return 'storage' in self.advanced_settings.get('product', {}) @property def advanced_settings(self): """Get settings from the base api/settings endpoint for appliance""" if self.version > '5.9': return self.rest_api.get(self.rest_api.collections.settings._href) else: writeout = self.ssh_client.run_rails_command( '"\'/tmp/yam_dump.yaml\', \'w\') ' '{|f| f.write(Settings.to_hash.deep_stringify_keys.to_yaml) }"' ) if writeout.rc: logger.error("Config couldn't be found") logger.error(writeout.output) raise Exception('Error obtaining config') base_data = self.ssh_client.run_command('cat /tmp/yam_dump.yaml') if base_data.rc: logger.error("Config couldn't be found") logger.error(base_data.output) raise Exception('Error obtaining config') try: return yaml.load(base_data.output) except Exception: logger.debug(base_data.output) raise
[docs] def update_advanced_settings(self, settings_dict): """PATCH settings from the master server's api/server/:id/settings endpoint Uses REST API for CFME 5.9+, uses rails console on lower versions Will automatically update existing settings dictionary with settings_dict Args: data_dict: dictionary of the changes to be made to the yaml configuration JSON dumps data_dict to pass as raw hash data to rest_api session Raises: ApplianceException when server_id isn't set """ # Can only modify through server ID, raise if that's not set yet if self.version < '5.9': data_dict_base = self.advanced_settings data_dict_base.update(settings_dict) temp_yaml = NamedTemporaryFile() dest_yaml = '/tmp/conf.yaml' yaml.dump(data_dict_base, temp_yaml, default_flow_style=False) self.ssh_client.put_file(, dest_yaml) # Build and send ruby script dest_ruby = '/tmp/set_conf.rb' ruby_template = data_path.join('utils', 'cfmedb_set_config.rbt') ruby_replacements = { 'config_file': dest_yaml } temp_ruby = load_data_file(ruby_template.strpath, ruby_replacements) self.ssh_client.put_file(, dest_ruby) # Run it result = self.ssh_client.run_rails_command(dest_ruby) if not result: raise Exception('Unable to set config: {!r}:{!r}'.format(result.rc, result.output)) else: if self.server_id() is None: raise ApplianceException('No server id is set, cannot modify yaml config via REST') self.server.update_advanced_settings(settings_dict)
[docs] def set_proxy(self, host, port, user=None, password=None, prov_type=None): vmdb_config = self.advanced_settings proxy_type = prov_type or 'default' settings = {'host': host, 'port': port, 'user': user, 'password': password} try: vmdb_config['http_proxy'][proxy_type] = settings except KeyError as ex: logger.error('Incorrect provider type') logger.exception(ex) raise ApplianceException('Impossible to create proxy with current provider type') self.update_advanced_settings(vmdb_config)
[docs] def set_session_timeout(self, timeout=86400, quiet=True): """Sets the timeout of UI timeout. Args: timeout: Timeout in seconds quiet: Whether to ignore any errors """ try: session_config = self.advanced_settings.get('session', {}) if session_config.get('timeout') != timeout: session_config['timeout'] = timeout self.update_advanced_settings({'session': session_config}) except Exception as ex: logger.error('Setting session timeout failed:') logger.exception(ex) if not quiet: raise
[docs] def delete_all_providers(self):'Destroying all appliance providers') for prov in self.rest_api.collections.providers: prov.action.delete()
[docs] def reset_automate_model(self): with self.ssh_client as ssh_client: ssh_client.run_rake_command("evm:automate:reset")
[docs] def clean_appliance(self): starttime = time() self.ssh_client.run_command('service evmserverd stop') self.ssh_client.run_command('sync; sync; echo 3 > /proc/sys/vm/drop_caches') self.ssh_client.run_command('service collectd stop') self.ssh_client.run_command('service {}-postgresql restart'.format( self.db.postgres_version)) self.ssh_client.run_command( 'cd /var/www/miq/vmdb; bin/rake evm:db:reset') self.ssh_client.run_rake_command('db:seed') self.ssh_client.run_command('service collectd start') self.ssh_client.run_command('rm -rf /var/www/miq/vmdb/log/*.log*') self.ssh_client.run_command('rm -rf /var/www/miq/vmdb/log/apache/*.log*') self.ssh_client.run_command('service evmserverd start') self.wait_for_evm_service() logger.debug('Cleaned appliance in: {}'.format(round(time() - starttime, 2)))
[docs] def set_full_refresh_threshold(self, threshold=100): yaml_data = {'ems_refresh': {'full_refresh_threshold': threshold}} self.update_advanced_settings(yaml_data)
[docs] def set_cap_and_util_all_via_rails(self): """Turns on Collect for All Clusters and Collect for all Datastores without using Web UI.""" command = ( 'Metric::Targets.perf_capture_always = {:storage=>true, :host_and_cluster=>true};') self.ssh_client.run_rails_console(command, timeout=None)
[docs] def set_cfme_server_relationship(self, vm_name, server_id=1): """Set MiqServer record to the id of a VM by name, effectively setting the CFME Server Relationship without using the Web UI.""" command = ('miq_server = MiqServer.find_by(id: {});' 'miq_server.vm_id = Vm.find_by(name: \'{}\').id;' ''.format(server_id, vm_name)) self.ssh_client.run_rails_console(command, timeout=None)
[docs] def set_pglogical_replication(self, replication_type=':none'): """Set pglogical replication type (:none, :remote, :global) without using the Web UI.""" command = ('MiqRegion.replication_type = {}'.format(replication_type)) self.ssh_client.run_rails_console(command, timeout=None)
[docs] def add_pglogical_replication_subscription(self, host): """Add a pglogical replication subscription without using the Web UI.""" user = conf.credentials['ssh']['username'] password = conf.credentials['ssh']['password'] dbname = 'vmdb_production' port = 5432 command = ('sub =;' 'sub.dbname = \'{}\';' ' = \'{}\';' 'sub.user = \'{}\';' 'sub.password = \'{}\';' 'sub.port = {};' ''.format(dbname, host, user, password, port)) self.ssh_client.run_rails_console(command, timeout=None)
[docs] def set_rubyrep_replication(self, host, port=5432, database='vmdb_production', username='root', password=None): """Sets up rubyrep replication via advanced configuration settings yaml.""" password = password or self._encrypt_string(conf.credentials['ssh']['password']) yaml_data = {'workers': {'worker_base': {'replication_worker': {'replication': { 'destination': {}}}}} } dest = yaml_data['workers']['worker_base']['replication_worker']['replication'][ 'destination'] dest['database'] = database dest['username'] = username dest['password'] = password dest['port'] = port dest['host'] = host logger.debug('Dest: {}'.format(yaml_data)) self.update_advanced_settings(yaml_data)
[docs] def wait_for_miq_server_workers_started(self, evm_tail=None, poll_interval=5): """Waits for the CFME's workers to be started by tailing evm.log for: 'INFO -- : MIQ(MiqServer#wait_for_started_workers) All workers have been started' """ if evm_tail is None:'Opening /var/www/miq/vmdb/log/evm.log for tail') evm_tail = SSHTail('/var/www/miq/vmdb/log/evm.log') evm_tail.set_initial_file_end() attempts = 0 detected = False max_attempts = 60 while (not detected and attempts < max_attempts): logger.debug('Attempting to detect MIQ Server workers started: {}'.format(attempts)) for line in evm_tail: if 'MiqServer#wait_for_started_workers' in line: if ('All workers have been started' in line):'Detected MIQ Server is ready.') detected = True break sleep(poll_interval) # Allow more log lines to accumulate attempts += 1 if not (attempts < max_attempts): logger.error('Could not detect MIQ Server workers started in {}s.'.format( poll_interval * max_attempts)) evm_tail.close()
@logger_wrap("Setting dev branch: {}") def use_dev_branch(self, repo, branch, log_callback=None): """Sets up an exitsing appliance to change the branch to specified one and reset it. Args: repo: URL to the repo branch: Branch of that repo """ with self.ssh_client as ssh_client: dev_branch_cmd = 'cd /var/www/miq/vmdb; git remote add dev_branch {}'.format(repo) if not ssh_client.run_command(dev_branch_cmd): ssh_client.run_command('cd /var/www/miq/vmdb; git remote remove dev_branch') if not ssh_client.run_command(dev_branch_cmd): raise Exception('Could not add the dev_branch remote') # We now have the repo and now let's update it ssh_client.run_command('cd /var/www/miq/vmdb; git remote update') self.evmserverd.stop() ssh_client.run_command( 'cd /var/www/miq/vmdb; git checkout dev_branch/{}'.format(branch)) ssh_client.run_command('cd /var/www/miq/vmdb; bin/update') self.start_evm_service() self.wait_for_evm_service() self.wait_for_web_ui()
[docs] def check_domain_enabled(self, domain): namespaces = self.db.client["miq_ae_namespaces"] q = self.db.client.session.query(namespaces).filter( namespaces.parent_id == None, == domain) # NOQA (for is/==) try: return list(q)[0].enabled except IndexError: raise KeyError("No such Domain: {}".format(domain))
@logger_wrap('Configuring openldap external auth provider') def configure_openldap(self, auth_provider, log_callback=None): """This method changes the /etc/sssd/sssd.conf and /etc/openldap/ldap.conf files to set up the appliance for an external authentication with OpenLdap. Apache file configurations are updated, for webui to take effect. Args: auth_provider: auth provider object derived from cfme.utils.auth.MIQAuthProvider """ # write /etc/hosts entry for ldap hostname TODO DNS for key in ['ipaddress', 'cert_filename', 'cert_filepath', 'ldap_conf', 'sssd_conf']: if not auth_provider.get(key): # either not set, or None raise ValueError('Auth Provider object {} needs attribute {} for external openldap' .format(auth_provider, key)) self.ssh_client.run_command('echo "{} {}" >> /etc/hosts' .format(auth_provider.ipaddress, auth_provider.host1)) # place cert from local conf directory on ldap server self.ssh_client.put_file(local_file=conf_path.join(auth_provider.cert_filename).strpath, remote_file=auth_provider.cert_filepath) # configure ldap and sssd with conf file content from yaml assert self.ssh_client.run_command('echo "{s}" > {c}' .format(s=auth_provider.ldap_conf, c=self.CONF_FILES['openldap'])) assert self.ssh_client.run_command('echo "{s}" > {c}' .format(s=auth_provider.sssd_conf, c=self.CONF_FILES['sssd'])) assert self.ssh_client.run_command('chown -R root:root {}').format(self.CONF_FILES['sssd']) assert self.ssh_client.run_command('chmod 600 {}').format(self.CONF_FILES['sssd']) # copy miq/cfme template files for httpd ext auth config template_dir = self.CONF_FILES.get('downstream_templates' if self.is_downstream else 'upstream_templates') # pam httpd-auth and httpd remote-user.conf for conf_file in [self.CONF_FILES['pam_httpd_auth'], self.CONF_FILES['httpd_remote_user']]: assert self.ssh_client.run_command('cp {t}{c} {c}'.format(t=template_dir, c=conf_file)) # https external-auth conf, template has extra '.erb' suffix assert self.ssh_client.run_command('cp {t}{c}.erb {c}' .format(t=template_dir, c=self.CONF_FILES['httpd_ext_auth'])) assert self.ssh_client.run_command('setenforce 0') self.sssd.restart() self.httpd.restart() self.wait_for_web_ui() # UI configuration of auth provider type self.server.authentication.configure(auth_mode='external', auth_provider=auth_provider) @logger_wrap('Disabling openldap external auth provider') def disable_openldap(self, log_callback=None): self.server.authentication.configure_auth() files_to_remove = [ self.CONF_FILES['sssd'], self.CONF_FILES['pam_httpd_auth'], self.CONF_FILES['httpd_ext_auth'], self.CONF_FILES['httpd_remote_user'] ] for conf_file in files_to_remove: assert self.ssh_client.run_command('rm -f $(ls {})'.format(conf_file)) self.evmserverd.restart() self.httpd.restart() self.wait_for_web_ui() self.server.authentication.configure_auth(auth_mode='database') @logger_wrap('Configuring freeipa external auth provider') def configure_freeipa(self, auth_provider, log_callback=None): """Configure appliance UI and backend for freeIPA Args: auth_provider: An auth provider class derived from cfme.utils.auth.BaseAuthProvider Notes: Completes backend config via appliance_console_cli Completes UI configuration for external auth mode """ if self.is_pod: # appliance_console_cli fails when calls hostnamectl --host. it seems docker issue # raise BZ ? assert str(self.ssh_client.run_command('hostname')).rstrip() == self.fqdn # First, clear any existing ipa config, runs clean if not configured self.appliance_console_cli.uninstall_ipa_client() self.wait_for_web_ui() # httpd restart in uninstall-ipa # ext auth ipa requires NTP sync if auth_provider.host1 not in self.server.settings.ntp_servers_values: self.server.settings.update_ntp_servers({'ntp_server_1': auth_provider.host1}) # backend appliance configuration of ext auth provider self.appliance_console_cli.configure_ipa(**auth_provider.as_external_value()) # UI configuration of auth provider type self.server.authentication.configure(auth_mode='external', auth_provider=auth_provider) # restart httpd self.httpd.restart() @logger_wrap('Disabling freeipa external auth provider') def disable_freeipa(self, log_callback=None): """Switch UI back to database authentication, and run --uninstall-ipa on appliance""" self.appliance_console_cli.uninstall_ipa_client() self.server.authentication.configure(auth_mode='database') self.wait_for_web_ui() # httpd restart in uninstall-ipa @logger_wrap("Configuring VM Console: {}") def configure_vm_console_cert(self, log_callback=None): """This method generates a self signed SSL cert and installs it in the miq/vmdb/certs dir. This cert will be used by the HTML 5 VM Console feature. Note evmserverd needs to be restarted after running this. """ log_callback('Installing SSL certificate') cert = conf.cfme_data['vm_console'].get('cert') if cert is None: raise Exception('vm_console:cert does not exist in cfme_data.yaml') cert_file = os.path.join(cert.install_dir, 'server.cer') key_file = os.path.join(cert.install_dir, 'server.cer.key') cert_generator = scripts_path.join('').strpath remote_cert_generator = os.path.join('/usr/bin', '') # Copy self signed SSL certificate generator to the appliance # because it needs to get the FQDN for the cert it generates. self.ssh_client.put_file(cert_generator, remote_cert_generator) # Generate cert command = ( '{cert_generator}' ' --C "{country}"' ' --ST "{state}"' ' --L "{city}"' ' --O "{organization}"' ' --OU "{organizational_unit}"' ' --keyFile "{key}"' ' --certFile "{cert}"' .format( cert_generator=remote_cert_generator,, state=cert.state,, organization=cert.organization, organizational_unit=cert.organizational_unit, key=key_file, cert=cert_file, ) ) result = self.ssh_client.run_command(command) if not result == 0: raise Exception( 'Failed to generate self-signed SSL cert on appliance: {}'.format( result.output ) ) @logger_wrap('Get the resolvable hostname') def get_resolvable_hostname(self, log_callback=None): """Lookup the hostname based on self.hostname""" # Example lookups with self.hostname as IP and self.hostname as resolvable name # [root@host-192-168-55-85 ~]# host # domain name pointer # [root@host-192-168-55-85 ~]# host # has address host_check = self.ssh_client.run_command('host {}'.format(self.hostname), ensure_host=True) log_callback('Parsing for resolvable hostname from "{}"'.format(host_check.output)) fqdn = None # just in case the logic gets weird if host_check.success and 'domain name pointer' in host_check.output: # resolvable and reverse lookup # parse out the resolved hostname fqdn = host_check.output.split(' ')[-1].rstrip('\n').rstrip('.') log_callback('Found FQDN by appliance IP: "{}"'.format(fqdn)) elif host_check.success and 'has address' in host_check.output: # resolvable and address returned fqdn = self.hostname log_callback('appliance hostname attr is FQDN: "{}"'.format(fqdn)) else: logger.warning('Bad RC from host command or unexpected output,' ' no resolvable hostname found') return fqdn @logger_wrap('Configuring resolvable hostname') def set_resolvable_hostname(self, log_callback=None): """Try to lookup the hostname based on self.hostname, which is generally an IP Intended use is for appliances hosted on openstack, which will likely not have a FQDN hostname after deploying the image. Can be called against any appliance type. If a hostname is resolved for the IP, its used to set the hostname via appliance_cli If an IP is resolved for self.hostname, no action is taken If nothing is resolved, no action is taken Returns: Boolean, True if hostname was set """ fqdn = self.get_resolvable_hostname() if fqdn is not None: log_callback('Setting FQDN "{}" via appliance_console_cli'.format(fqdn)) result = self.appliance_console_cli.set_hostname(str(fqdn)) return result.success else: logger.error('Unable to set resolvable hostname') return False
[docs] def remove_resolvable_hostname(self): """remove a resolvable hostname from /etc/hosts directly USE WITH CAUTION as it mangles /etc/hosts, recommended only for seal_for_templatizing with sprout appliances """ hosts_grep_cmd = 'grep {} /etc/hosts'.format(self.get_resolvable_hostname()) with self.ssh_client as ssh_client: if ssh_client.run_command(hosts_grep_cmd, ensure_host=True).success: # remove resolvable hostname from /etc/hosts # tuples of (loopback, replacement string) for each proto v6 = ('::1', re.escape('::1'.ljust(16))) v4 = ('', re.escape(''.ljust(16))) resolve_esc = re.escape(self.get_resolvable_hostname()) for addr, fill in [v6, v4]: # regex finds lines for loopback addrs where resolvable hostname set # sed replaces with the ljust (space padded) loopback addr ssh_client.run_command( "sed -i -r -e 's|({}\s..*){}|{}|' /etc/hosts".format(addr, resolve_esc, fill), ensure_host=True)
[docs] def provider_based_collection(self, provider, coll_type='vms'): """Given a provider, fetches a collection for the given collection type Some collections are provider based, like infra vms and cloud instances This provides an easy way to pick which one is right for your provider Args: provider: provider class/instance for lookup coll_type: which collection type to return based on the provider type Notes: Add coll_type support as there are collections dependent on a provider type Examples: # returns the infra_vms collection appliance.collections.provider_based_collection(rhevm_provider, 'vms') """ from import CloudProvider # for checking provider type if coll_type == 'vms': return getattr(self.collections, 'cloud_instances' if provider.one_of(CloudProvider) else 'infra_vms') if coll_type == 'templates': return getattr(self.collections, 'cloud_images' if provider.one_of(CloudProvider) else 'infra_templates') else: raise ValueError('No support for coll_type: "{}" collection name lookup' .format(coll_type))
def _switch_migration_ui(self, enable): self.update_advanced_settings({'product': {'transformation': enable}}) self.appliance.server.logout() self.restart_evm_service() self.wait_for_evm_service() self.wait_for_web_ui()
[docs] def enable_migration_ui(self): self._switch_migration_ui(True)
[docs] def disable_migration_ui(self): self. _switch_migration_ui(False)
[docs]class Appliance(IPAppliance): """Appliance represents an already provisioned cfme appliance vm **DO NOT INSTANTIATE DIRECTLY - USE :py:meth:`from_provider`** """ _default_name = 'EVM' @property def ipapp(self): # For backwards compat return self @classmethod
[docs] def from_provider(cls, provider_key, vm_name, name=None, **kwargs): """Constructor of this Appliance. Retrieves the IP address of the appliance from the provider and then instantiates it, adding some extra parameters that are required by this class. Args: provider_name: Name of the provider this appliance is running under vm_name: Name of the VM this appliance is running as browser_steal: Setting of the browser_steal attribute. """ from cfme.utils.providers import get_crud from cfme.containers.provider.openshift import OpenshiftProvider from cfme.utils import conf provider = get_crud(provider_key) app_kwargs = kwargs.copy() if 'hostname' not in app_kwargs: def is_ip_available(): try: # TODO: change after openshift wrapanapi refactor if provider.one_of(OpenshiftProvider): ip = provider.mgmt.get_ip_address(vm_name) else: vm = provider.mgmt.get_vm(vm_name) ip = vm.ip return ip or False # get_ip_address might return None except (AttributeError, VMInstanceNotFound): return False ec, tc = wait_for(is_ip_available, delay=5, num_sec=600) app_kwargs['hostname'] = str(ec) if provider.one_of(OpenshiftProvider): # there should also be present appliance hostname, container, db_host provider_creds = conf.credentials[provider.provider_data['credentials']] ssh_creds = conf.credentials[provider.provider_data['ssh_creds']] if not app_kwargs.get('project'): app_kwargs['project'] = vm_name app_kwargs['openshift_creds'] = { 'hostname': provider.provider_data['hostname'], 'username': provider_creds['username'], 'password': provider_creds['password'], 'ssh': { 'username': ssh_creds['username'], 'password': ssh_creds['password'], } } appliance = cls(**app_kwargs) appliance.vm_name = vm_name appliance.provider = provider.mgmt appliance.provider_key = provider_key = name or cls._default_name return appliance
def _custom_configure(self, **kwargs): log_callback = kwargs.pop( "log_callback", lambda msg:"Custom configure %s: %s", self.vm_name, msg)) region = kwargs.get('region', 0) db_address = kwargs.get('db_address') key_address = kwargs.get('key_address') db_username = kwargs.get('db_username') db_password = kwargs.get('ssh_password') ssh_password = kwargs.get('ssh_password') db_name = kwargs.get('db_name') on_openstack = kwargs.pop('on_openstack', False) if kwargs.get('fix_ntp_clock', True) is True: self.fix_ntp_clock(log_callback=log_callback) if self.is_downstream: # Upstream already has one. if kwargs.get('db_address') is None: # This is workaround for appliances to use only one disk for the VMDB # If they have been provisioned with a second disk in the infra, # 'self.unpartitioned_disks' should exist and therefore this won't run. if not self.unpartitioned_disks: self.db.create_db_lvm() self.db.enable_internal( region, key_address, db_password, ssh_password) else: self.db.enable_external( db_address, region, db_name, db_username, db_password) self.wait_for_web_ui(timeout=1800, log_callback=log_callback) if kwargs.get('loosen_pgssl', True) is True: self.db.loosen_pgssl() name_to_set = kwargs.get('name_to_set') if name_to_set is not None and name_to_set != self.rename(name_to_set) self.restart_evm_service(log_callback=log_callback) self.wait_for_web_ui(log_callback=log_callback) # Try to set a resolvable FQDN on openstack if on_openstack: self.set_resolvable_hostname(log_callback=log_callback) @logger_wrap("Configure Appliance: {}") def configure(self, setup_fleece=False, log_callback=None, **kwargs): """Configures appliance - database setup, rename, ntp sync Utility method to make things easier. Args: db_address: Address of external database if set, internal database if ``None`` (default ``None``) name_to_set: Name to set the appliance name to if not ``None`` (default ``None``) region: Number to assign to region (default ``0``) fix_ntp_clock: Fixes appliance time if ``True`` (default ``True``) loosen_pgssl: Loosens postgres connections if ``True`` (default ``True``) key_address: Fetch encryption key from this address if set, generate a new key if ``None`` (default ``None``) """ log_callback("Configuring appliance {} on {}".format(self.vm_name, self.provider_key)) if kwargs: with self: self._custom_configure(**kwargs) else: # Defer to the IPAppliance. super(Appliance, self).configure(log_callback=log_callback) # And do configure the fleecing if requested if setup_fleece: self.configure_fleecing(log_callback=log_callback) @logger_wrap("Configure fleecing: {}") def configure_fleecing(self, log_callback=None): with self(browser_steal=True): if self.is_on_vsphere: self.install_vddk(reboot=True, log_callback=log_callback) self.wait_for_web_ui(log_callback=log_callback) if self.is_on_rhev: self.add_rhev_direct_lun_disk() log_callback('Enabling smart proxy role...') roles = self.server.settings.server_roles_db if not roles["smartproxy"]: self.server.settings.enable_server_roles("smartproxy") # add provider log_callback('Setting up provider...') self.provider.setup() # credential hosts log_callback('Credentialing hosts...') if not RUNNING_UNDER_SPROUT: from cfme.utils.hosts import setup_providers_hosts_credentials setup_providers_hosts_credentials(self.provider_key, ignore_errors=True) # if rhev, set relationship if self.is_on_rhev: from cfme.infrastructure.virtual_machines import InfraVm log_callback('Setting up CFME VM relationship...') from cfme.utils.providers import get_crud vm = self.collections.infra_vms.instantiate(self.vm_name, get_crud(self.provider_key)) cfme_rel = InfraVm.CfmeRelationship(vm) cfme_rel.set_relationship(str(, self.server.sid) @cached_property def mgmt(self): return self.provider.mgmt.get_vm(self.vm_name)
[docs] def does_vm_exist(self): return self.provider.mgmt.does_vm_exist(self.vm_name)
[docs] def rename(self, new_name): """Changes appliance name Args: new_name: Name to set Note: Database must be up and running and evm service must be (re)started afterwards for the name change to take effect. """ vmdb_config = {'server': {'name': new_name}} self.update_advanced_settings(vmdb_config) = new_name
[docs] def destroy(self): """Destroys the VM this appliance is running as """ if self.is_on_rhev: # if rhev, try to remove direct_lun just in case it is detach self.remove_rhev_direct_lun_disk() if self.does_vm_exist(): self.mgmt.cleanup()
[docs] def stop(self): """Stops the VM this appliance is running as """ self.mgmt.ensure_state(VmState.STOPPED)
[docs] def start(self): """Starts the VM this appliance is running as """ self.mgmt.ensure_state(VmState.RUNNING)
[docs] def templatize(self, seal=True): """Marks the appliance as a template. Destroys the original VM in the process. By default it runs the sealing process. If you have done it differently, you can opt out. Args: seal: Whether to run the sealing process (making the VM 'universal'). """ if seal: if not self.is_running: self.start() self.seal_for_templatizing() self.stop() else: if self.is_running: self.stop() self.mgmt.mark_as_template()
@property def is_running(self): return self.mgmt.is_running @property def is_on_rhev(self): from cfme.infrastructure.provider.rhevm import RHEVMProvider return isinstance(self.provider, RHEVMProvider.mgmt_class) @property def is_on_vsphere(self): from cfme.infrastructure.provider.virtualcenter import VMwareProvider return isinstance(self.provider, VMwareProvider.mgmt_class) @property def is_on_openshift(self): from cfme.containers.provider.openshift import OpenshiftProvider return isinstance(self.provider, OpenshiftProvider.mgmt_class)
[docs] def set_ansible_url(self): if self.is_on_openshift: config_map = self.provider.get_appliance_tags(self.project) url = config_map['cfme-openshift-embedded-ansible']['url'] tag = config_map['cfme-openshift-embedded-ansible']['tag'] config = {'embedded_ansible': {'container': {'image_name': url, 'image_tag': tag}}} self.update_advanced_settings(config)
@property def _lun_name(self): return "{}LUNDISK".format(self.vm_name)
[docs] def add_rhev_direct_lun_disk(self, log_callback=None): if log_callback is None: log_callback = if not self.is_on_rhev: log_callback("appliance NOT on rhev, unable to connect direct_lun") raise ApplianceException("appliance NOT on rhev, unable to connect direct_lun") log_callback('Adding RHEV direct_lun hook...') self.wait_for_ssh() try: self.mgmt.connect_direct_lun(lun_name=self._lun_name) except Exception as e: log_callback("Appliance {} failed to connect RHEV direct LUN.".format(self.vm_name)) log_callback(str(e)) raise
@logger_wrap("Remove RHEV LUN: {}") def remove_rhev_direct_lun_disk(self, log_callback=None): if not self.is_on_rhev: msg = "appliance {} NOT on rhev, unable to disconnect direct_lun".format(self.vm_name) log_callback(msg) raise ApplianceException(msg) log_callback('Removing RHEV direct_lun hook...') self.wait_for_ssh() try: self.mgmt.disconnect_disk(self._lun_name) except Exception as e: log_callback("Appliance {} failed to connect RHEV direct LUN.".format(self.vm_name)) log_callback(str(e)) raise
[docs]def provision_appliance(version=None, vm_name_prefix='cfme', template=None, provider_name=None, vm_name=None): """Provisions fresh, unconfigured appliance of a specific version Note: Version must be mapped to template name under ``appliance_provisioning > versions`` in ``cfme_data.yaml``. If no matching template for given version is found, and trackerbot is set up, the latest available template of the same stream will be used. E.g.: if there is no template for but there is, it will be used instead. If both template name and version are specified, template name takes priority. Args: version: version of appliance to provision vm_name_prefix: name prefix to use when deploying the appliance vm Returns: Unconfigured appliance; instance of :py:class:`Appliance` Usage: my_appliance = provision_appliance('', 'my_tests') my_appliance.fix_ntp_clock() ...other configuration... my_appliance.db.enable_internal() my_appliance.wait_for_web_ui() or my_appliance = provision_appliance('', 'my_tests') my_appliance.configure() """ def _generate_vm_name(): if version is not None: version_digits = ''.join([letter for letter in version if letter.isdigit()]) return '{}_{}_{}'.format( vm_name_prefix, version_digits, fauxfactory.gen_alphanumeric(8)) else: return '{}_{}'.format(vm_name_prefix, fauxfactory.gen_alphanumeric(8)) def _get_latest_template(): from cfme.utils import trackerbot api = trackerbot.api() stream = get_stream(version) template_data = trackerbot.latest_template(api, stream, provider_name) return template_data.get('latest_template') if provider_name is None: provider_name = conf.cfme_data.get('appliance_provisioning', {})['default_provider'] if template is not None: template_name = template elif version is not None: templates_by_version = conf.cfme_data.get('appliance_provisioning', {}).get('versions', {}) try: template_name = templates_by_version[version] except KeyError: # We try to get the latest template from the same stream - if trackerbot is set up if conf.env.get('trackerbot', {}): template_name = _get_latest_template() if not template_name: raise ApplianceException('No template found for stream {} on provider {}' .format(get_stream(version), provider_name)) logger.warning('No template found matching version %s, using %s instead.', version, template_name) else: raise ApplianceException('No template found matching version {}'.format(version)) else: raise ApplianceException('Either version or template name must be specified') prov_data = conf.cfme_data.get('management_systems', {})[provider_name] from cfme.utils.providers import get_mgmt provider = get_mgmt(provider_name) if not vm_name: vm_name = _generate_vm_name() deploy_args = {} deploy_args['vm_name'] = vm_name if prov_data['type'] == 'rhevm': deploy_args['cluster'] = prov_data['default_cluster'] if prov_data["type"] == "virtualcenter": if "allowed_datastores" in prov_data: deploy_args["allowed_datastores"] = prov_data["allowed_datastores"] template = provider.get_template(template_name) template.deploy(**deploy_args) return Appliance(provider_name, vm_name)
[docs]class ApplianceStack(LocalStack):
[docs] def push(self, obj): was_before = super(ApplianceStack, self).push(obj)"Pushed appliance {} on stack (was {} before) ".format( obj.hostname, getattr(was_before, 'hostname', 'empty'))) if obj.browser_steal: from cfme.utils import browser browser.start()
[docs] def pop(self): was_before = super(ApplianceStack, self).pop() current = "Popped appliance {} from the stack (now there is {})".format( getattr(was_before, 'address', 'empty'), getattr(current, 'address', 'empty'))) if getattr(was_before, 'browser_steal', False): from cfme.utils import browser browser.start() return was_before
stack = ApplianceStack()
[docs]def load_appliances(appliance_list, global_kwargs): """Instantiate a list of appliances from configuration data. Args: appliance_list: List of dictionaries that contain parameters for :py:class:`IPAppliance` global_kwargs: Arguments that will be defined for each appliances. Appliance can override. Result: List of :py:class:`IPAppliance` """ result = [] for idx, appliance_kwargs in enumerate(appliance_list): kwargs = {} kwargs.update(global_kwargs) kwargs.update(appliance_kwargs) if kwargs.pop('dummy', False): result.append(DummyAppliance(**kwargs)) else: mapping = IPAppliance.CONFIG_MAPPING if not any(k in mapping for k in kwargs): raise ValueError( "No valid IPAppliance kwargs found in config for appliance #{}".format(idx) ) appliance = IPAppliance(**{mapping[k]: v for k, v in kwargs.items() if k in mapping}) result.append(appliance) return result
def _version_for_version_or_stream(version_or_stream, sprout_client=None): if version_or_stream is attr.NOTHING: return attr.fields(DummyAppliance).version.default if isinstance(version_or_stream, Version): return version_or_stream assert isinstance(version_or_stream, six.string_types), version_or_stream from cfme.test_framework.sprout.client import SproutClient sprout_client = SproutClient.from_config() if sprout_client is None else sprout_client if version_or_stream[0].isdigit(): # presume streams start with non-number return Version(version_or_stream) for version_str in sprout_client.available_cfme_versions(): version = Version(version_str) if == version_or_stream: return version raise LookupError(version_or_stream)
[docs]def collections_for_appliance(appliance): from cfme.modeling.base import EntityCollections return EntityCollections.for_appliance(appliance)
[docs]class DummyAppliance(object): """a dummy with minimal attribute set""" hostname = 'DummyApplianceHostname' browser_steal = False version = attr.ib(default=Version('5.8.0'), converter=_version_for_version_or_stream) is_downstream = True is_pod = False is_dev = False build = 'missing :)' managed_known_providers = [] collections = attr.ib(default=attr.Factory(collections_for_appliance, takes_self=True)) @classmethod
[docs] def from_config(cls, pytest_config): version = pytest_config.getoption('--dummy-appliance-version') return cls(version=(version or attr.NOTHING))
[docs] def set_session_timeout(self, *k): pass
[docs]def find_appliance(obj, require=True): if isinstance(obj, NavigatableMixin): return obj.appliance # duck type - either is the config of pytest, or holds it config = getattr(obj, 'config', obj) from cfme.test_framework.appliance import PLUGIN_KEY holder = config.pluginmanager.get_plugin(PLUGIN_KEY) if holder or require: assert holder return holder.held_appliance
[docs]def load_appliances_from_config(config): """ Instantiate IPAppliance objects based on data in ``appliances`` section of config. The ``config`` contains some global values and ``appliances`` key which contains a list of dicts that have the same keys as ``IPAppliance.CONFIG_MAPPING``'s keys. The global values in the root of the dict have lesser priority than the values in appliance definitions themselves Args: config: A dictionary with the configuration """ if 'appliances' not in config: raise ValueError("Invalid config: missing an 'appliances' section") appliances = config['appliances'] global_kwargs = { k: config[k] for k in IPAppliance.CONFIG_MAPPING.keys() if k not in IPAppliance.CONFIG_NONGLOBAL and k in config} return load_appliances(appliances, global_kwargs)
[docs]class ApplianceSummoningWarning(Warning): """to ease filtering/erroring on magical appliance creation based on script vs code"""
[docs]def get_or_create_current_appliance(): if is None: warnings.warn( "magical creation of appliance objects has been deprecated," " please obtain a appliance object directly", category=ApplianceSummoningWarning, ) stack.push(load_appliances_from_config(conf.env)[0]) return
current_appliance = LocalProxy(get_or_create_current_appliance) class _CurrentAppliance(object): def __get__(self, instance, owner): return get_or_create_current_appliance() warnings.simplefilter('ignore', NavigatableDeprecationWarning) @removals.removed_class( "Navigatable", message=("Navigatable is being deprecated in favour of using Collections " "objects with the NavigatableMixin"), category=NavigatableDeprecationWarning, )
[docs]class MiqImplementationContext(sentaku.ImplementationContext): """ Our context for Sentaku""" pass