OpenStack Testing (Tempest) of an existing cloud
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
tempest/tempest/common/compute.py

450 lines
19 KiB

# Copyright (c) 2015 Hewlett-Packard Development Company, L.P.
# All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import base64
import socket
import ssl
import struct
import textwrap
from urllib import parse as urlparse
from oslo_log import log as logging
from oslo_utils import excutils
from tempest.common import waiters
from tempest import config
from tempest.lib.common import fixed_network
from tempest.lib.common import rest_client
from tempest.lib.common.utils import data_utils
CONF = config.CONF
LOG = logging.getLogger(__name__)
def is_scheduler_filter_enabled(filter_name):
"""Check the list of enabled compute scheduler filters from config.
This function checks whether the given compute scheduler filter is enabled
in the nova config file. If the scheduler_enabled_filters option is set to
'all' in tempest.conf then, this function returns True with assumption that
requested filter 'filter_name' is one of the enabled filters in nova
("nova.scheduler.filters.all_filters").
"""
filters = CONF.compute_feature_enabled.scheduler_enabled_filters
if not filters:
return False
if 'all' in filters:
return True
if filter_name in filters:
return True
return False
def create_test_server(clients, validatable=False, validation_resources=None,
tenant_network=None, wait_until=None,
volume_backed=False, name=None, flavor=None,
image_id=None, wait_for_sshable=True, **kwargs):
"""Common wrapper utility returning a test server.
This method is a common wrapper returning a test server that can be
pingable or sshable.
:param clients: Client manager which provides OpenStack Tempest clients.
:param validatable: Whether the server will be pingable or sshable.
:param validation_resources: Resources created for the connection to the
server. Include a keypair, a security group and an IP.
:param tenant_network: Tenant network to be used for creating a server.
:param wait_until: Server status to wait for the server to reach after
its creation.
:param volume_backed: Whether the server is volume backed or not.
If this is true, a volume will be created and create server will be
requested with 'block_device_mapping_v2' populated with below values:
.. code-block:: python
bd_map_v2 = [{
'uuid': volume['volume']['id'],
'source_type': 'volume',
'destination_type': 'volume',
'boot_index': 0,
'delete_on_termination': True}]
kwargs['block_device_mapping_v2'] = bd_map_v2
If server needs to be booted from volume with other combination of bdm
inputs than mentioned above, then pass the bdm inputs explicitly as
kwargs and image_id as empty string ('').
:param name: Name of the server to be provisioned. If not defined a random
string ending with '-instance' will be generated.
:param flavor: Flavor of the server to be provisioned. If not defined,
CONF.compute.flavor_ref will be used instead.
:param image_id: ID of the image to be used to provision the server. If not
defined, CONF.compute.image_ref will be used instead.
:param wait_for_sshable: Check server's console log and wait until it will
be ready to login.
:returns: a tuple
"""
# TODO(jlanoux) add support of wait_until PINGABLE/SSHABLE
if name is None:
name = data_utils.rand_name(__name__ + "-instance")
if flavor is None:
flavor = CONF.compute.flavor_ref
if image_id is None:
image_id = CONF.compute.image_ref
kwargs = fixed_network.set_networks_kwarg(
tenant_network, kwargs) or {}
multiple_create_request = (max(kwargs.get('min_count', 0),
kwargs.get('max_count', 0)) > 1)
if CONF.validation.run_validation and validatable:
# As a first implementation, multiple pingable or sshable servers will
# not be supported
if multiple_create_request:
msg = ("Multiple pingable or sshable servers not supported at "
"this stage.")
raise ValueError(msg)
LOG.debug("Provisioning test server with validation resources %s",
validation_resources)
if 'security_groups' in kwargs:
kwargs['security_groups'].append(
{'name': validation_resources['security_group']['name']})
else:
try:
kwargs['security_groups'] = [
{'name': validation_resources['security_group']['name']}]
except KeyError:
LOG.debug("No security group provided.")
if 'key_name' not in kwargs:
try:
kwargs['key_name'] = validation_resources['keypair']['name']
except KeyError:
LOG.debug("No key provided.")
if CONF.validation.connect_method == 'floating':
if wait_until is None:
wait_until = 'ACTIVE'
if 'user_data' not in kwargs:
# If nothing overrides the default user data script then run
# a simple script on the host to print networking info. This is
# to aid in debugging ssh failures.
script = '''
#!/bin/sh
echo "Printing {user} user authorized keys"
cat ~{user}/.ssh/authorized_keys || true
'''.format(user=CONF.validation.image_ssh_user)
script_clean = textwrap.dedent(script).lstrip().encode('utf8')
script_b64 = base64.b64encode(script_clean)
kwargs['user_data'] = script_b64
if volume_backed:
volume_name = data_utils.rand_name(__name__ + '-volume')
volumes_client = clients.volumes_client_latest
params = {'name': volume_name,
'imageRef': image_id,
'size': CONF.volume.volume_size}
if CONF.compute.compute_volume_common_az:
params.setdefault('availability_zone',
CONF.compute.compute_volume_common_az)
volume = volumes_client.create_volume(**params)
try:
waiters.wait_for_volume_resource_status(volumes_client,
volume['volume']['id'],
'available')
except Exception:
with excutils.save_and_reraise_exception():
try:
volumes_client.delete_volume(volume['volume']['id'])
volumes_client.wait_for_resource_deletion(
volume['volume']['id'])
except Exception as exc:
LOG.exception("Deleting volume %s failed, exception %s",
volume['volume']['id'], exc)
bd_map_v2 = [{
'uuid': volume['volume']['id'],
'source_type': 'volume',
'destination_type': 'volume',
'boot_index': 0,
'delete_on_termination': True}]
kwargs['block_device_mapping_v2'] = bd_map_v2
# Since this is boot from volume an image does not need
# to be specified.
image_id = ''
if CONF.compute.compute_volume_common_az:
kwargs.setdefault('availability_zone',
CONF.compute.compute_volume_common_az)
body = clients.servers_client.create_server(name=name, imageRef=image_id,
flavorRef=flavor,
**kwargs)
# handle the case of multiple servers
if multiple_create_request:
# Get servers created which name match with name param.
body_servers = clients.servers_client.list_servers()
servers = \
[s for s in body_servers['servers'] if s['name'].startswith(name)]
else:
body = rest_client.ResponseBody(body.response, body['server'])
servers = [body]
def _setup_validation_fip():
if CONF.service_available.neutron:
ifaces = clients.interfaces_client.list_interfaces(server['id'])
validation_port = None
for iface in ifaces['interfaceAttachments']:
if iface['net_id'] == tenant_network['id']:
validation_port = iface['port_id']
break
if not validation_port:
# NOTE(artom) This will get caught by the catch-all clause in
# the wait_until loop below
raise ValueError('Unable to setup floating IP for validation: '
'port not found on tenant network')
clients.floating_ips_client.update_floatingip(
validation_resources['floating_ip']['id'],
port_id=validation_port)
else:
fip_client = clients.compute_floating_ips_client
fip_client.associate_floating_ip_to_server(
floating_ip=validation_resources['floating_ip']['ip'],
server_id=servers[0]['id'])
if wait_until:
for server in servers:
try:
waiters.wait_for_server_status(
clients.servers_client, server['id'], wait_until)
# Multiple validatable servers are not supported for now. Their
# creation will fail with the condition above.
if CONF.validation.run_validation and validatable:
if CONF.validation.connect_method == 'floating':
_setup_validation_fip()
except Exception:
with excutils.save_and_reraise_exception():
for server in servers:
try:
clients.servers_client.delete_server(
server['id'])
except Exception:
LOG.exception('Deleting server %s failed',
server['id'])
for server in servers:
# NOTE(artom) If the servers were booted with volumes
# and with delete_on_termination=False we need to wait
# for the servers to go away before proceeding with
# cleanup, otherwise we'll attempt to delete the
# volumes while they're still attached to servers that
# are in the process of being deleted.
try:
waiters.wait_for_server_termination(
clients.servers_client, server['id'])
except Exception:
LOG.exception('Server %s failed to delete in time',
server['id'])
if (validatable and CONF.compute_feature_enabled.console_output and
wait_for_sshable):
waiters.wait_for_guest_os_boot(clients.servers_client, server['id'])
return body, servers
def shelve_server(servers_client, server_id, force_shelve_offload=False):
"""Common wrapper utility to shelve server.
This method is a common wrapper to make server in 'SHELVED'
or 'SHELVED_OFFLOADED' state.
:param servers_clients: Compute servers client instance.
:param server_id: Server to make in shelve state
:param force_shelve_offload: Forcefully offload shelve server if it
is configured not to offload server
automatically after offload time.
"""
servers_client.shelve_server(server_id)
offload_time = CONF.compute.shelved_offload_time
if offload_time >= 0:
waiters.wait_for_server_status(servers_client, server_id,
'SHELVED_OFFLOADED',
extra_timeout=offload_time)
else:
waiters.wait_for_server_status(servers_client, server_id, 'SHELVED')
if force_shelve_offload:
servers_client.shelve_offload_server(server_id)
waiters.wait_for_server_status(servers_client, server_id,
'SHELVED_OFFLOADED')
def create_websocket(url):
url = urlparse.urlparse(url)
# NOTE(mnaser): It is possible that there is no port specified, so fall
# back to the default port based on the scheme.
port = url.port or (443 if url.scheme == 'https' else 80)
for res in socket.getaddrinfo(url.hostname, port,
socket.AF_UNSPEC, socket.SOCK_STREAM):
af, socktype, proto, _, sa = res
client_socket = socket.socket(af, socktype, proto)
if url.scheme == 'https':
client_socket = ssl.wrap_socket(client_socket)
client_socket.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
try:
client_socket.connect(sa)
except socket.error:
client_socket.close()
continue
break
else:
raise socket.error('WebSocket creation failed')
# Turn the Socket into a WebSocket to do the communication
return _WebSocket(client_socket, url)
class _WebSocket(object):
def __init__(self, client_socket, url):
"""Contructor for the WebSocket wrapper to the socket."""
self._socket = client_socket
# cached stream for early frames.
self.cached_stream = b''
# Upgrade the HTTP connection to a WebSocket
self._upgrade(url)
def _recv(self, recv_size):
"""Wrapper to receive data from the cached stream or socket."""
if recv_size <= 0:
return None
data_from_cached = b''
data_from_socket = b''
if len(self.cached_stream) > 0:
read_from_cached = min(len(self.cached_stream), recv_size)
data_from_cached += self.cached_stream[:read_from_cached]
self.cached_stream = self.cached_stream[read_from_cached:]
recv_size -= read_from_cached
if recv_size > 0:
data_from_socket = self._socket.recv(recv_size)
return data_from_cached + data_from_socket
def receive_frame(self):
"""Wrapper for receiving data to parse the WebSocket frame format"""
# We need to loop until we either get some bytes back in the frame
# or no data was received (meaning the socket was closed). This is
# done to handle the case where we get back some empty frames
while True:
header = self._recv(2)
# If we didn't receive any data, just return None
if not header:
return None
# We will make the assumption that we are only dealing with
# frames less than 125 bytes here (for the negotiation) and
# that only the 2nd byte contains the length, and since the
# server doesn't do masking, we can just read the data length
if int(header[1]) & 127 > 0:
return self._recv(int(header[1]) & 127)
def send_frame(self, data):
"""Wrapper for sending data to add in the WebSocket frame format."""
frame_bytes = list()
# For the first byte, want to say we are sending binary data (130)
frame_bytes.append(130)
# Only sending negotiation data so don't need to worry about > 125
# We do need to add the bit that says we are masking the data
frame_bytes.append(len(data) | 128)
# We don't really care about providing a random mask for security
# So we will just hard-code a value since a test program
mask = [7, 2, 1, 9]
for i in range(len(mask)):
frame_bytes.append(mask[i])
# Mask each of the actual data bytes that we are going to send
for i in range(len(data)):
frame_bytes.append(int(data[i]) ^ mask[i % 4])
# Convert our integer list to a binary array of bytes
frame_bytes = struct.pack('!%iB' % len(frame_bytes), * frame_bytes)
self._socket.sendall(frame_bytes)
def close(self):
"""Helper method to close the connection."""
# Close down the real socket connection and exit the test program
if self._socket is not None:
self._socket.shutdown(1)
self._socket.close()
self._socket = None
def _upgrade(self, url):
"""Upgrade the HTTP connection to a WebSocket and verify."""
# It is possible to pass the path as a query parameter in the request,
# so use it if present
# Given noVNC format
# https://x.com/vnc_auto.html?path=%3Ftoken%3Dxxx,
# url format is
# ParseResult(scheme='https', netloc='x.com',
# path='/vnc_auto.html', params='',
# query='path=%3Ftoken%3Dxxx', fragment='').
# qparams format is {'path': ['?token=xxx']}
qparams = urlparse.parse_qs(url.query)
# according to references
# https://docs.python.org/3/library/urllib.parse.html
# https://tools.ietf.org/html/rfc3986#section-3.4
# qparams['path'][0] format is '?token=xxx' without / prefix
# remove / in /websockify to comply to references.
path = qparams['path'][0] if 'path' in qparams else 'websockify'
# Fix websocket request format by adding / prefix.
# Updated request format: GET /?token=xxx HTTP/1.1
# or GET /websockify HTTP/1.1
reqdata = 'GET /%s HTTP/1.1\r\n' % path
reqdata += 'Host: %s' % url.hostname
# Add port only if we have one specified
if url.port:
reqdata += ':%s' % url.port
# Line-ending for Host header
reqdata += '\r\n'
# Tell the HTTP Server to Upgrade the connection to a WebSocket
reqdata += 'Upgrade: websocket\r\nConnection: Upgrade\r\n'
# The token=xxx is sent as a Cookie not in the URI for noVNC < v1.1.0
reqdata += 'Cookie: %s\r\n' % url.query
# Use a hard-coded WebSocket key since a test program
reqdata += 'Sec-WebSocket-Key: x3JJHMbDL1EzLkh9GBhXDw==\r\n'
reqdata += 'Sec-WebSocket-Version: 13\r\n'
# We are choosing to use binary even though browser may do Base64
reqdata += 'Sec-WebSocket-Protocol: binary\r\n\r\n'
# Send the HTTP GET request and get the response back
self._socket.sendall(reqdata.encode('utf8'))
self.response = data = self._socket.recv(4096)
# Loop through & concatenate all of the data in the response body
end_loc = self.response.find(b'\r\n\r\n')
while data and end_loc < 0:
data = self._socket.recv(4096)
self.response += data
end_loc = self.response.find(b'\r\n\r\n')
if len(self.response) > end_loc + 4:
# In case some frames (e.g. the first RFP negotiation) have
# arrived, cache it for next reading.
self.cached_stream = self.response[end_loc + 4:]
# ensure response ends with '\r\n\r\n'.
self.response = self.response[:end_loc + 4]