def run(test, params, env):
"""
Test the virsh pool commands
(1) Define a given type pool
(2) List pool with '--inactive --type' options
(3) Dumpxml for the pool
(4) Undefine the pool
(5) Define pool by using the XML file in step (3)
(6) Build the pool(except 'disk' type pool
For 'fs' type pool, cover --overwrite and --no-overwrite options
(7) Start the pool
(8) List pool with '--persistent --type' options
(9) Mark pool autostart
(10) List pool with '--autostart --type' options
(11) Restart libvirtd and list pool with '--autostart --persistent' options
(12) Destroy the pool
(13) Unmark pool autostart
(14) Repeat step (11)
(15) Start the pool
(16) Get pool info
(17) Get pool uuid by name
(18) Get pool name by uuid
(19) Refresh the pool
For 'dir' type pool, touch a file under target path and refresh again
to make the new file show in vol-list.
(20) Check pool 'Capacity', 'Allocation' and 'Available'
Create a over size vol in pool(expect fail), then check these values
(21) Undefine the pool, and this should fail as pool is still active
(22) Destroy the pool
(23) Delete pool for 'dir' type pool. After the command, the pool object
will still exist but target path will be deleted
(24) Undefine the pool
"""
# Initialize the variables
pool_name = params.get("pool_name", "temp_pool_1")
pool_type = params.get("pool_type", "dir")
pool_target = params.get("pool_target", "")
source_format = params.get("source_format", "")
source_name = params.get("pool_source_name", "gluster-vol1")
source_path = params.get("pool_source_path", "/")
new_pool_name = params.get("new_pool_name", "")
build_option = params.get("build_option", "")
iscsi_initiator = params.get("iscsi_initiator", "")
same_source_test = "yes" == params.get("same_source_test", "no")
customize_initiator_iqn = "yes" == params.get("customize_initiator_iqn",
"no")
# The file for dumped pool xml
poolxml = os.path.join(data_dir.get_tmp_dir(), "pool.xml.tmp")
if os.path.dirname(pool_target) is "":
pool_target = os.path.join(data_dir.get_tmp_dir(), pool_target)
vol_name = params.get("volume_name", "temp_vol_1")
# Use pool name as VG name
status_error = "yes" == params.get("status_error", "no")
vol_path = os.path.join(pool_target, vol_name)
ip_protocal = params.get('ip_protocal', 'ipv4')
if not libvirt_version.version_compare(1, 0, 0):
if pool_type == "gluster":
test.cancel("Gluster pool is not supported in current"
" libvirt version.")
if not libvirt_version.version_compare(4, 7, 0):
if pool_type == "iscsi-direct":
test.cancel("iSCSI-direct pool is not supported in current"
"libvirt version.")
def check_pool_list(pool_name, option="--all", expect_error=False):
"""
Check pool by running pool-list command with given option.
:param pool_name: Name of the pool
:param option: option for pool-list command
:param expect_error: Boolean value, expect command success or fail
"""
found = False
# Get the list stored in a variable
result = virsh.pool_list(option, ignore_status=True)
utlv.check_exit_status(result, False)
output = re.findall(r"(\S+)\ +(\S+)\ +(\S+)",
str(result.stdout.strip()))
for item in output:
if pool_name in item[0]:
found = True
break
if found:
logging.debug("Find pool '%s' in pool list.", pool_name)
else:
logging.debug("Not find pool %s in pool list.", pool_name)
if expect_error and found:
test.fail("Unexpect pool '%s' exist." % pool_name)
if not expect_error and not found:
test.fail("Expect pool '%s' doesn't exist." % pool_name)
def check_vol_list(vol_name, pool_name):
"""
Check volume from the list
:param vol_name: Name of the volume
:param pool_name: Name of the pool
"""
found = False
# Get the volume list stored in a variable
result = virsh.vol_list(pool_name, ignore_status=True)
utlv.check_exit_status(result)
output = re.findall(r"(\S+)\ +(\S+)", str(result.stdout.strip()))
for item in output:
if vol_name in item[0]:
found = True
break
if found:
logging.debug("Find volume '%s' in pool '%s'.", vol_name,
pool_name)
else:
test.fail("Not find volume '%s' in pool '%s'." %
(vol_name, pool_name))
def is_in_range(actual, expected, error_percent):
deviation = 100 - (100 * (float(actual) / float(expected)))
logging.debug("Deviation: %0.2f%%", float(deviation))
return float(deviation) <= float(error_percent)
def check_pool_info(pool_info, check_point, value):
"""
Check the pool name, uuid, etc.
:param pool_info: A dict include pool's information
:param key: Key of pool info dict, available value: Name, UUID, State
Persistent, Autostart, Capacity, Allocation, Available
:param value: Expect value of pool_info[key]
"""
if pool_info is None:
test.fail("Pool info dictionary is needed.")
val_tup = ('Capacity', 'Allocation', 'Available')
if check_point in val_tup and float(value.split()[0]):
# As from bytes to GiB, could cause deviation, and it should not
# exceed 1 percent.
if is_in_range(float(pool_info[check_point].split()[0]),
float(value.split()[0]), 1):
logging.debug("Pool '%s' is '%s'.", check_point, value)
else:
test.fail("Pool '%s' isn't '%s'." % (check_point, value))
else:
if pool_info[check_point] == value:
logging.debug("Pool '%s' is '%s'.", check_point, value)
else:
test.fail("Pool '%s' isn't '%s'." % (check_point, value))
# Stop multipathd to avoid start pool fail(For fs like pool, the new add
# disk may in use by device-mapper, so start pool will report disk already
# mounted error).
multipathd = service.Factory.create_service("multipathd")
multipathd_status = multipathd.status()
if multipathd_status:
multipathd.stop()
# Run Testcase
pvt = utlv.PoolVolumeTest(test, params)
kwargs = {
'image_size': '1G',
'pre_disk_vol': ['100M'],
'source_name': source_name,
'source_path': source_path,
'source_format': source_format,
'persistent': True,
'ip_protocal': ip_protocal,
'emulated_image': "emulated-image",
'pool_target': pool_target,
'iscsi_initiator': iscsi_initiator
}
params.update(kwargs)
try:
_pool = libvirt_storage.StoragePool()
# Step (1)
# Pool define
pvt.pre_pool(**params)
# Step (2)
# Pool list
option = "--inactive --type %s" % pool_type
check_pool_list(pool_name, option)
# Step (3)
# Pool dumpxml
xml = virsh.pool_dumpxml(pool_name, to_file=poolxml)
logging.debug("Pool '%s' XML:\n%s", pool_name, xml)
# Update pool name
if new_pool_name:
if "/" in new_pool_name:
new_pool_name = new_pool_name.replace("/", "\/")
logging.debug(new_pool_name)
p_xml = pool_xml.PoolXML.new_from_dumpxml(pool_name)
p_xml.name = new_pool_name
del p_xml.uuid
poolxml = p_xml.xml
logging.debug("XML after update pool name:\n%s" % p_xml)
# Update host name
if same_source_test:
s_xml = p_xml.get_source()
s_xml.host_name = "192.168.1.1"
p_xml.set_source(s_xml)
poolxml = p_xml.xml
logging.debug("XML after update host name:\n%s" % p_xml)
if customize_initiator_iqn:
initiator_iqn = params.get("initiator_iqn",
"iqn.2018-07.com.virttest:pool.target")
p_xml = pool_xml.PoolXML.new_from_dumpxml(pool_name)
s_node = p_xml.xmltreefile.find('/source')
i_node = ET.SubElement(s_node, 'initiator')
ET.SubElement(i_node, 'iqn', {'name': initiator_iqn})
p_xml.xmltreefile.write()
poolxml = p_xml.xml
logging.debug('XML after add Multi-IQN:\n%s' % p_xml)
# Step (4)
# Undefine pool
if not same_source_test:
result = virsh.pool_undefine(pool_name)
utlv.check_exit_status(result)
check_pool_list(pool_name, "--all", True)
# Step (5)
# Define pool from XML file
result = virsh.pool_define(poolxml, debug=True)
# Give error msg when exit status is not expected
if "/" in new_pool_name and not result.exit_status:
error_msg = "https://bugzilla.redhat.com/show_bug.cgi?id=639923 "
error_msg += "is helpful for tracing this bug."
logging.error(error_msg)
if "." in new_pool_name and result.exit_status:
error_msg = "https://bugzilla.redhat.com/show_bug.cgi?id=1333248 "
error_msg += "is helpful for tracing this bug."
logging.error(error_msg)
if same_source_test and not result.exit_status:
error_msg = "https://bugzilla.redhat.com/show_bug.cgi?id=1171984 "
error_msg += "is helpful for tracing this bug."
logging.error(error_msg)
utlv.check_exit_status(result, status_error)
if not result.exit_status:
# Step (6)
# Buid pool
# '--overwrite/--no-overwrite' just for fs/disk/logiacl type pool
# disk/fs pool: as prepare step already make label and create filesystem
# for the disk, use '--overwrite' is necessary
# logical_pool: build pool will fail if VG already exist, BZ#1373711
if new_pool_name:
pool_name = new_pool_name
if pool_type != "logical":
result = virsh.pool_build(pool_name,
build_option,
ignore_status=True)
utlv.check_exit_status(result)
# Step (7)
# Pool start
result = virsh.pool_start(pool_name,
debug=True,
ignore_status=True)
utlv.check_exit_status(result)
# Step (8)
# Pool list
option = "--persistent --type %s" % pool_type
check_pool_list(pool_name, option)
# Step (9)
# Pool autostart
result = virsh.pool_autostart(pool_name, ignore_status=True)
utlv.check_exit_status(result)
# Step (10)
# Pool list
option = "--autostart --type %s" % pool_type
check_pool_list(pool_name, option)
# Step (11)
# Restart libvirtd and check the autostart pool
utils_libvirtd.libvirtd_restart()
option = "--autostart --persistent"
check_pool_list(pool_name, option)
# Step (12)
# Pool destroy
if virsh.pool_destroy(pool_name):
logging.debug("Pool %s destroyed.", pool_name)
else:
test.fail("Destroy pool % failed." % pool_name)
# Step (13)
# Pool autostart disable
result = virsh.pool_autostart(pool_name,
"--disable",
ignore_status=True)
utlv.check_exit_status(result)
# Step (14)
# Repeat step (11)
utils_libvirtd.libvirtd_restart()
option = "--autostart"
check_pool_list(pool_name, option, True)
# Step (15)
# Pool start
# When libvirtd starts up, it'll check to see if any of the storage
# pools have been activated externally. If so, then it'll mark the
# pool as active. This is independent of autostart.
# So a directory based storage pool is thus pretty much always active,
# and so as the SCSI pool.
if pool_type not in ["dir", 'scsi']:
result = virsh.pool_start(pool_name, ignore_status=True)
utlv.check_exit_status(result)
# Step (16)
# Pool info
pool_info = _pool.pool_info(pool_name)
logging.debug("Pool '%s' info:\n%s", pool_name, pool_info)
# Step (17)
# Pool UUID
result = virsh.pool_uuid(pool_info["Name"], ignore_status=True)
utlv.check_exit_status(result)
check_pool_info(pool_info, "UUID", result.stdout.strip())
# Step (18)
# Pool Name
result = virsh.pool_name(pool_info["UUID"], ignore_status=True)
utlv.check_exit_status(result)
check_pool_info(pool_info, "Name", result.stdout.strip())
# Step (19)
# Pool refresh for 'dir' type pool
if pool_type == "dir":
os.mknod(vol_path)
result = virsh.pool_refresh(pool_name)
utlv.check_exit_status(result)
check_vol_list(vol_name, pool_name)
# Step (20)
# Create an over size vol in pool(expect fail), then check pool:
# 'Capacity', 'Allocation' and 'Available'
# For NFS type pool, there's a bug(BZ#1077068) about allocate volume,
# and glusterfs pool not support create volume, so not test them
if pool_type != "netfs":
vol_capacity = "10000G"
vol_allocation = "10000G"
result = virsh.vol_create_as("oversize_vol", pool_name,
vol_capacity, vol_allocation,
"raw")
utlv.check_exit_status(result, True)
new_info = _pool.pool_info(pool_name)
check_items = ["Capacity", "Allocation", "Available"]
for i in check_items:
check_pool_info(pool_info, i, new_info[i])
# Step (21)
# Undefine pool, this should fail as the pool is active
result = virsh.pool_undefine(pool_name, ignore_status=True)
utlv.check_exit_status(result, expect_error=True)
check_pool_list(pool_name, "", False)
# Step (22)
# Pool destroy
if virsh.pool_destroy(pool_name):
logging.debug("Pool %s destroyed.", pool_name)
else:
test.fail("Destroy pool % failed." % pool_name)
# Step (23)
# Pool delete for 'dir' type pool
if pool_type == "dir":
for f in os.listdir(pool_target):
os.remove(os.path.join(pool_target, f))
result = virsh.pool_delete(pool_name, ignore_status=True)
utlv.check_exit_status(result)
option = "--inactive --type %s" % pool_type
check_pool_list(pool_name, option)
if os.path.exists(pool_target):
test.fail("The target path '%s' still exist." %
pool_target)
result = virsh.pool_start(pool_name,
ignore_status=True)
utlv.check_exit_status(result, True)
# Step (24)
# Pool undefine
result = virsh.pool_undefine(pool_name, ignore_status=True)
utlv.check_exit_status(result)
check_pool_list(pool_name, "--all", True)
finally:
# Clean up
try:
pvt.cleanup_pool(**params)
utlv.setup_or_cleanup_iscsi(False)
except exceptions.TestFail as detail:
logging.error(str(detail))
if multipathd_status:
multipathd.start()
if os.path.exists(poolxml):
os.remove(poolxml)
def run(test, params, env):
"""
Test the virsh pool commands
(1) Define a given type pool
(2) List pool with '--inactive --type' options
(3) Dumpxml for the pool
(4) Undefine the pool
(5) Define pool by using the XML file in step (3)
(6) Build the pool(except 'disk' type pool
For 'fs' type pool, cover --overwrite and --no-overwrite options
(7) Start the pool
(8) List pool with '--persistent --type' options
(9) Mark pool autostart
(10) List pool with '--autostart --type' options
(11) Restart libvirtd and list pool with '--autostart --persistent' options
(12) Destroy the pool
(13) Unmark pool autostart
(14) Repeat step (11)
(15) Start the pool
(16) Get pool info
(17) Get pool uuid by name
(18) Get pool name by uuid
(19) Refresh the pool
For 'dir' type pool, touch a file under target path and refresh again
to make the new file show in vol-list.
(20) Check pool 'Capacity', 'Allocation' and 'Available'
Create a over size vol in pool(expect fail), then check these values
(21) Undefine the pool, and this should fail as pool is still active
(22) Destroy the pool
(23) Delete pool for 'dir' type pool. After the command, the pool object
will still exist but target path will be deleted
(24) Undefine the pool
"""
# Initialize the variables
pool_name = params.get("pool_name", "temp_pool_1")
pool_type = params.get("pool_type", "dir")
pool_target = params.get("pool_target", "")
source_format = params.get("source_format", "")
source_name = params.get("pool_source_name", "gluster-vol1")
source_path = params.get("pool_source_path", "/")
# The file for dumped pool xml
pool_xml = os.path.join(test.tmpdir, "pool.xml.tmp")
if os.path.dirname(pool_target) is "":
pool_target = os.path.join(test.tmpdir, pool_target)
vol_name = params.get("vol_name", "temp_vol_1")
# Use pool name as VG name
status_error = "yes" == params.get("status_error", "no")
vol_path = os.path.join(pool_target, vol_name)
ip_protocal = params.get('ip_protocal', 'ipv4')
if not libvirt_version.version_compare(1, 0, 0):
if pool_type == "gluster":
raise error.TestNAError("Gluster pool is not supported in current"
" libvirt version.")
def check_pool_list(pool_name, option="--all", expect_error=False):
"""
Check pool by running pool-list command with given option.
:param pool_name: Name of the pool
:param option: option for pool-list command
:param expect_error: Boolean value, expect command success or fail
"""
found = False
# Get the list stored in a variable
result = virsh.pool_list(option, ignore_status=True)
utlv.check_exit_status(result, False)
output = re.findall(r"(\S+)\ +(\S+)\ +(\S+)[\ +\n]",
str(result.stdout))
for item in output:
if pool_name in item[0]:
found = True
break
if found:
logging.debug("Find pool '%s' in pool list.", pool_name)
else:
logging.debug("Not find pool %s in pool list.", pool_name)
if expect_error and found:
raise error.TestFail("Unexpect pool '%s' exist." % pool_name)
if not expect_error and not found:
raise error.TestFail("Expect pool '%s' doesn't exist." % pool_name)
def check_vol_list(vol_name, pool_name):
"""
Check volume from the list
:param vol_name: Name of the volume
:param pool_name: Name of the pool
"""
found = False
# Get the volume list stored in a variable
result = virsh.vol_list(pool_name, ignore_status=True)
utlv.check_exit_status(result)
output = re.findall(r"(\S+)\ +(\S+)[\ +\n]", str(result.stdout))
for item in output:
if vol_name in item[0]:
found = True
break
if found:
logging.debug("Find volume '%s' in pool '%s'.", vol_name,
pool_name)
else:
raise error.TestFail("Not find volume '%s' in pool '%s'." %
(vol_name, pool_name))
def is_in_range(actual, expected, error_percent):
deviation = 100 - (100 * (float(actual) / float(expected)))
logging.debug("Deviation: %0.2f%%", float(deviation))
return float(deviation) <= float(error_percent)
def check_pool_info(pool_info, check_point, value):
"""
Check the pool name, uuid, etc.
:param pool_info: A dict include pool's information
:param key: Key of pool info dict, available value: Name, UUID, State
Persistent, Autostart, Capacity, Allocation, Available
:param value: Expect value of pool_info[key]
"""
if pool_info is None:
raise error.TestFail("Pool info dictionary is needed.")
val_tup = ('Capacity', 'Allocation', 'Available')
if check_point in val_tup and float(value.split()[0]):
# As from bytes to GiB, could cause deviation, and it should not
# exceed 1 percent.
if is_in_range(float(pool_info[check_point].split()[0]),
float(value.split()[0]), 1):
logging.debug("Pool '%s' is '%s'.", check_point, value)
else:
raise error.TestFail("Pool '%s' isn't '%s'." %
(check_point, value))
else:
if pool_info[check_point] == value:
logging.debug("Pool '%s' is '%s'.", check_point, value)
else:
raise error.TestFail("Pool '%s' isn't '%s'." %
(check_point, value))
# Stop multipathd to avoid start pool fail(For fs like pool, the new add
# disk may in use by device-mapper, so start pool will report disk already
# mounted error).
multipathd = service.Factory.create_service("multipathd")
multipathd_status = multipathd.status()
if multipathd_status:
multipathd.stop()
# Run Testcase
pvt = utlv.PoolVolumeTest(test, params)
emulated_image = "emulated-image"
kwargs = {
'image_size': '1G',
'pre_disk_vol': ['1M'],
'source_name': source_name,
'source_path': source_path,
'source_format': source_format,
'persistent': True,
'ip_protocal': ip_protocal
}
try:
_pool = libvirt_storage.StoragePool()
# Step (1)
# Pool define
pvt.pre_pool(pool_name, pool_type, pool_target, emulated_image,
**kwargs)
# Step (2)
# Pool list
option = "--inactive --type %s" % pool_type
check_pool_list(pool_name, option)
# Step (3)
# Pool dumpxml
xml = virsh.pool_dumpxml(pool_name, to_file=pool_xml)
logging.debug("Pool '%s' XML:\n%s", pool_name, xml)
# Step (4)
# Undefine pool
result = virsh.pool_undefine(pool_name, ignore_status=True)
utlv.check_exit_status(result)
check_pool_list(pool_name, "--all", True)
# Step (5)
# Define pool from XML file
result = virsh.pool_define(pool_xml)
utlv.check_exit_status(result, status_error)
# Step (6)
# Buid pool, this step may fail for 'disk' and 'logical' types pool
if pool_type not in ["disk", "logical"]:
option = ""
# Options --overwrite and --no-overwrite can only be used to
# build a filesystem pool, but it will fail for now
# if pool_type == "fs":
# option = '--overwrite'
result = virsh.pool_build(pool_name, option, ignore_status=True)
utlv.check_exit_status(result)
# Step (7)
# Pool start
result = virsh.pool_start(pool_name, ignore_status=True)
utlv.check_exit_status(result)
# Step (8)
# Pool list
option = "--persistent --type %s" % pool_type
check_pool_list(pool_name, option)
# Step (9)
# Pool autostart
result = virsh.pool_autostart(pool_name, ignore_status=True)
utlv.check_exit_status(result)
# Step (10)
# Pool list
option = "--autostart --type %s" % pool_type
check_pool_list(pool_name, option)
# Step (11)
# Restart libvirtd and check the autostart pool
utils_libvirtd.libvirtd_restart()
option = "--autostart --persistent"
check_pool_list(pool_name, option)
# Step (12)
# Pool destroy
if virsh.pool_destroy(pool_name):
logging.debug("Pool %s destroyed.", pool_name)
else:
raise error.TestFail("Destroy pool % failed." % pool_name)
# Step (13)
# Pool autostart disable
result = virsh.pool_autostart(pool_name,
"--disable",
ignore_status=True)
utlv.check_exit_status(result)
# Step (14)
# Repeat step (11)
utils_libvirtd.libvirtd_restart()
option = "--autostart"
check_pool_list(pool_name, option, True)
# Step (15)
# Pool start
# When libvirtd starts up, it'll check to see if any of the storage
# pools have been activated externally. If so, then it'll mark the
# pool as active. This is independent of autostart.
# So a directory based storage pool is thus pretty much always active,
# and so as the SCSI pool.
if pool_type not in ["dir", 'scsi']:
result = virsh.pool_start(pool_name, ignore_status=True)
utlv.check_exit_status(result)
# Step (16)
# Pool info
pool_info = _pool.pool_info(pool_name)
logging.debug("Pool '%s' info:\n%s", pool_name, pool_info)
# Step (17)
# Pool UUID
result = virsh.pool_uuid(pool_info["Name"], ignore_status=True)
utlv.check_exit_status(result)
check_pool_info(pool_info, "UUID", result.stdout.strip())
# Step (18)
# Pool Name
result = virsh.pool_name(pool_info["UUID"], ignore_status=True)
utlv.check_exit_status(result)
check_pool_info(pool_info, "Name", result.stdout.strip())
# Step (19)
# Pool refresh for 'dir' type pool
if pool_type == "dir":
os.mknod(vol_path)
result = virsh.pool_refresh(pool_name)
utlv.check_exit_status(result)
check_vol_list(vol_name, pool_name)
# Step (20)
# Create an over size vol in pool(expect fail), then check pool:
# 'Capacity', 'Allocation' and 'Available'
# For NFS type pool, there's a bug(BZ#1077068) about allocate volume,
# and glusterfs pool not support create volume, so not test them
if pool_type != "netfs":
vol_capacity = "10000G"
vol_allocation = "10000G"
result = virsh.vol_create_as("oversize_vol", pool_name,
vol_capacity, vol_allocation, "raw")
utlv.check_exit_status(result, True)
new_info = _pool.pool_info(pool_name)
check_pool_info(pool_info, "Capacity", new_info['Capacity'])
check_pool_info(pool_info, "Allocation", new_info['Allocation'])
check_pool_info(pool_info, "Available", new_info['Available'])
# Step (21)
# Undefine pool, this should fail as the pool is active
result = virsh.pool_undefine(pool_name, ignore_status=True)
utlv.check_exit_status(result, expect_error=True)
check_pool_list(pool_name, "", False)
# Step (22)
# Pool destroy
if virsh.pool_destroy(pool_name):
logging.debug("Pool %s destroyed.", pool_name)
else:
raise error.TestFail("Destroy pool % failed." % pool_name)
# Step (23)
# Pool delete for 'dir' type pool
if pool_type == "dir":
for f in os.listdir(pool_target):
os.remove(os.path.join(pool_target, f))
result = virsh.pool_delete(pool_name, ignore_status=True)
utlv.check_exit_status(result)
option = "--inactive --type %s" % pool_type
check_pool_list(pool_name, option)
if os.path.exists(pool_target):
raise error.TestFail("The target path '%s' still exist." %
pool_target)
result = virsh.pool_start(pool_name, ignore_status=True)
utlv.check_exit_status(result, True)
# Step (24)
# Pool undefine
result = virsh.pool_undefine(pool_name, ignore_status=True)
utlv.check_exit_status(result)
check_pool_list(pool_name, "--all", True)
finally:
# Clean up
try:
pvt.cleanup_pool(pool_name, pool_type, pool_target, emulated_image,
**kwargs)
except error.TestFail, detail:
logging.error(str(detail))
if multipathd_status:
multipathd.start()
if os.path.exists(pool_xml):
os.remove(pool_xml)
def run(test, params, env):
"""
Test the virsh pool commands
(1) Define a given type pool
(2) List pool with '--inactive --type' options
(3) Dumpxml for the pool
(4) Undefine the pool
(5) Define pool by using the XML file in step (3)
(6) Build the pool(except 'disk' type pool
For 'fs' type pool, cover --overwrite and --no-overwrite options
(7) Start the pool
(8) List pool with '--persistent --type' options
(9) Mark pool autostart
(10) List pool with '--autostart --type' options
(11) Restart libvirtd and list pool with '--autostart --persistent' options
(12) Destroy the pool
(13) Unmark pool autostart
(14) Repeat step (11)
(15) Start the pool
(16) Get pool info
(17) Get pool uuid by name
(18) Get pool name by uuid
(19) Refresh the pool
For 'dir' type pool, touch a file under target path and refresh again
to make the new file show in vol-list.
(20) Destroy the pool
(21) Delete pool for 'dir' type pool. After the command, the pool object
will still exist but target path will be deleted
(22) Undefine the pool
"""
# Initialize the variables
pool_name = params.get("pool_name", "temp_pool_1")
pool_type = params.get("pool_type", "dir")
pool_target = params.get("pool_target", "")
# The file for dumped pool xml
pool_xml = os.path.join(test.tmpdir, "pool.xml.tmp")
if os.path.dirname(pool_target) is "":
pool_target = os.path.join(test.tmpdir, pool_target)
vol_name = params.get("vol_name", "temp_vol_1")
# Use pool name as VG name
vg_name = pool_name
status_error = "yes" == params.get("status_error", "no")
vol_path = os.path.join(pool_target, vol_name)
# Clean up flags:
# cleanup_env[0] for nfs, cleanup_env[1] for iscsi, cleanup_env[2] for lvm
cleanup_env = [False, False, False]
def check_exit_status(result, expect_error=False):
"""
Check the exit status of virsh commands.
:param result: Virsh command result object
:param expect_error: Boolean value, expect command success or fail
"""
if not expect_error:
if result.exit_status != 0:
raise error.TestFail(result.stderr)
else:
logging.debug("Command output:\n%s", result.stdout.strip())
elif expect_error and result.exit_status == 0:
raise error.TestFail("Expect fail, but run successfully.")
def check_pool_list(pool_name, option="--all", expect_error=False):
"""
Check pool by running pool-list command with given option.
:param pool_name: Name of the pool
:param option: option for pool-list command
:param expect_error: Boolean value, expect command success or fail
"""
found = False
# Get the list stored in a variable
result = virsh.pool_list(option, ignore_status=True)
check_exit_status(result, False)
output = re.findall(r"(\S+)\ +(\S+)\ +(\S+)[\ +\n]",
str(result.stdout))
for item in output:
if pool_name in item[0]:
found = True
break
if found:
logging.debug("Find pool '%s' in pool list.", pool_name)
else:
logging.debug("Not find pool %s in pool list.", pool_name)
if expect_error and found:
raise error.TestFail("Unexpect pool '%s' exist." % pool_name)
if not expect_error and not found:
raise error.TestFail("Expect pool '%s' doesn't exist." % pool_name)
def check_vol_list(vol_name, pool_name):
"""
Check volume from the list
:param vol_name: Name of the volume
:param pool_name: Name of the pool
"""
found = False
# Get the volume list stored in a variable
result = virsh.vol_list(pool_name, ignore_status=True)
check_exit_status(result)
output = re.findall(r"(\S+)\ +(\S+)[\ +\n]", str(result.stdout))
for item in output:
if vol_name in item[0]:
found = True
break
if found:
logging.debug(
"Find volume '%s' in pool '%s'.", vol_name, pool_name)
else:
raise error.TestFail(
"Not find volume '%s' in pool '%s'." %
(vol_name, pool_name))
def check_pool_info(pool_info, check_point, value):
"""
Check the pool name, uuid, etc.
:param pool_info: A dict include pool's information
:param key: Key of pool info dict, available value: Name, UUID, State
Persistent, Autostart, Capacity, Allocation, Available
:param value: Expect value of pool_info[key]
"""
if pool_info is None:
raise error.TestFail("Pool info dictionary is needed.")
if pool_info[check_point] == value:
logging.debug("Pool '%s' is '%s'.", check_point, value)
else:
raise error.TestFail("Pool '%s' isn't '%s'." % (check_point,
value))
# Run Testcase
try:
_pool = libvirt_storage.StoragePool()
# Step (1)
# Pool define
result = utils_test.libvirt.define_pool(pool_name, pool_type,
pool_target, cleanup_env)
check_exit_status(result, status_error)
# Step (2)
# Pool list
option = "--inactive --type %s" % pool_type
check_pool_list(pool_name, option)
# Step (3)
# Pool dumpxml
xml = virsh.pool_dumpxml(pool_name, to_file=pool_xml)
logging.debug("Pool '%s' XML:\n%s", pool_name, xml)
# Step (4)
# Undefine pool
result = virsh.pool_undefine(pool_name, ignore_status=True)
check_exit_status(result)
check_pool_list(pool_name, "--all", True)
# Step (5)
# Define pool from XML file
result = virsh.pool_define(pool_xml)
check_exit_status(result, status_error)
# Step (6)
# Buid pool, this step may fail for 'disk' and 'logical' types pool
if pool_type not in ["disk", "logical"]:
option = ""
# Options --overwrite and --no-overwrite can only be used to
# build a filesystem pool, but it will fail for now
# if pool_type == "fs":
# option = '--overwrite'
result = virsh.pool_build(pool_name, option, ignore_status=True)
check_exit_status(result)
# Step (7)
# Pool start
result = virsh.pool_start(pool_name, ignore_status=True)
check_exit_status(result)
# Step (8)
# Pool list
option = "--persistent --type %s" % pool_type
check_pool_list(pool_name, option)
# Step (9)
# Pool autostart
result = virsh.pool_autostart(pool_name, ignore_status=True)
check_exit_status(result)
# Step (10)
# Pool list
option = "--autostart --type %s" % pool_type
check_pool_list(pool_name, option)
# Step (11)
# Restart libvirtd and check the autostart pool
utils_libvirtd.libvirtd_restart()
option = "--autostart --persistent"
check_pool_list(pool_name, option)
# Step (12)
# Pool destroy
if virsh.pool_destroy(pool_name):
logging.debug("Pool %s destroyed.", pool_name)
else:
raise error.TestFail("Destroy pool % failed." % pool_name)
# Step (13)
# Pool autostart disable
result = virsh.pool_autostart(pool_name, "--disable",
ignore_status=True)
check_exit_status(result)
# Step (14)
# Repeat step (11)
utils_libvirtd.libvirtd_restart()
option = "--autostart"
check_pool_list(pool_name, option, True)
# Step (15)
# Pool start
# If the filesystem cntaining the directory is mounted, then the
# directory will show as running, which means the local 'dir' pool
# don't need start after restart libvirtd
if pool_type != "dir":
result = virsh.pool_start(pool_name, ignore_status=True)
check_exit_status(result)
# Step (16)
# Pool info
pool_info = _pool.pool_info(pool_name)
logging.debug("Pool '%s' info:\n%s", pool_name, pool_info)
# Step (17)
# Pool UUID
result = virsh.pool_uuid(pool_info["Name"], ignore_status=True)
check_exit_status(result)
check_pool_info(pool_info, "UUID", result.stdout.strip())
# Step (18)
# Pool Name
result = virsh.pool_name(pool_info["UUID"], ignore_status=True)
check_exit_status(result)
check_pool_info(pool_info, "Name", result.stdout.strip())
# Step (19)
# Pool refresh for 'dir' type pool
if pool_type == "dir":
os.mknod(vol_path)
result = virsh.pool_refresh(pool_name)
check_exit_status(result)
check_vol_list(vol_name, pool_name)
# Step(20)
# Pool destroy
if virsh.pool_destroy(pool_name):
logging.debug("Pool %s destroyed.", pool_name)
else:
raise error.TestFail("Destroy pool % failed." % pool_name)
# Step (21)
# Pool delete for 'dir' type pool
if pool_type == "dir":
if os.path.exists(vol_path):
os.remove(vol_path)
result = virsh.pool_delete(pool_name, ignore_status=True)
check_exit_status(result)
option = "--inactive --type %s" % pool_type
check_pool_list(pool_name, option)
if os.path.exists(pool_target):
raise error.TestFail("The target path '%s' still exist." %
pool_target)
result = virsh.pool_start(pool_name, ignore_status=True)
check_exit_status(result, True)
# Step (22)
# Pool undefine
result = virsh.pool_undefine(pool_name, ignore_status=True)
check_exit_status(result)
check_pool_list(pool_name, "--all", True)
finally:
# Clean up
if os.path.exists(pool_xml):
os.remove(pool_xml)
if not _pool.delete_pool(pool_name):
logging.error("Can't delete pool: %s", pool_name)
if cleanup_env[2]:
cmd = "pvs |grep %s |awk '{print $1}'" % vg_name
pv_name = utils.system_output(cmd)
lv_utils.vg_remove(vg_name)
utils.run("pvremove %s" % pv_name)
if cleanup_env[1]:
utils_test.libvirt.setup_or_cleanup_iscsi(False)
if cleanup_env[0]:
utils_test.libvirt.setup_or_cleanup_nfs(False)
def run(test, params, env):
"""
Test the virsh pool commands
(1) Define a given type pool
(2) List pool with '--inactive --type' options
(3) Dumpxml for the pool
(4) Undefine the pool
(5) Define pool by using the XML file in step (3)
(6) Build the pool(except 'disk' type pool
For 'fs' type pool, cover --overwrite and --no-overwrite options
(7) Start the pool
(8) List pool with '--persistent --type' options
(9) Mark pool autostart
(10) List pool with '--autostart --type' options
(11) Restart libvirtd and list pool with '--autostart --persistent' options
(12) Destroy the pool
(13) Unmark pool autostart
(14) Repeat step (11)
(15) Start the pool
(16) Get pool info
(17) Get pool uuid by name
(18) Get pool name by uuid
(19) Refresh the pool
For 'dir' type pool, touch a file under target path and refresh again
to make the new file show in vol-list.
(20) Check pool 'Capacity', 'Allocation' and 'Available'
Create a over size vol in pool(expect fail), then check these values
(21) Undefine the pool, and this should fail as pool is still active
(22) Destroy the pool
(23) Delete pool for 'dir' type pool. After the command, the pool object
will still exist but target path will be deleted
(24) Undefine the pool
"""
# Initialize the variables
pool_name = params.get("pool_name", "temp_pool_1")
pool_type = params.get("pool_type", "dir")
pool_target = params.get("pool_target", "")
source_format = params.get("source_format", "")
source_name = params.get("pool_source_name", "gluster-vol1")
source_path = params.get("pool_source_path", "/")
# The file for dumped pool xml
pool_xml = os.path.join(test.tmpdir, "pool.xml.tmp")
if os.path.dirname(pool_target) is "":
pool_target = os.path.join(test.tmpdir, pool_target)
vol_name = params.get("vol_name", "temp_vol_1")
# Use pool name as VG name
status_error = "yes" == params.get("status_error", "no")
vol_path = os.path.join(pool_target, vol_name)
ip_protocal = params.get("ip_protocal", "ipv4")
if not libvirt_version.version_compare(1, 0, 0):
if pool_type == "gluster":
raise error.TestNAError("Gluster pool is not supported in current" " libvirt version.")
def check_exit_status(result, expect_error=False):
"""
Check the exit status of virsh commands.
:param result: Virsh command result object
:param expect_error: Boolean value, expect command success or fail
"""
if not expect_error:
if result.exit_status != 0:
raise error.TestFail(result.stderr)
else:
logging.debug("Command output:\n%s", result.stdout.strip())
elif expect_error and result.exit_status == 0:
raise error.TestFail("Expect fail, but run successfully.")
def check_pool_list(pool_name, option="--all", expect_error=False):
"""
Check pool by running pool-list command with given option.
:param pool_name: Name of the pool
:param option: option for pool-list command
:param expect_error: Boolean value, expect command success or fail
"""
found = False
# Get the list stored in a variable
result = virsh.pool_list(option, ignore_status=True)
check_exit_status(result, False)
output = re.findall(r"(\S+)\ +(\S+)\ +(\S+)[\ +\n]", str(result.stdout))
for item in output:
if pool_name in item[0]:
found = True
break
if found:
logging.debug("Find pool '%s' in pool list.", pool_name)
else:
logging.debug("Not find pool %s in pool list.", pool_name)
if expect_error and found:
raise error.TestFail("Unexpect pool '%s' exist." % pool_name)
if not expect_error and not found:
raise error.TestFail("Expect pool '%s' doesn't exist." % pool_name)
def check_vol_list(vol_name, pool_name):
"""
Check volume from the list
:param vol_name: Name of the volume
:param pool_name: Name of the pool
"""
found = False
# Get the volume list stored in a variable
result = virsh.vol_list(pool_name, ignore_status=True)
check_exit_status(result)
output = re.findall(r"(\S+)\ +(\S+)[\ +\n]", str(result.stdout))
for item in output:
if vol_name in item[0]:
found = True
break
if found:
logging.debug("Find volume '%s' in pool '%s'.", vol_name, pool_name)
else:
raise error.TestFail("Not find volume '%s' in pool '%s'." % (vol_name, pool_name))
def check_pool_info(pool_info, check_point, value):
"""
Check the pool name, uuid, etc.
:param pool_info: A dict include pool's information
:param key: Key of pool info dict, available value: Name, UUID, State
Persistent, Autostart, Capacity, Allocation, Available
:param value: Expect value of pool_info[key]
"""
if pool_info is None:
raise error.TestFail("Pool info dictionary is needed.")
if pool_info[check_point] == value:
logging.debug("Pool '%s' is '%s'.", check_point, value)
else:
raise error.TestFail("Pool '%s' isn't '%s'." % (check_point, value))
# Stop multipathd to avoid start pool fail(For fs like pool, the new add
# disk may in use by device-mapper, so start pool will report disk already
# mounted error).
multipathd = service.Factory.create_service("multipathd")
multipathd_status = multipathd.status()
if multipathd_status:
multipathd.stop()
# Run Testcase
pvt = utlv.PoolVolumeTest(test, params)
emulated_image = "emulated-image"
kwargs = {
"image_size": "1G",
"pre_disk_vol": ["1M"],
"source_name": source_name,
"source_path": source_path,
"source_format": source_format,
"persistent": True,
"ip_protocal": ip_protocal,
}
try:
_pool = libvirt_storage.StoragePool()
# Step (1)
# Pool define
pvt.pre_pool(pool_name, pool_type, pool_target, emulated_image, **kwargs)
# Step (2)
# Pool list
option = "--inactive --type %s" % pool_type
check_pool_list(pool_name, option)
# Step (3)
# Pool dumpxml
xml = virsh.pool_dumpxml(pool_name, to_file=pool_xml)
logging.debug("Pool '%s' XML:\n%s", pool_name, xml)
# Step (4)
# Undefine pool
result = virsh.pool_undefine(pool_name, ignore_status=True)
check_exit_status(result)
check_pool_list(pool_name, "--all", True)
# Step (5)
# Define pool from XML file
result = virsh.pool_define(pool_xml)
check_exit_status(result, status_error)
# Step (6)
# Buid pool, this step may fail for 'disk' and 'logical' types pool
if pool_type not in ["disk", "logical"]:
option = ""
# Options --overwrite and --no-overwrite can only be used to
# build a filesystem pool, but it will fail for now
# if pool_type == "fs":
# option = '--overwrite'
result = virsh.pool_build(pool_name, option, ignore_status=True)
check_exit_status(result)
# Step (7)
# Pool start
result = virsh.pool_start(pool_name, ignore_status=True)
check_exit_status(result)
# Step (8)
# Pool list
option = "--persistent --type %s" % pool_type
check_pool_list(pool_name, option)
# Step (9)
# Pool autostart
result = virsh.pool_autostart(pool_name, ignore_status=True)
check_exit_status(result)
# Step (10)
# Pool list
option = "--autostart --type %s" % pool_type
check_pool_list(pool_name, option)
# Step (11)
# Restart libvirtd and check the autostart pool
utils_libvirtd.libvirtd_restart()
option = "--autostart --persistent"
check_pool_list(pool_name, option)
# Step (12)
# Pool destroy
if virsh.pool_destroy(pool_name):
logging.debug("Pool %s destroyed.", pool_name)
else:
raise error.TestFail("Destroy pool % failed." % pool_name)
# Step (13)
# Pool autostart disable
result = virsh.pool_autostart(pool_name, "--disable", ignore_status=True)
check_exit_status(result)
# Step (14)
# Repeat step (11)
utils_libvirtd.libvirtd_restart()
option = "--autostart"
check_pool_list(pool_name, option, True)
# Step (15)
# Pool start
# When libvirtd starts up, it'll check to see if any of the storage
# pools have been activated externally. If so, then it'll mark the
# pool as active. This is independent of autostart.
# So a directory based storage pool is thus pretty much always active,
# and so as the SCSI pool.
if pool_type != "scsi" and (pool_type != "dir" or libvirt_version.version_compare(1, 2, 15)):
result = virsh.pool_start(pool_name, ignore_status=True)
check_exit_status(result)
# Step (16)
# Pool info
pool_info = _pool.pool_info(pool_name)
logging.debug("Pool '%s' info:\n%s", pool_name, pool_info)
# Step (17)
# Pool UUID
result = virsh.pool_uuid(pool_info["Name"], ignore_status=True)
check_exit_status(result)
check_pool_info(pool_info, "UUID", result.stdout.strip())
# Step (18)
# Pool Name
result = virsh.pool_name(pool_info["UUID"], ignore_status=True)
check_exit_status(result)
check_pool_info(pool_info, "Name", result.stdout.strip())
# Step (19)
# Pool refresh for 'dir' type pool
if pool_type == "dir":
os.mknod(vol_path)
result = virsh.pool_refresh(pool_name)
check_exit_status(result)
check_vol_list(vol_name, pool_name)
# Step (20)
# Create an over size vol in pool(expect fail), then check pool:
# 'Capacity', 'Allocation' and 'Available'
# For NFS type pool, there's a bug(BZ#1077068) about allocate volume,
# and glusterfs pool not support create volume, so not test them
if pool_type != "netfs":
vol_capacity = "10000G"
vol_allocation = "10000G"
result = virsh.vol_create_as("oversize_vol", pool_name, vol_capacity, vol_allocation, "raw")
check_exit_status(result, True)
new_info = _pool.pool_info(pool_name)
check_pool_info(pool_info, "Capacity", new_info["Capacity"])
check_pool_info(pool_info, "Allocation", new_info["Allocation"])
check_pool_info(pool_info, "Available", new_info["Available"])
# Step (21)
# Undefine pool, this should fail as the pool is active
result = virsh.pool_undefine(pool_name, ignore_status=True)
check_exit_status(result, expect_error=True)
check_pool_list(pool_name, "", False)
# Step (22)
# Pool destroy
if virsh.pool_destroy(pool_name):
logging.debug("Pool %s destroyed.", pool_name)
else:
raise error.TestFail("Destroy pool % failed." % pool_name)
# Step (23)
# Pool delete for 'dir' type pool
if pool_type == "dir":
for f in os.listdir(pool_target):
os.remove(os.path.join(pool_target, f))
result = virsh.pool_delete(pool_name, ignore_status=True)
check_exit_status(result)
option = "--inactive --type %s" % pool_type
check_pool_list(pool_name, option)
if os.path.exists(pool_target):
raise error.TestFail("The target path '%s' still exist." % pool_target)
result = virsh.pool_start(pool_name, ignore_status=True)
check_exit_status(result, True)
# Step (24)
# Pool undefine
result = virsh.pool_undefine(pool_name, ignore_status=True)
check_exit_status(result)
check_pool_list(pool_name, "--all", True)
finally:
# Clean up
try:
pvt.cleanup_pool(pool_name, pool_type, pool_target, emulated_image, **kwargs)
except error.TestFail, detail:
logging.error(str(detail))
if multipathd_status:
multipathd.start()
if os.path.exists(pool_xml):
os.remove(pool_xml)
def run(test, params, env):
"""
Test the virsh pool commands
(1) Define a given type pool
(2) List pool with '--inactive --type' options
(3) Dumpxml for the pool
(4) Undefine the pool
(5) Define pool by using the XML file in step (3)
(6) Build the pool(except 'disk' type pool
For 'fs' type pool, cover --overwrite and --no-overwrite options
(7) Start the pool
(8) List pool with '--persistent --type' options
(9) Mark pool autostart
(10) List pool with '--autostart --type' options
(11) Restart libvirtd and list pool with '--autostart --persistent' options
(12) Destroy the pool
(13) Unmark pool autostart
(14) Repeat step (11)
(15) Start the pool
(16) Get pool info
(17) Get pool uuid by name
(18) Get pool name by uuid
(19) Refresh the pool
For 'dir' type pool, touch a file under target path and refresh again
to make the new file show in vol-list.
(20) Check pool 'Capacity', 'Allocation' and 'Available'
Create a over size vol in pool(expect fail), then check these values
(21) Undefine the pool, and this should fail as pool is still active
(22) Destroy the pool
(23) Delete pool for 'dir' type pool. After the command, the pool object
will still exist but target path will be deleted
(24) Undefine the pool
"""
# Initialize the variables
pool_name = params.get("pool_name", "temp_pool_1")
pool_type = params.get("pool_type", "dir")
pool_target = params.get("pool_target", "")
source_format = params.get("source_format", "")
source_name = params.get("pool_source_name", "gluster-vol1")
source_path = params.get("pool_source_path", "/")
new_pool_name = params.get("new_pool_name", "")
build_option = params.get("build_option", "")
same_source_test = "yes" == params.get("same_source_test", "no")
customize_initiator_iqn = "yes" == params.get("customize_initiator_iqn", "no")
# The file for dumped pool xml
poolxml = os.path.join(data_dir.get_tmp_dir(), "pool.xml.tmp")
if os.path.dirname(pool_target) is "":
pool_target = os.path.join(data_dir.get_tmp_dir(), pool_target)
vol_name = params.get("vol_name", "temp_vol_1")
# Use pool name as VG name
status_error = "yes" == params.get("status_error", "no")
vol_path = os.path.join(pool_target, vol_name)
ip_protocal = params.get('ip_protocal', 'ipv4')
if not libvirt_version.version_compare(1, 0, 0):
if pool_type == "gluster":
test.cancel("Gluster pool is not supported in current"
" libvirt version.")
def check_pool_list(pool_name, option="--all", expect_error=False):
"""
Check pool by running pool-list command with given option.
:param pool_name: Name of the pool
:param option: option for pool-list command
:param expect_error: Boolean value, expect command success or fail
"""
found = False
# Get the list stored in a variable
result = virsh.pool_list(option, ignore_status=True)
utlv.check_exit_status(result, False)
output = re.findall(r"(\S+)\ +(\S+)\ +(\S+)",
str(result.stdout.strip()))
for item in output:
if pool_name in item[0]:
found = True
break
if found:
logging.debug("Find pool '%s' in pool list.", pool_name)
else:
logging.debug("Not find pool %s in pool list.", pool_name)
if expect_error and found:
test.fail("Unexpect pool '%s' exist." % pool_name)
if not expect_error and not found:
test.fail("Expect pool '%s' doesn't exist." % pool_name)
def check_vol_list(vol_name, pool_name):
"""
Check volume from the list
:param vol_name: Name of the volume
:param pool_name: Name of the pool
"""
found = False
# Get the volume list stored in a variable
result = virsh.vol_list(pool_name, ignore_status=True)
utlv.check_exit_status(result)
output = re.findall(r"(\S+)\ +(\S+)", str(result.stdout.strip()))
for item in output:
if vol_name in item[0]:
found = True
break
if found:
logging.debug(
"Find volume '%s' in pool '%s'.", vol_name, pool_name)
else:
test.fail(
"Not find volume '%s' in pool '%s'." %
(vol_name, pool_name))
def is_in_range(actual, expected, error_percent):
deviation = 100 - (100 * (float(actual) / float(expected)))
logging.debug("Deviation: %0.2f%%", float(deviation))
return float(deviation) <= float(error_percent)
def check_pool_info(pool_info, check_point, value):
"""
Check the pool name, uuid, etc.
:param pool_info: A dict include pool's information
:param key: Key of pool info dict, available value: Name, UUID, State
Persistent, Autostart, Capacity, Allocation, Available
:param value: Expect value of pool_info[key]
"""
if pool_info is None:
test.fail("Pool info dictionary is needed.")
val_tup = ('Capacity', 'Allocation', 'Available')
if check_point in val_tup and float(value.split()[0]):
# As from bytes to GiB, could cause deviation, and it should not
# exceed 1 percent.
if is_in_range(float(pool_info[check_point].split()[0]),
float(value.split()[0]), 1):
logging.debug("Pool '%s' is '%s'.", check_point, value)
else:
test.fail("Pool '%s' isn't '%s'." %
(check_point, value))
else:
if pool_info[check_point] == value:
logging.debug("Pool '%s' is '%s'.", check_point, value)
else:
test.fail("Pool '%s' isn't '%s'." %
(check_point, value))
# Stop multipathd to avoid start pool fail(For fs like pool, the new add
# disk may in use by device-mapper, so start pool will report disk already
# mounted error).
multipathd = service.Factory.create_service("multipathd")
multipathd_status = multipathd.status()
if multipathd_status:
multipathd.stop()
# Run Testcase
pvt = utlv.PoolVolumeTest(test, params)
emulated_image = "emulated-image"
kwargs = {'image_size': '1G', 'pre_disk_vol': ['100M'],
'source_name': source_name, 'source_path': source_path,
'source_format': source_format, 'persistent': True,
'ip_protocal': ip_protocal}
try:
_pool = libvirt_storage.StoragePool()
# Step (1)
# Pool define
pvt.pre_pool(pool_name, pool_type, pool_target, emulated_image,
**kwargs)
# Step (2)
# Pool list
option = "--inactive --type %s" % pool_type
check_pool_list(pool_name, option)
# Step (3)
# Pool dumpxml
xml = virsh.pool_dumpxml(pool_name, to_file=poolxml)
logging.debug("Pool '%s' XML:\n%s", pool_name, xml)
# Update pool name
if new_pool_name:
if "/" in new_pool_name:
new_pool_name = new_pool_name.replace("/", "\/")
logging.debug(new_pool_name)
p_xml = pool_xml.PoolXML.new_from_dumpxml(pool_name)
p_xml.name = new_pool_name
del p_xml.uuid
poolxml = p_xml.xml
logging.debug("XML after update pool name:\n%s" % p_xml)
# Update host name
if same_source_test:
s_xml = p_xml.get_source()
s_xml.host_name = "192.168.1.1"
p_xml.set_source(s_xml)
poolxml = p_xml.xml
logging.debug("XML after update host name:\n%s" % p_xml)
if customize_initiator_iqn:
initiator_iqn = params.get("initiator_iqn",
"iqn.2018-07.com.virttest:pool.target")
p_xml = pool_xml.PoolXML.new_from_dumpxml(pool_name)
s_node = p_xml.xmltreefile.find('/source')
i_node = ET.SubElement(s_node, 'initiator')
ET.SubElement(i_node, 'iqn', {'name': initiator_iqn})
p_xml.xmltreefile.write()
poolxml = p_xml.xml
logging.debug('XML after add Multi-IQN:\n%s' % p_xml)
# Step (4)
# Undefine pool
if not same_source_test:
result = virsh.pool_undefine(pool_name)
utlv.check_exit_status(result)
check_pool_list(pool_name, "--all", True)
# Step (5)
# Define pool from XML file
result = virsh.pool_define(poolxml, debug=True)
# Give error msg when exit status is not expected
if "/" in new_pool_name and not result.exit_status:
error_msg = "https://bugzilla.redhat.com/show_bug.cgi?id=639923 "
error_msg += "is helpful for tracing this bug."
logging.error(error_msg)
if "." in new_pool_name and result.exit_status:
error_msg = "https://bugzilla.redhat.com/show_bug.cgi?id=1333248 "
error_msg += "is helpful for tracing this bug."
logging.error(error_msg)
if same_source_test and not result.exit_status:
error_msg = "https://bugzilla.redhat.com/show_bug.cgi?id=1171984 "
error_msg += "is helpful for tracing this bug."
logging.error(error_msg)
utlv.check_exit_status(result, status_error)
if not result.exit_status:
# Step (6)
# Buid pool
# '--overwrite/--no-overwrite' just for fs/disk/logiacl type pool
# disk/fs pool: as prepare step already make label and create filesystem
# for the disk, use '--overwrite' is necessary
# logical_pool: build pool will fail if VG already exist, BZ#1373711
if new_pool_name:
pool_name = new_pool_name
if pool_type != "logical":
result = virsh.pool_build(pool_name, build_option, ignore_status=True)
utlv.check_exit_status(result)
# Step (7)
# Pool start
result = virsh.pool_start(pool_name, debug=True, ignore_status=True)
utlv.check_exit_status(result)
# Step (8)
# Pool list
option = "--persistent --type %s" % pool_type
check_pool_list(pool_name, option)
# Step (9)
# Pool autostart
result = virsh.pool_autostart(pool_name, ignore_status=True)
utlv.check_exit_status(result)
# Step (10)
# Pool list
option = "--autostart --type %s" % pool_type
check_pool_list(pool_name, option)
# Step (11)
# Restart libvirtd and check the autostart pool
utils_libvirtd.libvirtd_restart()
option = "--autostart --persistent"
check_pool_list(pool_name, option)
# Step (12)
# Pool destroy
if virsh.pool_destroy(pool_name):
logging.debug("Pool %s destroyed.", pool_name)
else:
test.fail("Destroy pool % failed." % pool_name)
# Step (13)
# Pool autostart disable
result = virsh.pool_autostart(pool_name, "--disable",
ignore_status=True)
utlv.check_exit_status(result)
# Step (14)
# Repeat step (11)
utils_libvirtd.libvirtd_restart()
option = "--autostart"
check_pool_list(pool_name, option, True)
# Step (15)
# Pool start
# When libvirtd starts up, it'll check to see if any of the storage
# pools have been activated externally. If so, then it'll mark the
# pool as active. This is independent of autostart.
# So a directory based storage pool is thus pretty much always active,
# and so as the SCSI pool.
if pool_type not in ["dir", 'scsi']:
result = virsh.pool_start(pool_name, ignore_status=True)
utlv.check_exit_status(result)
# Step (16)
# Pool info
pool_info = _pool.pool_info(pool_name)
logging.debug("Pool '%s' info:\n%s", pool_name, pool_info)
# Step (17)
# Pool UUID
result = virsh.pool_uuid(pool_info["Name"], ignore_status=True)
utlv.check_exit_status(result)
check_pool_info(pool_info, "UUID", result.stdout.strip())
# Step (18)
# Pool Name
result = virsh.pool_name(pool_info["UUID"], ignore_status=True)
utlv.check_exit_status(result)
check_pool_info(pool_info, "Name", result.stdout.strip())
# Step (19)
# Pool refresh for 'dir' type pool
if pool_type == "dir":
os.mknod(vol_path)
result = virsh.pool_refresh(pool_name)
utlv.check_exit_status(result)
check_vol_list(vol_name, pool_name)
# Step (20)
# Create an over size vol in pool(expect fail), then check pool:
# 'Capacity', 'Allocation' and 'Available'
# For NFS type pool, there's a bug(BZ#1077068) about allocate volume,
# and glusterfs pool not support create volume, so not test them
if pool_type != "netfs":
vol_capacity = "10000G"
vol_allocation = "10000G"
result = virsh.vol_create_as("oversize_vol", pool_name,
vol_capacity, vol_allocation, "raw")
utlv.check_exit_status(result, True)
new_info = _pool.pool_info(pool_name)
check_items = ["Capacity", "Allocation", "Available"]
for i in check_items:
check_pool_info(pool_info, i, new_info[i])
# Step (21)
# Undefine pool, this should fail as the pool is active
result = virsh.pool_undefine(pool_name, ignore_status=True)
utlv.check_exit_status(result, expect_error=True)
check_pool_list(pool_name, "", False)
# Step (22)
# Pool destroy
if virsh.pool_destroy(pool_name):
logging.debug("Pool %s destroyed.", pool_name)
else:
test.fail("Destroy pool % failed." % pool_name)
# Step (23)
# Pool delete for 'dir' type pool
if pool_type == "dir":
for f in os.listdir(pool_target):
os.remove(os.path.join(pool_target, f))
result = virsh.pool_delete(pool_name, ignore_status=True)
utlv.check_exit_status(result)
option = "--inactive --type %s" % pool_type
check_pool_list(pool_name, option)
if os.path.exists(pool_target):
test.fail("The target path '%s' still exist." %
pool_target)
result = virsh.pool_start(pool_name, ignore_status=True)
utlv.check_exit_status(result, True)
# Step (24)
# Pool undefine
result = virsh.pool_undefine(pool_name, ignore_status=True)
utlv.check_exit_status(result)
check_pool_list(pool_name, "--all", True)
finally:
# Clean up
try:
pvt.cleanup_pool(pool_name, pool_type, pool_target,
emulated_image, **kwargs)
utlv.setup_or_cleanup_iscsi(False)
except exceptions.TestFail as detail:
logging.error(str(detail))
if multipathd_status:
multipathd.start()
if os.path.exists(poolxml):
os.remove(poolxml)
