def run(test, params, env):
"""
Test the pull-mode backup function
Steps:
1. craete a vm with extra disk vdb
2. create some data on vdb
3. start a pull mode full backup on vdb
4. create some data on vdb
5. start a pull mode incremental backup
6. repeat step 5 to 7
7. check the full/incremental backup file data
"""
# Basic case config
hotplug_disk = "yes" == params.get("hotplug_disk", "no")
original_disk_size = params.get("original_disk_size", "100M")
original_disk_type = params.get("original_disk_type", "local")
original_disk_target = params.get("original_disk_target", "vdb")
local_hostname = params.get("loal_hostname", "localhost")
local_ip = params.get("local_ip", "127.0.0.1")
local_user_name = params.get("local_user_name", "root")
local_user_password = params.get("local_user_password", "redhat")
tmp_dir = data_dir.get_tmp_dir()
# Backup config
scratch_type = params.get("scratch_type", "file")
reuse_scratch_file = "yes" == params.get("reuse_scratch_file")
prepare_scratch_file = "yes" == params.get("prepare_scratch_file")
scratch_blkdev_path = params.get("scratch_blkdev_path")
scratch_blkdev_size = params.get("scratch_blkdev_size", original_disk_size)
prepare_scratch_blkdev = "yes" == params.get("prepare_scratch_blkdev")
backup_rounds = int(params.get("backup_rounds", 3))
backup_error = "yes" == params.get("backup_error")
expect_backup_canceled = "yes" == params.get("expect_backup_canceled")
# NBD service config
nbd_protocol = params.get("nbd_protocol", "unix")
nbd_socket = params.get("nbd_socket", "/tmp/pull_backup.socket")
nbd_tcp_port = params.get("nbd_tcp_port", "10809")
nbd_hostname = local_hostname
set_exportname = "yes" == params.get("set_exportname")
set_exportbitmap = "yes" == params.get("set_exportbitmap")
# TLS service config
tls_enabled = "yes" == params.get("tls_enabled")
tls_x509_verify = "yes" == params.get("tls_x509_verify")
custom_pki_path = "yes" == params.get("custom_pki_path")
tls_client_ip = tls_server_ip = local_ip
tls_client_cn = tls_server_cn = local_hostname
tls_client_user = tls_server_user = local_user_name
tls_client_pwd = tls_server_pwd = local_user_password
tls_provide_client_cert = "yes" == params.get("tls_provide_client_cert")
tls_error = "yes" == params.get("tls_error")
# LUKS config
scratch_luks_encrypted = "yes" == params.get("scratch_luks_encrypted")
luks_passphrase = params.get("luks_passphrase", "password")
# Cancel the test if libvirt support related functions
if not libvirt_version.version_compare(6, 0, 0):
test.cancel("Current libvirt version doesn't support "
"incremental backup.")
if tls_enabled and not libvirt_version.version_compare(6, 6, 0):
test.cancel("Current libvirt version doesn't support pull mode "
"backup with tls nbd.")
try:
vm_name = params.get("main_vm")
vm = env.get_vm(vm_name)
# Make sure there is no checkpoint metadata before test
utils_backup.clean_checkpoints(vm_name)
# Backup vm xml
vmxml = vm_xml.VMXML.new_from_inactive_dumpxml(vm_name)
vmxml_backup = vmxml.copy()
utils_backup.enable_inc_backup_for_vm(vm)
# Prepare tls env
if tls_enabled:
# Prepare pki
tls_config = {
"qemu_tls": "yes",
"auto_recover": "yes",
"client_ip": tls_client_ip,
"server_ip": tls_server_ip,
"client_cn": tls_client_cn,
"server_cn": tls_server_cn,
"client_user": tls_client_user,
"server_user": tls_server_user,
"client_pwd": tls_client_pwd,
"server_pwd": tls_server_pwd,
}
if custom_pki_path:
pki_path = os.path.join(tmp_dir, "inc_bkup_pki")
else:
pki_path = "/etc/pki/libvirt-backup/"
if tls_x509_verify:
tls_config["client_ip"] = tls_client_ip
tls_config["custom_pki_path"] = pki_path
tls_obj = TLSConnection(tls_config)
tls_obj.conn_setup(True, tls_provide_client_cert)
logging.debug("TLS certs in: %s" % pki_path)
# Set qemu.conf
qemu_config = LibvirtQemuConfig()
if tls_x509_verify:
qemu_config.backup_tls_x509_verify = True
else:
qemu_config.backup_tls_x509_verify = False
if custom_pki_path:
qemu_config.backup_tls_x509_cert_dir = pki_path
utils_libvirtd.Libvirtd().restart()
# Prepare libvirt secret
if scratch_luks_encrypted:
utils_secret.clean_up_secrets()
luks_secret_uuid = libvirt.create_secret(params)
virsh.secret_set_value(luks_secret_uuid,
luks_passphrase,
encode=True,
debug=True)
# Prepare the disk to be backuped.
disk_params = {}
disk_path = ""
if original_disk_type == "local":
image_name = "{}_image.qcow2".format(original_disk_target)
disk_path = os.path.join(tmp_dir, image_name)
libvirt.create_local_disk("file", disk_path, original_disk_size,
"qcow2")
disk_params = {
"device_type": "disk",
"type_name": "file",
"driver_type": "qcow2",
"target_dev": original_disk_target,
"source_file": disk_path
}
if original_disk_target:
disk_params["target_dev"] = original_disk_target
elif original_disk_type == "iscsi":
iscsi_host = '127.0.0.1'
iscsi_target, lun_num = libvirt.setup_or_cleanup_iscsi(
is_setup=True,
is_login=False,
image_size=original_disk_size,
portal_ip=iscsi_host)
disk_path = ("iscsi://[%s]/%s/%s" %
(iscsi_host, iscsi_target, lun_num))
process.run("qemu-img create -f qcow2 %s %s" %
(disk_path, original_disk_size),
shell=True,
verbose=True)
disk_params = {
'device_type': 'disk',
'type_name': 'network',
"driver_type": "qcow2",
'target_dev': original_disk_target
}
disk_params_src = {
'source_protocol': 'iscsi',
'source_name': iscsi_target + "/%s" % lun_num,
'source_host_name': iscsi_host,
'source_host_port': '3260'
}
disk_params.update(disk_params_src)
elif original_disk_type == "gluster":
gluster_vol_name = "gluster_vol"
gluster_pool_name = "gluster_pool"
gluster_img_name = "gluster.qcow2"
gluster_host_ip = gluster.setup_or_cleanup_gluster(
is_setup=True,
vol_name=gluster_vol_name,
pool_name=gluster_pool_name,
**params)
disk_path = 'gluster://%s/%s/%s' % (
gluster_host_ip, gluster_vol_name, gluster_img_name)
process.run("qemu-img create -f qcow2 %s %s" %
(disk_path, original_disk_size),
shell=True,
verbose=True)
disk_params = {
'device_type': 'disk',
'type_name': 'network',
"driver_type": "qcow2",
'target_dev': original_disk_target
}
disk_params_src = {
'source_protocol': 'gluster',
'source_name': gluster_vol_name + "/%s" % gluster_img_name,
'source_host_name': gluster_host_ip,
'source_host_port': '24007'
}
disk_params.update(disk_params_src)
else:
test.error("The disk type '%s' not supported in this script.",
original_disk_type)
if hotplug_disk:
vm.start()
session = vm.wait_for_login().close()
disk_xml = libvirt.create_disk_xml(disk_params)
virsh.attach_device(vm_name, disk_xml, debug=True)
else:
disk_xml = libvirt.create_disk_xml(disk_params)
virsh.attach_device(vm.name,
disk_xml,
flagstr="--config",
debug=True)
vm.start()
session = vm.wait_for_login()
new_disks_in_vm = list(utils_disk.get_linux_disks(session).keys())
session.close()
if len(new_disks_in_vm) != 1:
test.fail("Test disk not prepared in vm")
# Use the newly added disk as the test disk
test_disk_in_vm = "/dev/" + new_disks_in_vm[0]
vmxml = vm_xml.VMXML.new_from_dumpxml(vm_name)
vm_disks = list(vmxml.get_disk_all().keys())
checkpoint_list = []
is_incremental = False
backup_file_list = []
for backup_index in range(backup_rounds):
# Prepare backup xml
backup_params = {"backup_mode": "pull"}
if backup_index > 0:
is_incremental = True
backup_params["backup_incremental"] = "checkpoint_" + str(
backup_index - 1)
# Set libvirt default nbd export name and bitmap name
nbd_export_name = original_disk_target
nbd_bitmap_name = "backup-" + original_disk_target
backup_server_dict = {}
if nbd_protocol == "unix":
backup_server_dict["transport"] = "unix"
backup_server_dict["socket"] = nbd_socket
else:
backup_server_dict["name"] = nbd_hostname
backup_server_dict["port"] = nbd_tcp_port
if tls_enabled:
backup_server_dict["tls"] = "yes"
backup_params["backup_server"] = backup_server_dict
backup_disk_xmls = []
for vm_disk in vm_disks:
backup_disk_params = {"disk_name": vm_disk}
if vm_disk != original_disk_target:
backup_disk_params["enable_backup"] = "no"
else:
backup_disk_params["enable_backup"] = "yes"
backup_disk_params["disk_type"] = scratch_type
# Custom nbd export name and bitmap name if required
if set_exportname:
nbd_export_name = original_disk_target + "_custom_exp"
backup_disk_params["exportname"] = nbd_export_name
if set_exportbitmap:
nbd_bitmap_name = original_disk_target + "_custom_bitmap"
backup_disk_params["exportbitmap"] = nbd_bitmap_name
# Prepare nbd scratch file/dev params
scratch_params = {"attrs": {}}
scratch_path = None
if scratch_type == "file":
scratch_file_name = "scratch_file_%s" % backup_index
scratch_path = os.path.join(tmp_dir, scratch_file_name)
if prepare_scratch_file:
libvirt.create_local_disk("file", scratch_path,
original_disk_size,
"qcow2")
scratch_params["attrs"]["file"] = scratch_path
elif scratch_type == "block":
if prepare_scratch_blkdev:
scratch_path = libvirt.setup_or_cleanup_iscsi(
is_setup=True, image_size=scratch_blkdev_size)
scratch_params["attrs"]["dev"] = scratch_path
else:
test.fail(
"We do not support backup scratch type: '%s'" %
scratch_type)
if scratch_luks_encrypted:
encryption_dict = {
"encryption": "luks",
"secret": {
"type": "passphrase",
"uuid": luks_secret_uuid
}
}
scratch_params["encryption"] = encryption_dict
logging.debug("scratch params: %s", scratch_params)
backup_disk_params["backup_scratch"] = scratch_params
backup_disk_xml = utils_backup.create_backup_disk_xml(
backup_disk_params)
backup_disk_xmls.append(backup_disk_xml)
logging.debug("disk list %s", backup_disk_xmls)
backup_xml = utils_backup.create_backup_xml(
backup_params, backup_disk_xmls)
logging.debug("ROUND_%s Backup Xml: %s", backup_index, backup_xml)
# Prepare checkpoint xml
checkpoint_name = "checkpoint_%s" % backup_index
checkpoint_list.append(checkpoint_name)
cp_params = {"checkpoint_name": checkpoint_name}
cp_params["checkpoint_desc"] = params.get(
"checkpoint_desc", "desc of cp_%s" % backup_index)
disk_param_list = []
for vm_disk in vm_disks:
cp_disk_param = {"name": vm_disk}
if vm_disk != original_disk_target:
cp_disk_param["checkpoint"] = "no"
else:
cp_disk_param["checkpoint"] = "bitmap"
cp_disk_bitmap = params.get("cp_disk_bitmap")
if cp_disk_bitmap:
cp_disk_param["bitmap"] = cp_disk_bitmap + str(
backup_index)
disk_param_list.append(cp_disk_param)
checkpoint_xml = utils_backup.create_checkpoint_xml(
cp_params, disk_param_list)
logging.debug("ROUND_%s Checkpoint Xml: %s", backup_index,
checkpoint_xml)
# Create some data in vdb
dd_count = "1"
dd_seek = str(backup_index * 10 + 10)
dd_bs = "1M"
session = vm.wait_for_login()
utils_disk.dd_data_to_vm_disk(session, test_disk_in_vm, dd_bs,
dd_seek, dd_count)
session.close()
# Start backup
backup_options = backup_xml.xml + " " + checkpoint_xml.xml
if reuse_scratch_file:
backup_options += " --reuse-external"
backup_result = virsh.backup_begin(vm_name,
backup_options,
ignore_status=True,
debug=True)
if backup_result.exit_status:
raise utils_backup.BackupBeginError(
backup_result.stderr.strip())
# If required, do some error operations during backup job
error_operation = params.get("error_operation")
if error_operation:
if "destroy_vm" in error_operation:
vm.destroy(gracefully=False)
if "kill_qemu" in error_operation:
utils_misc.safe_kill(vm.get_pid(), signal.SIGKILL)
if utils_misc.wait_for(
lambda: utils_backup.is_backup_canceled(vm_name),
timeout=5):
raise utils_backup.BackupCanceledError()
elif expect_backup_canceled:
test.fail("Backup job should be canceled but not.")
backup_file_path = os.path.join(
tmp_dir, "backup_file_%s.qcow2" % str(backup_index))
backup_file_list.append(backup_file_path)
nbd_params = {
"nbd_protocol": nbd_protocol,
"nbd_export": nbd_export_name
}
if nbd_protocol == "unix":
nbd_params["nbd_socket"] = nbd_socket
elif nbd_protocol == "tcp":
nbd_params["nbd_hostname"] = nbd_hostname
nbd_params["nbd_tcp_port"] = nbd_tcp_port
if tls_enabled:
nbd_params["tls_dir"] = pki_path
nbd_params["tls_server_ip"] = tls_server_ip
if not is_incremental:
# Do full backup
try:
utils_backup.pull_full_backup_to_file(
nbd_params, backup_file_path)
except Exception as details:
if tls_enabled and tls_error:
raise utils_backup.BackupTLSError(details)
else:
test.fail("Fail to get full backup data: %s" % details)
logging.debug("Full backup to: %s", backup_file_path)
else:
# Do incremental backup
utils_backup.pull_incremental_backup_to_file(
nbd_params, backup_file_path, nbd_bitmap_name,
original_disk_size)
# Check if scratch file encrypted
if scratch_luks_encrypted and scratch_path:
cmd = "qemu-img info -U %s" % scratch_path
result = process.run(cmd, shell=True,
verbose=True).stdout_text.strip()
if (not re.search("format.*luks", result, re.IGNORECASE)
or not re.search("encrypted.*yes", result,
re.IGNORECASE)):
test.fail("scratch file/dev is not encrypted by LUKS")
virsh.domjobabort(vm_name, debug=True)
for checkpoint_name in checkpoint_list:
virsh.checkpoint_delete(vm_name, checkpoint_name, debug=True)
if vm.is_alive():
vm.destroy(gracefully=False)
# Compare the backup data and original data
original_data_file = os.path.join(tmp_dir, "original_data.qcow2")
cmd = "qemu-img convert -f qcow2 %s -O qcow2 %s" % (disk_path,
original_data_file)
process.run(cmd, shell=True, verbose=True)
for backup_file in backup_file_list:
if not utils_backup.cmp_backup_data(original_data_file,
backup_file):
test.fail("Backup and original data are not identical for"
"'%s' and '%s'" % (disk_path, backup_file))
else:
logging.debug("'%s' contains correct backup data", backup_file)
except utils_backup.BackupBeginError as detail:
if backup_error:
logging.debug("Backup failed as expected.")
else:
test.fail("Backup failed to start: %s" % detail)
except utils_backup.BackupTLSError as detail:
if tls_error:
logging.debug("Failed to get backup data as expected.")
else:
test.fail("Failed to get tls backup data: %s" % detail)
except utils_backup.BackupCanceledError as detail:
if expect_backup_canceled:
logging.debug("Backup canceled as expected.")
if not vm.is_alive():
logging.debug("Check if vm can be started again when backup "
"canceled.")
vm.start()
vm.wait_for_login().close()
else:
test.fail("Backup job canceled: %s" % detail)
finally:
# Remove checkpoints
clean_checkpoint_metadata = not vm.is_alive()
if "error_operation" in locals() and error_operation is not None:
if "kill_qemu" in error_operation:
clean_checkpoint_metadata = True
utils_backup.clean_checkpoints(
vm_name, clean_metadata=clean_checkpoint_metadata)
if vm.is_alive():
vm.destroy(gracefully=False)
# Restoring vm
vmxml_backup.sync()
# Remove iscsi devices
if original_disk_type == "iscsi" or scratch_type == "block":
libvirt.setup_or_cleanup_iscsi(False)
# Remove gluster devices
if original_disk_type == "gluster":
gluster.setup_or_cleanup_gluster(is_setup=False,
vol_name=gluster_vol_name,
pool_name=gluster_pool_name,
**params)
# Recover qemu.conf
if "qemu_config" in locals():
qemu_config.restore()
# Remove tls object
if "tls_obj" in locals():
del tls_obj
# Remove libvirt secret
if "luks_secret_uuid" in locals():
virsh.secret_undefine(luks_secret_uuid, ignore_status=True)
def run(test, params, env):
"""
Test push-mode incremental backup
Steps:
1. create a vm with extra disk vdb
2. create some data on vdb in vm
3. start a push mode full backup on vdb
4. create some data on vdb in vm
5. start a push mode incremental backup
6. repeat step 4 and 5 as required
7. check the full/incremental backup file data
"""
def backup_job_done(vm_name, vm_disk):
"""
Check if a backup job for a vm's specific disk is finished.
:param vm_name: vm's name
:param vm_disk: the disk to be checked, such as 'vdb'
:return: 'True' means job finished
"""
result = virsh.blockjob(vm_name, vm_disk, debug=True)
if "no current block job" in result.stdout_text.strip().lower():
return True
# Cancel the test if libvirt version is too low
if not libvirt_version.version_compare(6, 0, 0):
test.cancel("Current libvirt version doesn't support "
"incremental backup.")
hotplug_disk = "yes" == params.get("hotplug_disk", "no")
original_disk_size = params.get("original_disk_size", "100M")
original_disk_type = params.get("original_disk_type", "local")
original_disk_target = params.get("original_disk_target", "vdb")
target_driver = params.get("target_driver", "qcow2")
target_type = params.get("target_type", "file")
target_blkdev_path = params.get("target_blkdev_path")
target_blkdev_size = params.get("target_blkdev_size", original_disk_size)
reuse_target_file = "yes" == params.get("reuse_target_file")
prepare_target_file = "yes" == params.get("prepare_target_file")
prepare_target_blkdev = "yes" == params.get("prepare_target_blkdev")
backup_rounds = int(params.get("backup_rounds", 3))
backup_error = "yes" == params.get("backup_error")
expect_backup_canceled = "yes" == params.get("expect_backup_canceled")
tmp_dir = data_dir.get_data_dir()
virsh_dargs = {'debug': True, 'ignore_status': True}
try:
vm_name = params.get("main_vm")
vm = env.get_vm(vm_name)
# Make sure there is no checkpoint metadata before test
utils_backup.clean_checkpoints(vm_name)
# Backup vm xml
vmxml = vm_xml.VMXML.new_from_inactive_dumpxml(vm_name)
vmxml_backup = vmxml.copy()
utils_backup.enable_inc_backup_for_vm(vm)
# Prepare the disk to be backuped.
disk_params = {}
disk_path = ""
if original_disk_type == "local":
image_name = "{}_image.qcow2".format(original_disk_target)
disk_path = os.path.join(tmp_dir, image_name)
if os.path.exists(disk_path):
os.remove(disk_path)
libvirt.create_local_disk("file", disk_path, original_disk_size,
"qcow2")
disk_params = {"device_type": "disk",
"type_name": "file",
"driver_type": "qcow2",
"target_dev": original_disk_target,
"source_file": disk_path}
if original_disk_target:
disk_params["target_dev"] = original_disk_target
elif original_disk_type == "ceph":
ceph_mon_host = params.get("ceph_mon_host", "EXAMPLE_MON_HOST_AUTHX")
ceph_host_port = params.get("ceph_host_port", "EXAMPLE_PORT")
ceph_pool_name = params.get("ceph_pool_name", "EXAMPLE_POOL")
ceph_file_name = params.get("ceph_file_name", "EXAMPLE_FILE")
ceph_disk_name = ceph_pool_name + "/" + ceph_file_name
ceph_client_name = params.get("ceph_client_name", "EXAMPLE_CLIENT_NAME")
ceph_client_key = params.get("ceph_client_key", "EXAMPLE_CLIENT_KEY")
ceph_auth_user = params.get("ceph_auth_user", "EXAMPLE_AUTH_USER")
ceph_auth_key = params.get("ceph_auth_key", "EXAMPLE_AUTH_KEY")
auth_sec_usage_type = "ceph"
enable_auth = "yes" == params.get("enable_auth", "yes")
key_file = os.path.join(tmp_dir, "ceph.key")
key_opt = ""
# Prepare a blank params to confirm if delete the configure at the end of the test
ceph_cfg = ""
if not utils_package.package_install(["ceph-common"]):
test.error("Failed to install ceph-common")
# Create config file if it doesn't exist
ceph_cfg = ceph.create_config_file(ceph_mon_host)
if enable_auth:
# If enable auth, prepare a local file to save key
if ceph_client_name and ceph_client_key:
with open(key_file, 'w') as f:
f.write("[%s]\n\tkey = %s\n" %
(ceph_client_name, ceph_client_key))
key_opt = "--keyring %s" % key_file
auth_sec_dict = {"sec_usage": auth_sec_usage_type,
"sec_name": "ceph_auth_secret"}
auth_sec_uuid = libvirt.create_secret(auth_sec_dict)
virsh.secret_set_value(auth_sec_uuid, ceph_auth_key,
debug=True)
disk_params_auth = {"auth_user": ceph_auth_user,
"secret_type": auth_sec_usage_type,
"secret_uuid": auth_sec_uuid,
"auth_in_source": True}
else:
test.error("No ceph client name/key provided.")
disk_path = "rbd:%s:mon_host=%s:keyring=%s" % (ceph_disk_name,
ceph_mon_host,
key_file)
ceph.rbd_image_rm(ceph_mon_host, ceph_pool_name,
ceph_file_name, ceph_cfg, key_file)
process.run("qemu-img create -f qcow2 %s %s" % (disk_path, original_disk_size),
shell=True, verbose=True)
disk_params = {'device_type': 'disk',
'type_name': 'network',
"driver_type": "qcow2",
'target_dev': original_disk_target}
disk_params_src = {'source_protocol': 'rbd',
'source_name': ceph_disk_name,
'source_host_name': ceph_mon_host,
'source_host_port': ceph_host_port}
disk_params.update(disk_params_src)
disk_params.update(disk_params_auth)
else:
test.error("The disk type '%s' not supported in this script." %
original_disk_type)
if hotplug_disk:
vm.start()
session = vm.wait_for_login().close()
disk_xml = libvirt.create_disk_xml(disk_params)
virsh.attach_device(vm_name, disk_xml, debug=True)
else:
disk_xml = libvirt.create_disk_xml(disk_params)
virsh.attach_device(vm.name, disk_xml,
flagstr="--config", debug=True)
vm.start()
session = vm.wait_for_login()
new_disks_in_vm = list(utils_disk.get_linux_disks(session).keys())
session.close()
if len(new_disks_in_vm) != 1:
test.fail("Test disk not prepared in vm")
# Use the newly added disk as test disk
test_disk_in_vm = "/dev/" + new_disks_in_vm[0]
vmxml = vm_xml.VMXML.new_from_dumpxml(vm_name)
vm_disks = list(vmxml.get_disk_all().keys())
checkpoint_list = []
is_incremental = False
backup_path_list = []
for backup_index in range(backup_rounds):
# Prepare backup xml
backup_params = {"backup_mode": "push"}
if backup_index > 0:
is_incremental = True
backup_params["backup_incremental"] = "checkpoint_" + str(backup_index - 1)
backup_disk_xmls = []
for vm_disk in vm_disks:
backup_disk_params = {"disk_name": vm_disk}
if vm_disk != original_disk_target:
backup_disk_params["enable_backup"] = "no"
else:
backup_disk_params["enable_backup"] = "yes"
backup_disk_params["disk_type"] = target_type
target_params = {"attrs": {}}
if target_type == "file":
target_file_name = "target_file_%s" % backup_index
target_file_path = os.path.join(tmp_dir, target_file_name)
if os.path.exists(target_file_path):
os.remove(target_file_path)
if prepare_target_file:
libvirt.create_local_disk("file", target_file_path,
original_disk_size, target_driver)
target_params["attrs"]["file"] = target_file_path
backup_path_list.append(target_file_path)
elif target_type == "block":
if prepare_target_blkdev:
target_blkdev_path = libvirt.setup_or_cleanup_iscsi(
is_setup=True, image_size=target_blkdev_size)
target_params["attrs"]["dev"] = target_blkdev_path
backup_path_list.append(target_blkdev_path)
else:
test.fail("We do not support backup target type: '%s'"
% target_type)
logging.debug("target params: %s", target_params)
backup_disk_params["backup_target"] = target_params
driver_params = {"type": target_driver}
backup_disk_params["backup_driver"] = driver_params
backup_disk_xml = utils_backup.create_backup_disk_xml(
backup_disk_params)
backup_disk_xmls.append(backup_disk_xml)
logging.debug("disk list %s", backup_disk_xmls)
backup_xml = utils_backup.create_backup_xml(backup_params,
backup_disk_xmls)
logging.debug("ROUND_%s Backup Xml: %s", backup_index, backup_xml)
# Prepare checkpoint xml
checkpoint_name = "checkpoint_%s" % backup_index
checkpoint_list.append(checkpoint_name)
cp_params = {"checkpoint_name": checkpoint_name}
cp_params["checkpoint_desc"] = params.get("checkpoint_desc",
"desc of cp_%s" % backup_index)
disk_param_list = []
for vm_disk in vm_disks:
cp_disk_param = {"name": vm_disk}
if vm_disk != original_disk_target:
cp_disk_param["checkpoint"] = "no"
else:
cp_disk_param["checkpoint"] = "bitmap"
cp_disk_bitmap = params.get("cp_disk_bitmap")
if cp_disk_bitmap:
cp_disk_param["bitmap"] = cp_disk_bitmap + str(backup_index)
disk_param_list.append(cp_disk_param)
checkpoint_xml = utils_backup.create_checkpoint_xml(cp_params,
disk_param_list)
logging.debug("ROUND_%s Checkpoint Xml: %s",
backup_index, checkpoint_xml)
# Start backup
backup_options = backup_xml.xml + " " + checkpoint_xml.xml
# Create some data in vdb
dd_count = "1"
if expect_backup_canceled:
# Generate more data to extend the backup job duration
dd_count = "100"
dd_seek = str(backup_index * 10 + 10)
dd_bs = "1M"
session = vm.wait_for_login()
utils_disk.dd_data_to_vm_disk(session, test_disk_in_vm, dd_bs,
dd_seek, dd_count)
session.close()
if reuse_target_file:
backup_options += " --reuse-external"
backup_result = virsh.backup_begin(vm_name, backup_options,
debug=True)
if backup_result.exit_status:
raise utils_backup.BackupBeginError(backup_result.stderr.strip())
# If required, do some error operations during backup job
error_operation = params.get("error_operation")
if error_operation:
if "destroy_vm" in error_operation:
vm.destroy(gracefully=False)
if "kill_qemu" in error_operation:
utils_misc.safe_kill(vm.get_pid(), signal.SIGKILL)
if utils_misc.wait_for(lambda: utils_backup.is_backup_canceled(vm_name),
timeout=5):
raise utils_backup.BackupCanceledError()
elif expect_backup_canceled:
test.fail("Backup job should be canceled but not.")
# Wait for the backup job actually finished
if not utils_misc.wait_for(
lambda: backup_job_done(vm_name, original_disk_target), 60):
test.fail("Backup job not finished in 60s")
for checkpoint_name in checkpoint_list:
virsh.checkpoint_delete(vm_name, checkpoint_name, debug=True)
if vm.is_alive():
vm.destroy(gracefully=False)
# Compare the backup data and original data
original_data_file = os.path.join(tmp_dir, "original_data.qcow2")
cmd = "qemu-img convert -f qcow2 %s -O qcow2 %s" % (disk_path, original_data_file)
process.run(cmd, shell=True, verbose=True)
for backup_path in backup_path_list:
if target_driver == "qcow2":
# Clear backup image's backing file before comparison
qemu_cmd = ("qemu-img rebase -u -f qcow2 -b '' -F qcow2 %s"
% backup_path)
process.run(qemu_cmd, shell=True, verbose=True)
if not utils_backup.cmp_backup_data(original_data_file, backup_path,
backup_file_driver=target_driver):
test.fail("Backup and original data are not identical for"
"'%s' and '%s'" % (disk_path, backup_path))
else:
logging.debug("'%s' contains correct backup data", backup_path)
except utils_backup.BackupBeginError as details:
if backup_error:
logging.debug("Backup failed as expected.")
else:
test.fail(details)
except utils_backup.BackupCanceledError as detail:
if expect_backup_canceled:
logging.debug("Backup canceled as expected.")
if not vm.is_alive():
logging.debug("Check if vm can be started again when backup "
"canceled.")
vm.start()
vm.wait_for_login().close()
else:
test.fail("Backup job canceled: %s" % detail)
finally:
# Remove checkpoints
utils_backup.clean_checkpoints(vm_name,
clean_metadata=not vm.is_alive())
if vm.is_alive():
vm.destroy(gracefully=False)
# Restoring vm
vmxml_backup.sync()
# Remove local backup file
if "target_file_path" in locals():
if os.path.exists(target_file_path):
os.remove(target_file_path)
# Remove test disk's local image file
if original_disk_type == "local":
if "disk_path" in locals() and os.path.exists(disk_path):
os.remove(disk_path)
# Remove iscsi devices
libvirt.setup_or_cleanup_iscsi(False)
# Remove ceph related data
if original_disk_type == "ceph":
ceph.rbd_image_rm(ceph_mon_host, ceph_pool_name,
ceph_file_name, ceph_cfg, key_file)
if "auth_sec_uuid" in locals() and auth_sec_uuid:
virsh.secret_undefine(auth_sec_uuid)
if "ceph_cfg" in locals() and os.path.exists(ceph_cfg):
os.remove(ceph_cfg)
if os.path.exists(key_file):
os.remove(key_file)