def storage(args):
job_manager = JobManager()
params = '-d {} '.format(args.d)
for j in job_manager.get_running_job_list():
if j.type == JobType.STORAGELOAD :
logger.info("Cannot start storage load job for 'sar',")
print("-1")
return
print( job_manager.add_job(JobType.STORAGELOAD,params))
logger.info("Start storage load job for 'sar'")
sys.exit(0)
def client(args):
job_manager = JobManager()
params = '-d {} -t {} -b {} -m {} -p {}'.format(args.d, args.t, args.b,
args.m, args.p)
for j in job_manager.get_running_job_list():
if j.type == JobType.CLIENTSTRESS:
logger.info("Cannot start client stress job for 'rados',")
print("-1")
return
print(job_manager.add_job(JobType.CLIENTSTRESS, params))
logger.info("Start client stress job for rados")
sys.exit(0)
def delete_cache(args):
if not configuration().get_node_info().is_storage:
print("-1")
return
job_manager = JobManager()
params = '-disk_name {}'.format(args.disk_name)
for j in job_manager.get_running_job_list():
if j.type == JobType.DELETEOSD or j.type == JobType.ADDDISK or \
j.type == JobType.ADDJOURNAL or j.type == JobType.DELETEJOURNAL or \
j.type == JobType.ADDCACHE or j.type == JobType.DELETECACHE:
logger.info("Cannot start delete job for cache of disk {},There ara running jobs. ".format(args.disk_name))
print("-1")
return
print(job_manager.add_job(JobType.DELETECACHE,params))
logger.info("Start delete job for cache of disk {}".format(args.disk_name))
sys.exit()
def start(self, type, duration_sec, threads, clients, pool, cleanup):
job_manager = JobManager()
clients = "" + ",".join(clients) + ""
for j in job_manager.get_running_job_list():
if j.type == JobType.BENCHMANAGER:
logger.info(
"Cannot start benchmark manager there is a job already running."
)
return -1
cleanup_val = 1
if not cleanup:
cleanup_val = 0
params = '-d {} -t {} -type {} -c {} -p {} --cleanup {}'.format(
duration_sec, threads, type, clients, pool, cleanup_val)
id = job_manager.add_job(JobType.BENCHMANAGER, params)
return id
def add_journal(args):
if not configuration().get_node_info().is_storage:
print("-1")
return
job_manager = JobManager()
params = '-disk_name {}'.format(args.disk_name)
for j in job_manager.get_running_job_list():
if j.type == JobType.DELETEOSD or j.type == JobType.ADDDISK or\
j.type == JobType.ADDJOURNAL or j.type == JobType.DELETEJOURNAL:
logger.info(
"Cannot start add journal job to create journal for disk.{}.There ara running jobs. "
.format(args.disk_name))
print("-1")
return
print(job_manager.add_job(JobType.ADDJOURNAL, params))
logger.info("Start add journal job for disk {}.".format(args.disk_name))
sys.exit()
def add_osd(args):
if not configuration().get_node_info().is_storage:
print("-1")
return
job_manager = JobManager()
# If no journal and no cache :
if str(args.journal) == "" and str(args.cache) == "":
params = '-disk_name {}'.format(args.disk_name)
# If journal but no cache :
elif str(args.journal) != "" and str(args.cache) == "":
params = '-disk_name {} -journal {}'.format(args.disk_name,
args.journal)
# If cache but no journal :
elif str(args.journal) == "" and str(args.cache) != "" and str(
args.cache_type) != "":
params = '-disk_name {} -cache {} -cache_type {}'.format(
args.disk_name, args.cache, args.cache_type)
# If both journal and cache :
else:
params = '-disk_name {} -journal {} -cache {} -cache_type {}'.format(
args.disk_name, args.journal, args.cache, args.cache_type)
# Getting all running jobs :
for j in job_manager.get_running_job_list():
if j.type == JobType.DELETEOSD or j.type == JobType.ADDDISK or \
j.type == JobType.ADDJOURNAL or j.type == JobType.DELETEJOURNAL or \
j.type == JobType.ADDCACHE or j.type == JobType.DELETECACHE:
logger.info(
"Cannot start add job to create osd for disk : {}. There ara running jobs."
.format(args.disk_name))
print("-1")
return
print(job_manager.add_job(JobType.ADDDISK, params))
logger.info("Start add osd job for disk {}.".format(args.disk_name))
sys.exit()
def delete_disk(self, disk_id, pool):
ceph_api = CephAPI()
consul_api = ConsulAPI()
ls = ceph_api.get_disks_meta_for_pool(pool)
try:
for disk in ls:
if disk_id == disk.id:
if disk and hasattr(disk, "paths") and not disk.paths:
disk_status = DisplayDiskStatus.unattached
elif disk and hasattr(disk, "paths") and disk.paths:
data = consul_api.find_disk(disk.id)
if data is not None:
disk_status = DisplayDiskStatus.started
if str(data.Flags) == "1":
disk_status = DisplayDiskStatus.stopping
else:
disk_status = DisplayDiskStatus.stopped
break
disk_status = None
except:
return StopDiskStatus.error
if disk_status == DisplayDiskStatus.started or disk_status == DisplayDiskStatus.stopping:
return StopDiskStatus.working
elif disk_status is None:
return StopDiskStatus.error
elif disk_status == DisplayDiskStatus.stopped or disk_status == DisplayDiskStatus.unattached:
# return ceph_api.delete_disk(disk_id,pool)
# start: delete disk as a job
__image_name_prefix = ConfigAPI().get_image_name_prefix()
# set image_name by disk_id :
image_name = disk_id
# if PetaSAN disk :
if disk_id.isdigit() and (len(disk_id) == 5):
image_name = __image_name_prefix + str(disk_id)
jm = JobManager()
try:
id = jm.add_job(JobType.DELETE_DISK, image_name + ' ' + pool)
print("Start Delete image: ", image_name)
if id > 0:
logger.info(
"Deleting disk: {} has been started as a job".format(
image_name))
return id
except Exception as ex:
logger.error("Error Deleting disk: {}".format(image_name))
# end: delete disk as a job #
else:
return StopDiskStatus.error
def clear_disk(args):
disk_id = args.disk_id
image_name = "image-" + disk_id
try:
# Get which ceph user is using this function & get his keyring file path #
# ---------------------------------------------------------------------- #
ceph_auth = CephAuthenticator()
config = configuration()
cluster_name = config.get_cluster_name()
# Get disk metadata :
# -------------------
ceph_api = CephAPI()
disk_metadata = ceph_api.get_diskmeta(disk_id)
# Get pool name :
# ---------------
pool_name = disk_metadata.pool
data_pool = ""
# Check if disk has been created on replicated pool or erasure pool :
# -------------------------------------------------------------------
if len(disk_metadata.data_pool) > 0:
data_pool = disk_metadata.data_pool
tmp_image_name = "tmp_disk_" + disk_metadata.id
# (1.) Check if a previous tmp image for this disk is still existed :
# ===================================================================
images_list = ceph_api.get_all_images(pool_name)
for image in images_list:
if tmp_image_name in image:
# Delete image #
cmd = "rbd rm {}/{} {} --cluster {}".format(
pool_name, image, ceph_auth.get_authentication_string(),
cluster_name)
if not call_cmd(cmd):
print(
"Error : clear_disk.py script : cannot remove tmp image ,\ncmd : "
+ cmd)
sys.exit(-1)
print(
"Stage 1 :\n\tCheck if a previous tmp image for this disk is still existed > (Completed)"
)
logger.info(
"Stage 1 :\n\tCheck if a previous tmp image for this disk is still existed > (Completed)"
)
# (2.) Stop old disk :
# ====================
consul_api = ConsulAPI()
kv = consul_api.find_disk(disk_id)
if kv is not None:
manage_disk = ManageDisk()
status = manage_disk.stop(disk_id)
if status != Status.done:
print('Error : Cannot stop disk , id = ' + disk_id)
sys.exit(-1)
print("Stage 2 :\n\tStop old disk > (Completed)")
logger.info("Stage 2 :\n\tStop old disk > (Completed)")
time.sleep(3)
# (3.) Check if old disk is stopped or not :
# ==========================================
if len(data_pool) > 0:
pool_type = "erasure"
_confirm_disk_stopped(data_pool, disk_id, pool_type)
else:
pool_type = "replicated"
_confirm_disk_stopped(pool_name, disk_id, pool_type)
print(
"Stage 3 :\n\tConfirm that disk is completely stopped > (Completed)"
)
logger.info(
"Stage 3 :\n\tConfirm that disk is completely stopped > (Completed)"
)
else:
print("Stage 2 :\n\tStop old disk > (Completed)")
logger.info("Stage 2 :\n\tStop old disk > (Completed)")
print(
"Stage 3 :\n\tConfirm that disk is completely stopped > (Completed)"
)
logger.info(
"Stage 3 :\n\tConfirm that disk is completely stopped > (Completed)"
)
print('\tclear_disk.py script : disk {} is already stopped'.format(
disk_id))
# (4.) Create a tmp image (not PetaSAN image) :
# =============================================
# Generate a random value between 1 and 99999 #
random_no = str(random.randint(1, 100000))
tmp_image_name = tmp_image_name + "_" + str(random_no)
image_size = disk_metadata.size * 1024
if len(data_pool) > 0:
cmd = "rbd create {}/{} --size {} --data-pool {} {} --cluster {}".format(
pool_name, tmp_image_name, image_size, data_pool,
ceph_auth.get_authentication_string(), cluster_name)
else:
cmd = "rbd create {}/{} --size {} {} --cluster {}".format(
pool_name, tmp_image_name, image_size,
ceph_auth.get_authentication_string(), cluster_name)
if not call_cmd(cmd):
print(
"Error : clear_disk.py script : cannot create new tmp image ,\ncmd : "
+ cmd)
sys.exit(-1)
print("Stage 4 :\n\tCreate a tmp image called ( " + tmp_image_name +
" ) > (Completed)")
logger.info("Stage 4 :\n\tCreate a tmp image called ( " +
tmp_image_name + " ) > (Completed)")
# (5.) Run script to copy "old disk" metadata to new "tmp_disk" :
# ===============================================================
metadata_script_file = ConfigAPI().get_disk_meta_script_path()
# Function : read_disks_metadata :
parser_key_1 = "read"
arg_1 = "--image"
arg_2 = "--pool"
# Function : set_disk_metadata :
parser_key_2 = "write"
arg_3 = "--file"
cmd = metadata_script_file + " " + parser_key_1 + " " + arg_1 + " " + image_name + " " + arg_2 + " " + pool_name +\
" | " + metadata_script_file + " " + parser_key_2 + " " + arg_1 + " " + tmp_image_name + " " + arg_2 + " " + pool_name
if not call_cmd(cmd):
print(
"Error : clear_disk.py script : cannot copy metadata from old disk to new tmp image ,\ncmd : "
+ cmd)
sys.exit(-1)
print(
"Stage 5 :\n\tRun script to copy 'old disk' metadata to new 'tmp_disk' > (Completed)"
)
logger.info(
"Stage 5 :\n\tRun script to copy 'old disk' metadata to new 'tmp_disk' > (Completed)"
)
time.sleep(3)
# (6.) Remove metadata of old disk :
# ===========================================================
old_image_name = str(ceph_api.conf_api.get_image_name_prefix() +
disk_metadata.id)
confirm = ceph_api.remove_disk_metadata(old_image_name,
disk_metadata.pool)
if not confirm:
print(
"Error : clear_disk.py script : cannot remove metadata of old disk"
)
# sys.exit(-1)
print("Stage 6 :\n\tRemove metadata of old disk > (Completed)")
logger.info("Stage 6 :\n\tRemove metadata of old disk > (Completed)")
# (7.) Rename old disk image name with "deleted-" + disk_id + random_no:
# ======================================================================
new_image_name = "deleted-" + disk_metadata.id + "-" + random_no
cmd = "rbd mv {}/{} {} {} --cluster {}".format(
pool_name, image_name, new_image_name,
ceph_auth.get_authentication_string(), cluster_name)
if not call_cmd(cmd):
print(
"Error : clear_disk.py script : cannot rename old image from {} to {} ,\ncmd : {}"
.format(image_name, new_image_name, cmd))
sys.exit(-1)
print("Stage 7 :\n\tRename old disk image name with ( " +
new_image_name + " ) > (Completed)")
logger.info("Stage 7 :\n\tRename old disk image name with ( " +
new_image_name + " ) > (Completed)")
time.sleep(5)
# (8.) Rename "tmp_disk" with old disk image name :
# =================================================
cmd = "rbd mv {}/{} {} {} --cluster {}".format(
pool_name, tmp_image_name, image_name,
ceph_auth.get_authentication_string(), cluster_name)
if not call_cmd(cmd):
print(
"Error : clear_disk.py script : cannot rename \"tmp_disk\" from {} to {} ,\ncmd : {}"
.format(tmp_image_name, image_name, cmd))
sys.exit(-1)
print(
"Stage 8 :\n\tRename 'tmp_disk' with old disk image name > (Completed)"
)
logger.info(
"Stage 8 :\n\tRename 'tmp_disk' with old disk image name > (Completed)"
)
time.sleep(5)
jm = JobManager()
id = jm.add_job(JobType.DELETE_DISK, new_image_name + ' ' + pool_name)
print("Stage 9 :\n\tStart a job to remove old disk image , job id = " +
str(id))
logger.info(
"Stage 9 :\n\tStart a job to remove old disk image , job id = " +
str(id))
sys.exit(0)
except PoolException as e:
print("Error : PoolException , {}".format(e.message))
logger.error("Clear Disk Error : PoolException , {}".format(e.message))
sys.exit(-1)
except DiskListException as e:
print("Error : DiskListException , {}".format(e.message))
logger.error("Clear Disk Error : DiskListException , {}".format(
e.message))
sys.exit(-1)
except CephException as e:
if e.id == CephException.GENERAL_EXCEPTION:
print("Error : CephException , {}".format(e.message))
logger.error("Clear Disk Error : CephException , {}".format(e.message))
sys.exit(-1)
except MetadataException as e:
print("Error : MetadataException , {}".format(e.message))
logger.error("Clear Disk Error : MetadataException , {}".format(
e.message))
sys.exit(-1)
except Exception as e:
print("Error : Exception , {}".format(e.message))
logger.error("Clear Disk Error : Exception , {}".format(e.message))
sys.exit(-1)
def delete_pool(self, pool_name):
jm = JobManager()
id = jm.add_job(JobType.DELETE_POOL,
pool_name + ' --yes-i-really-really-mean-it')
return id