sbcli-dev 10.4.0__tar.gz → 10.7.1__tar.gz

{sbcli_dev-10.4.0 → sbcli_dev-10.7.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: sbcli-dev
-Version: 10.4.0
+Version: 10.7.1
 Summary: CLI for managing SimplyBlock cluster
 Home-page: https://www.simplyblock.io/
 Author: Hamdy

{sbcli_dev-10.4.0 → sbcli_dev-10.7.1}/sbcli_dev.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: sbcli-dev
-Version: 10.4.0
+Version: 10.7.1
 Summary: CLI for managing SimplyBlock cluster
 Home-page: https://www.simplyblock.io/
 Author: Hamdy

{sbcli_dev-10.4.0 → sbcli_dev-10.7.1}/simplyblock_cli/cli.py

@@ -47,6 +47,7 @@ class CLIWrapper:
         sub_command.add_argument("--max-prov", help='Maximum amount of GB to be provisioned via all storage nodes', dest='max_prov')
         sub_command.add_argument("--number-of-distribs", help='The number of distirbs to be created on the node', dest='number_of_distribs', type=int, default=2)
         sub_command.add_argument("--number-of-devices", help='Number of devices per storage node if it\'s not supported EC2 instance', dest='number_of_devices', type=int)
+        sub_command.add_argument("--size-of-device", help='Size of device per storage node', dest='partition_size')
         sub_command.add_argument("--cpu-mask", help='SPDK app CPU mask, default is all cores found', dest='spdk_cpu_mask')
 
         sub_command.add_argument("--spdk-image", help='SPDK image uri', dest='spdk_image')
@@ -58,7 +59,7 @@ class CLIWrapper:
         sub_command.add_argument("--disable-ha-jm", help='Disable HA JM for distrib creation', action='store_false', dest='enable_ha_jm', default=True)
         sub_command.add_argument("--is-secondary-node", help='add as secondary node', action='store_true', dest='is_secondary_node', default=False)
         sub_command.add_argument("--namespace", help='k8s namespace to deploy on',)
-
+        sub_command.add_argument("--id-device-by-nqn", help='Use device nqn to identify it instead of serial number', action='store_true', dest='id_device_by_nqn', default=False)
 
         # delete storage node
         sub_command = self.add_sub_command(subparser, "delete", 'Delete storage node obj')
@@ -798,6 +799,8 @@ class CLIWrapper:
                     number_of_distribs=number_of_distribs,
                     enable_ha_jm=enable_ha_jm,
                     is_secondary_node=args.is_secondary_node,
+                    id_device_by_nqn=args.id_device_by_nqn,
+                    partition_size=args.partition_size,
                 )
 
                 return out

{sbcli_dev-10.4.0 → sbcli_dev-10.7.1}/simplyblock_core/cluster_ops.py

@@ -130,6 +130,10 @@ def create_cluster(blk_size, page_size_in_blocks, cli_pass,
         if c.swarm.attrs and "ID" in c.swarm.attrs:
             logger.info("Docker swarm found, leaving swarm now")
             c.swarm.leave(force=True)
+            try:
+                c.volumes.get("monitoring_grafana_data").remove(force=True)
+            except:
+                pass
             time.sleep(3)
 
         c.swarm.init(DEV_IP)
@@ -160,7 +164,7 @@ def create_cluster(blk_size, page_size_in_blocks, cli_pass,
     if prov_cap_crit and prov_cap_crit > 0:
         c.prov_cap_crit = prov_cap_crit
     if distr_ndcs == 0 and distr_npcs == 0:
-        c.distr_ndcs = 4
+        c.distr_ndcs = 1
         c.distr_npcs = 1
     else:
         c.distr_ndcs = distr_ndcs
@@ -365,7 +369,7 @@ def cluster_activate(cl_id, force=False, force_lvstore_create=False):
         else:
             ret = storage_node_ops.create_lvstore(snode, cluster.distr_ndcs, cluster.distr_npcs, cluster.distr_bs,
                                               cluster.distr_chunk_bs, cluster.page_size_in_blocks, max_size, snodes)
-        if not ret:
+        if not ret and not force:
             logger.error("Failed to activate cluster")
             set_cluster_status(cl_id, ols_status)
             return False
@@ -377,7 +381,7 @@ def cluster_activate(cl_id, force=False, force_lvstore_create=False):
             continue
 
         ret = storage_node_ops.recreate_lvstore(snode)
-        if not ret:
+        if not ret and not force:
             logger.error("Failed to activate cluster")
             set_cluster_status(cl_id, ols_status)
             return False

{sbcli_dev-10.4.0 → sbcli_dev-10.7.1}/simplyblock_core/constants.py

@@ -73,10 +73,10 @@ SIMPLY_BLOCK_CLI_NAME = get_from_env_var_file(
         "SIMPLY_BLOCK_COMMAND_NAME", "sbcli")
 TASK_EXEC_INTERVAL_SEC = 10
 TASK_EXEC_RETRY_COUNT = 8
-SIMPLY_BLOCK_SPDK_CORE_IMAGE = "simplyblock/spdk-core:v24.05.x-latest"
+SIMPLY_BLOCK_SPDK_CORE_IMAGE = "simplyblock/spdk-core:v24.05-tag-latest"
 
 
-SIMPLY_BLOCK_SPDK_ULTRA_IMAGE = "simplyblock/spdk:lvolstore-ha-snapshot-latest"
+SIMPLY_BLOCK_SPDK_ULTRA_IMAGE = "simplyblock/spdk:main-latest"
 
 GELF_PORT = 12202
 
@@ -88,7 +88,7 @@ EXTRA_LARGE_POOL_COUNT = 16000
 # EXTRA_LARGE_POOL_COUNT = 0
 EXTRA_HUGE_PAGE_MEMORY = 2147483648
 #EXTRA_SYS_MEMORY = 2147483648
-EXTRA_SYS_MEMORY = 0
+EXTRA_SYS_MEMORY = 0.15
 
 INSTANCE_STORAGE_DATA = {
         'i4i.large': {'number_of_devices': 1, 'size_per_device_gb': 468},
@@ -128,4 +128,5 @@ LVOL_NVME_CONNECT_NR_IO_QUEUES=6
 QPAIR_COUNT=32
 NVME_TIMEOUT_US=20000000
 
-NVMF_MAX_SUBSYSTEMS=50000
+NVMF_MAX_SUBSYSTEMS=50000
+HA_JM_COUNT=3

{sbcli_dev-10.4.0 → sbcli_dev-10.7.1}/simplyblock_core/controllers/caching_node_controller.py

@@ -176,7 +176,12 @@ def add_node(cluster_id, node_ip, iface_name, data_nics_list, spdk_cpu_mask, spd
         logger.error("Pod is not running, exiting")
         return False
 
-    time.sleep(20)
+    time.sleep(3)
+
+    snode.namespace = namespace
+    snode.spdk_cpu_mask = spdk_cpu_mask
+    snode.spdk_mem = spdk_mem
+    snode.spdk_image = spdk_image
 
     # creating RPCClient instance
     rpc_client = RPCClient(
@@ -216,50 +221,50 @@ def add_node(cluster_id, node_ip, iface_name, data_nics_list, spdk_cpu_mask, spd
 
     cache_size = 0
     cache_bdev = None
-    if supported_ssd_size < ssd_size:
-        logger.info(f"SSD size is bigger than the supported size, will use split bdev: {split_factor}")
-        ret = rpc_client.bdev_split(ssd_dev.nvme_bdev, split_factor)
-        cache_bdev = ret[0]
-        cache_size = int(ssd_dev.size/split_factor)
-        snode.cache_split_factor = split_factor
-    else:
-        snode.cache_split_factor = 0
-        cache_bdev = ssd_dev.nvme_bdev
-        cache_size = ssd_dev.size
-
     # if supported_ssd_size < ssd_size:
-    #     logger.info(f"SSD size is bigger than the supported size, creating partition")
-    #
-    #     nbd_device = rpc_client.nbd_start_disk(ssd_dev.nvme_bdev)
-    #     time.sleep(3)
-    #     if not nbd_device:
-    #         logger.error(f"Failed to start nbd dev")
-    #         return False
-    #
-    #     jm_percent = int((supported_ssd_size/ssd_size) * 100)
-    #     result, error = cnode_api.make_gpt_partitions(nbd_device, jm_percent)
-    #     if error:
-    #         logger.error(f"Failed to make partitions")
-    #         logger.error(error)
-    #         return False
-    #     time.sleep(3)
-    #     rpc_client.nbd_stop_disk(nbd_device)
-    #     time.sleep(1)
-    #     rpc_client.bdev_nvme_detach_controller(ssd_dev.nvme_controller)
-    #     time.sleep(1)
-    #     rpc_client.bdev_nvme_controller_attach(ssd_dev.nvme_controller, ssd_dev.pcie_address)
-    #     time.sleep(1)
-    #     rpc_client.bdev_examine(ssd_dev.nvme_bdev)
-    #     time.sleep(1)
-    #
-    #     cache_bdev = f"{ssd_dev.nvme_bdev}p1"
-    #     cache_size = int(supported_ssd_size)
-    #
+    #     logger.info(f"SSD size is bigger than the supported size, will use split bdev: {split_factor}")
+    #     ret = rpc_client.bdev_split(ssd_dev.nvme_bdev, split_factor)
+    #     cache_bdev = ret[0]
+    #     cache_size = int(ssd_dev.size/split_factor)
+    #     snode.cache_split_factor = split_factor
     # else:
-    #
+    #     snode.cache_split_factor = 0
     #     cache_bdev = ssd_dev.nvme_bdev
     #     cache_size = ssd_dev.size
 
+    if supported_ssd_size < ssd_size:
+        logger.info(f"SSD size is bigger than the supported size, creating partition")
+
+        nbd_device = rpc_client.nbd_start_disk(ssd_dev.nvme_bdev)
+        time.sleep(3)
+        if not nbd_device:
+            logger.error(f"Failed to start nbd dev")
+            return False
+
+        jm_percent = int((supported_ssd_size/ssd_size) * 100)
+        result, error = cnode_api.make_gpt_partitions(nbd_device, jm_percent)
+        if error:
+            logger.error(f"Failed to make partitions")
+            logger.error(error)
+            return False
+        time.sleep(3)
+        rpc_client.nbd_stop_disk(nbd_device)
+        time.sleep(1)
+        rpc_client.bdev_nvme_detach_controller(ssd_dev.nvme_controller)
+        time.sleep(1)
+        rpc_client.bdev_nvme_controller_attach(ssd_dev.nvme_controller, ssd_dev.pcie_address)
+        time.sleep(1)
+        rpc_client.bdev_examine(ssd_dev.nvme_bdev)
+        time.sleep(1)
+
+        cache_bdev = f"{ssd_dev.nvme_bdev}p1"
+        cache_size = int(supported_ssd_size)
+
+    else:
+
+        cache_bdev = ssd_dev.nvme_bdev
+        cache_size = ssd_dev.size
+
     logger.info(f"Cache size: {utils.humanbytes(cache_size)}")
 
     snode.cache_bdev = cache_bdev
@@ -297,7 +302,32 @@ def recreate(node_id):
         return False
 
     logger.info(f"Recreating caching node: {node_id}, status: {snode.status}")
-    snode_api = CNodeClient(f"{snode.mgmt_ip}:5000")
+    cnode_api = CNodeClient(f"{snode.mgmt_ip}:5000")
+
+    if not cnode_api.spdk_process_is_up():
+        results, err = cnode_api.spdk_process_start(
+            snode.spdk_cpu_mask, snode.spdk_mem, snode.spdk_image, snode.mgmt_ip,
+            snode.rpc_port, snode.rpc_username, snode.rpc_password, snode.namespace)
+        if not results:
+            logger.error(f"Failed to start spdk: {err}")
+            return False
+
+        retries = 20
+        while retries > 0:
+            resp, _ = cnode_api.spdk_process_is_up()
+            if resp:
+                logger.info(f"Pod is up")
+                break
+            else:
+                logger.info("Pod is not running, waiting...")
+                time.sleep(3)
+                retries -= 1
+
+        if retries == 0:
+            logger.error("Pod is not running, exiting")
+            return False
+
+        time.sleep(5)
 
     # creating RPCClient instance
     rpc_client = RPCClient(
@@ -306,7 +336,7 @@ def recreate(node_id):
         timeout=60*5, retry=5)
 
     # get new node info after starting spdk
-    node_info, _ = snode_api.info()
+    node_info, _ = cnode_api.info()
     # adding devices
     nvme_devs = addNvmeDevices(rpc_client, node_info['spdk_pcie_list'], snode)
     if not nvme_devs:
@@ -337,7 +367,7 @@ def recreate(node_id):
 
     if snode.lvols:
         for lvol in snode.lvols:
-            ret = connect(snode.get_id(), lvol.lvol_id)
+            ret = connect(snode.get_id(), lvol.lvol_id, force=True)
             if ret:
                 logger.info(f"connecting lvol {lvol.lvol_id} ... ok")
             else:
@@ -351,7 +381,7 @@ def recreate(node_id):
     return True
 
 
-def connect(caching_node_id, lvol_id):
+def connect(caching_node_id, lvol_id, force=False):
     lvol = db_controller.get_lvol_by_id(lvol_id)
     if not lvol:
         logger.error(f"LVol not found: {lvol_id}")
@@ -383,7 +413,8 @@ def connect(caching_node_id, lvol_id):
     for clvol in cnode.lvols:
         if clvol.lvol_id == lvol_id:
             logger.info(f"Already connected, dev path: {clvol.device_path}")
-            return False
+            if not force:
+                return False
 
     if cnode.cluster_id != pool.cluster_id:
         logger.error("Caching node and LVol are in different clusters")
@@ -423,7 +454,8 @@ def connect(caching_node_id, lvol_id):
     logger.debug(ret)
     if not ret:
         logger.error("Failed to create OCF bdev")
-        return False
+        if not force:
+            return False
 
     # logger.info("Creating local subsystem")
     # create subsystem (local)
@@ -443,7 +475,8 @@ def connect(caching_node_id, lvol_id):
 
     if not ret:
         logger.error(f"Failed to add: {cach_bdev} to the subsystem: {subsystem_nqn}")
-        return False
+        if not force:
+            return False
 
     logger.info("Connecting to local subsystem")
     # make nvme connect to nqn
@@ -451,7 +484,8 @@ def connect(caching_node_id, lvol_id):
     ret, _ = cnode_client.connect_nvme('127.0.0.1', "4420", subsystem_nqn)
     if not ret:
         logger.error("Failed to connect to local subsystem")
-        return False
+        if not force:
+            return False
 
     if cnode.multipathing:
         snode = db_controller.get_storage_node_by_id(lvol.node_id)
@@ -460,7 +494,7 @@ def connect(caching_node_id, lvol_id):
             ret, _ = cnode_client.connect_nvme(ip, "4420", subsystem_nqn)
             break
 
-    time.sleep(5)
+    time.sleep(3)
     cnode_info, _ = cnode_client.info()
     nvme_devs = cnode_info['nvme_devices']
     dev_path = None
@@ -625,8 +659,8 @@ def deploy(ifname):
             f"DOCKER_IP={dev_ip}"
         ]
     )
-    logger.info("Pulling SPDK images")
-    node_docker.images.pull(constants.SIMPLY_BLOCK_DOCKER_IMAGE)
+    logger.info(f"Pulling SPDK image {constants.SIMPLY_BLOCK_SPDK_CORE_IMAGE}")
+    node_docker.images.pull(constants.SIMPLY_BLOCK_SPDK_CORE_IMAGE)
     return f"{dev_ip}:5000"