Storage Tiering implementation

simplyblock_cli/cli-reference.yaml

@@ -1085,6 +1085,41 @@ commands:
             dest: strict_node_anti_affinity
             type: bool
             action: store_true
+          - name: "--support-storage-tiering"
+            help: "Enable storage tiering for the cluster"
+            dest: storage_tiering
+            type: bool
+            default: false
+          - name: "--s3-endpoint"
+            help: "S3 endpoint for storage tiering"
+            dest: s3_endpoint
+            type: str
+            default: ""
+          - name: "--s3-bucket"
+            help: "S3 bucket for storage tiering"
+            dest: s3_bucket
+            type: str
+            default: ""
+          - name: "--s3-access-key"
+            help: "S3 access key for storage tiering"
+            dest: s3_access_key
+            type: str
+            default: ""
+          - name: "--s3-secret-key"
+            help: "S3 secret key for storage tiering"
+            dest: s3_secret_key
+            type: str
+            default: ""
+          - name: "--s3-workerpool-mask"
+            help: "S3 workerpool mask for storage tiering"
+            dest: s3_workerpool_mask
+            type:
+              regex: "^(0x|0X)?[a-fA-F0-9]+$"
+          - name: "--s3-workerpool-size"
+            help: "S3 workerpool size for storage tiering"
+            dest: s3_workerpool_size
+            type: int
+            default: 32
       - name: add
         help: "Adds a new cluster"
         arguments:
@@ -1185,6 +1220,41 @@ commands:
             dest: strict_node_anti_affinity
             type: bool
             action: store_true
+          - name: "--support-storage-tiering"
+            help: "Enable storage tiering for the cluster"
+            dest: storage_tiering
+            type: bool
+            default: false
+          - name: "--s3-endpoint"
+            help: "S3 endpoint for storage tiering"
+            dest: s3_endpoint
+            type: str
+            default: ""
+          - name: "--s3-bucket"
+            help: "S3 bucket for storage tiering"
+            dest: s3_bucket
+            type: str
+            default: ""
+          - name: "--s3-access-key"
+            help: "S3 access key for storage tiering"
+            dest: s3_access_key
+            type: str
+            default: ""
+          - name: "--s3-secret-key"
+            help: "S3 secret key for storage tiering"
+            dest: s3_secret_key
+            type: str
+            default: ""
+          - name: "--s3-workerpool-mask"
+            help: "S3 workerpool mask for storage tiering"
+            dest: s3_workerpool_mask
+            type:
+              regex: "^(0x|0X)?[a-fA-F0-9]+$"
+          - name: "--s3-workerpool-size"
+            help: "S3 workerpool size for storage tiering"
+            dest: s3_workerpool_size
+            type: int
+            default: 32
       - name: activate
         help: >
           Activates a cluster.
@@ -1986,6 +2056,13 @@ commands:
             dest: resize
             type: size
             default: "0"
+      - name: backup
+        help: "Backs up a snapshot to a s3 storage using storage tiering"
+        arguments:
+          - name: "snapshot_id"
+            help: "Snapshot id"
+            dest: snapshot_id
+            type: str
   - name: "caching-node"
     help: "Caching node commands"
     aliases:

simplyblock_cli/cli.py

@@ -440,6 +440,13 @@ def init_cluster__create(self, subparser):
         if self.developer_mode:
             argument = subcommand.add_argument('--enable-qos', help='Enable qos bdev for storage nodes, true by default', type=bool, default=False, dest='enable_qos')
         argument = subcommand.add_argument('--strict-node-anti-affinity', help='Enable strict node anti affinity for storage nodes. Never more than one chunk is placed on a node. This requires a minimum of _data-chunks-in-stripe + parity-chunks-in-stripe + 1_ nodes in the cluster.', dest='strict_node_anti_affinity', action='store_true')
+        argument = subcommand.add_argument('--support-storage-tiering', help='Enable storage tiering for the cluster', type=bool, default=False, dest='storage_tiering')
+        argument = subcommand.add_argument('--s3-endpoint', help='S3 endpoint for storage tiering', type=str, default='', dest='s3_endpoint')
+        argument = subcommand.add_argument('--s3-bucket', help='S3 bucket for storage tiering', type=str, default='', dest='s3_bucket')
+        argument = subcommand.add_argument('--s3-access-key', help='S3 access key for storage tiering', type=str, default='', dest='s3_access_key')
+        argument = subcommand.add_argument('--s3-secret-key', help='S3 secret key for storage tiering', type=str, default='', dest='s3_secret_key')
+        argument = subcommand.add_argument('--s3-workerpool-mask', help='S3 workerpool mask for storage tiering', type=regex_type(r'^(0x|0X)?[a-fA-F0-9]+$'), dest='s3_workerpool_mask')
+        argument = subcommand.add_argument('--s3-workerpool-size', help='S3 workerpool size for storage tiering', type=int, default=32, dest='s3_workerpool_size')
 
     def init_cluster__add(self, subparser):
         subcommand = self.add_sub_command(subparser, 'add', 'Adds a new cluster')
@@ -465,6 +472,13 @@ def init_cluster__add(self, subparser):
         if self.developer_mode:
             argument = subcommand.add_argument('--enable-qos', help='Enable qos bdev for storage nodes, default: true', type=bool, default=False, dest='enable_qos')
         argument = subcommand.add_argument('--strict-node-anti-affinity', help='Enable strict node anti affinity for storage nodes. Never more than one chunk is placed on a node. This requires a minimum of _data-chunks-in-stripe + parity-chunks-in-stripe + 1_ nodes in the cluster."', dest='strict_node_anti_affinity', action='store_true')
+        argument = subcommand.add_argument('--support-storage-tiering', help='Enable storage tiering for the cluster', type=bool, default=False, dest='storage_tiering')
+        argument = subcommand.add_argument('--s3-endpoint', help='S3 endpoint for storage tiering', type=str, default='', dest='s3_endpoint')
+        argument = subcommand.add_argument('--s3-bucket', help='S3 bucket for storage tiering', type=str, default='', dest='s3_bucket')
+        argument = subcommand.add_argument('--s3-access-key', help='S3 access key for storage tiering', type=str, default='', dest='s3_access_key')
+        argument = subcommand.add_argument('--s3-secret-key', help='S3 secret key for storage tiering', type=str, default='', dest='s3_secret_key')
+        argument = subcommand.add_argument('--s3-workerpool-mask', help='S3 workerpool mask for storage tiering', type=regex_type(r'^(0x|0X)?[a-fA-F0-9]+$'), dest='s3_workerpool_mask')
+        argument = subcommand.add_argument('--s3-workerpool-size', help='S3 workerpool size for storage tiering', type=int, default=32, dest='s3_workerpool_size')
 
     def init_cluster__activate(self, subparser):
         subcommand = self.add_sub_command(subparser, 'activate', 'Activates a cluster.')
@@ -785,6 +799,7 @@ def init_snapshot(self):
         self.init_snapshot__list(subparser)
         self.init_snapshot__delete(subparser)
         self.init_snapshot__clone(subparser)
+        self.init_snapshot__backup(subparser)
 
 
     def init_snapshot__add(self, subparser):
@@ -807,6 +822,10 @@ def init_snapshot__clone(self, subparser):
         subcommand.add_argument('lvol_name', help='Logical volume name', type=str)
         argument = subcommand.add_argument('--resize', help='New logical volume size: 10M, 10G, 10(bytes). Can only increase.', type=size_type(), default='0', dest='resize')
 
+    def init_snapshot__backup(self, subparser):
+        subcommand = self.add_sub_command(subparser, 'backup', 'Backs up a snapshot to a s3 storage using storage tiering')
+        subcommand.add_argument('snapshot_id', help='Snapshot id', type=str)
+
 
     def init_caching_node(self):
         subparser = self.add_command('caching-node', 'Caching node commands', aliases=['cn',])
@@ -1214,6 +1233,8 @@ def run(self):
                     ret = self.snapshot__delete(sub_command, args)
                 elif sub_command in ['clone']:
                     ret = self.snapshot__clone(sub_command, args)
+                elif sub_command in ['backup']:
+                    ret = self.snapshot__backup(sub_command, args)
                 else:
                     self.parser.print_help()
 

simplyblock_cli/clibase.py

@@ -628,6 +628,10 @@ def snapshot__clone(self, sub_command, args):
         success, details = snapshot_controller.clone(args.snapshot_id, args.lvol_name, new_size)
         return details
 
+    def snapshot__backup(self, sub_command, args):
+        snapshot_id = args.snapshot_id
+        return snapshot_controller.backup(snapshot_id)
+
     def caching_node__deploy(self, sub_command, args):
         return caching_node_controller.deploy(args.ifname)
 
@@ -688,18 +692,26 @@ def cluster_add(self, args):
         distr_bs = args.distr_bs
         distr_chunk_bs = args.distr_chunk_bs
         ha_type = args.ha_type
-
         enable_node_affinity = args.enable_node_affinity
         qpair_count = args.qpair_count
         max_queue_size = args.max_queue_size
         inflight_io_threshold = args.inflight_io_threshold
         enable_qos = args.enable_qos
         strict_node_anti_affinity = args.strict_node_anti_affinity
+        storage_tiering = args.storage_tiering
+        s3_endpoint = args.s3_endpoint
+        s3_bucket = args.s3_bucket
+        s3_access_key = args.s3_access_key
+        s3_secret_key = args.s3_secret_key
+        s3_workerpool_mask = args.s3_workerpool_mask
+        s3_workerpool_size = args.s3_workerpool_size
 
         return cluster_ops.add_cluster(
             blk_size, page_size_in_blocks, cap_warn, cap_crit, prov_cap_warn, prov_cap_crit,
             distr_ndcs, distr_npcs, distr_bs, distr_chunk_bs, ha_type, enable_node_affinity,
-            qpair_count, max_queue_size, inflight_io_threshold, enable_qos, strict_node_anti_affinity)
+            qpair_count, max_queue_size, inflight_io_threshold, enable_qos, strict_node_anti_affinity,
+            storage_tiering, s3_endpoint, s3_bucket, s3_access_key, s3_secret_key, s3_workerpool_mask, 
+            s3_workerpool_size)
 
     def cluster_deploy(self, args):
         grafana_endpoint = ""
@@ -819,6 +831,14 @@ def cluster_create(self, args):
         enable_qos = args.enable_qos
         strict_node_anti_affinity = args.strict_node_anti_affinity
 
+        # storage_tiering = args.storage_tiering
+        # s3_endpoint = args.s3_endpoint
+        # s3_bucket = args.s3_bucket
+        # s3_access_key = args.s3_access_key
+        # s3_secret_key = args.s3_secret_key
+        # s3_workerpool_mask = args.s3_workerpool_mask
+        # s3_workerpool_size = args.s3_workerpool_size
+
         return cluster_ops.create_cluster(
             blk_size, page_size_in_blocks,
             CLI_PASS, cap_warn, cap_crit, prov_cap_warn, prov_cap_crit,

simplyblock_core/cluster_ops.py

@@ -458,7 +458,8 @@ def deploy_cluster(storage_nodes,test,ha_type,distr_ndcs,distr_npcs,enable_qos,i
 
 def add_cluster(blk_size, page_size_in_blocks, cap_warn, cap_crit, prov_cap_warn, prov_cap_crit,
                 distr_ndcs, distr_npcs, distr_bs, distr_chunk_bs, ha_type, enable_node_affinity, qpair_count,
-                max_queue_size, inflight_io_threshold, enable_qos, strict_node_anti_affinity) -> str:
+                max_queue_size, inflight_io_threshold, enable_qos, strict_node_anti_affinity, storage_tiering, 
+                s3_endpoint, s3_bucket, s3_access_key, s3_secret_key, s3_workerpool_mask, s3_workerpool_size) -> str:
     db_controller = DBController()
     clusters = db_controller.get_clusters()
     if not clusters:
@@ -493,6 +494,14 @@ def add_cluster(blk_size, page_size_in_blocks, cap_warn, cap_crit, prov_cap_warn
     cluster.max_queue_size = max_queue_size
     cluster.inflight_io_threshold = inflight_io_threshold
     cluster.enable_qos = enable_qos
+    cluster.storage_tiering = storage_tiering
+    cluster.s3_endpoint = s3_endpoint
+    cluster.s3_bucket = s3_bucket
+    cluster.s3_access_key = s3_access_key
+    cluster.s3_secret_key = s3_secret_key
+    cluster.s3_workerpool_mask = s3_workerpool_mask
+    cluster.s3_workerpool_size = s3_workerpool_size
+
     if cap_warn and cap_warn > 0:
         cluster.cap_warn = cap_warn
     if cap_crit and cap_crit > 0:

simplyblock_core/controllers/snapshot_controller.py

@@ -2,6 +2,7 @@
 import logging as lg
 import time
 import uuid
+import datetime
 
 from simplyblock_core.controllers import lvol_controller, snapshot_events, pool_controller
 
@@ -227,6 +228,39 @@ def list(all=False):
     return utils.print_table(data)
 
 
+def backup(snapshot_uuid):
+    snap = db_controller.get_snapshot_by_id(snapshot_uuid)
+    if not snap:
+        msg = f"Snapshot not found: {snapshot_uuid}"
+        logger.error(msg)
+        return False, msg
+
+    snode = db_controller.get_storage_node_by_id(snap.lvol.node_id)
+    if not snode:
+        msg = f"Storage node not found: {snap.lvol.node_id}"
+        logger.error(msg)
+        return False, msg
+
+    rpc_client = RPCClient(snode.mgmt_ip, snode.rpc_port, snode.rpc_username, snode.rpc_password)
+    ret = rpc_client.bdev_lvol_backup_snapshot(snap.snap_uuid, timeout_us=1000000, dev_page_size=2 * 1024 * 1024, nmax_retries=4, nmax_flush_jobs=4)
+    if not ret:
+        msg = f"Failed to backup snapshot: {snap.snap_bdev}"
+        logger.error(msg)
+        return False, msg
+
+    times = 5
+    while times > 0:
+        resp = rpc_client.bdev_lvol_get_snapshot_backup_status(lvol_name=snap.snap_uuid)
+        time.sleep(3)
+        times = times - 1
+        if resp == "SUCCEEDED":
+            snap.backedup_at = str(datetime.datetime.now(datetime.timezone.utc))
+            snap.write_to_db(db_controller.kv_store)
+            logger.info("Done")
+            return True, ""
+
+    return False, f"Failed to backup snapshot: {snap.snap_bdev}, status: {resp}"
+
 def delete(snapshot_uuid, force_delete=False):
     snap = db_controller.get_snapshot_by_id(snapshot_uuid)
     if not snap:

simplyblock_core/env_var

@@ -1,6 +1,5 @@
 SIMPLY_BLOCK_COMMAND_NAME=sbcli-dev
 SIMPLY_BLOCK_VERSION=18.0.90
 
-
-SIMPLY_BLOCK_DOCKER_IMAGE=simplyblock/simplyblock:main
-SIMPLY_BLOCK_SPDK_ULTRA_IMAGE=simplyblock/spdk:main-latest
+SIMPLY_BLOCK_DOCKER_IMAGE=public.ecr.aws/simply-block/simplyblock:manohar-storage-tiering3
+SIMPLY_BLOCK_SPDK_ULTRA_IMAGE=simplyblock/spdk:storage-tiering-ha-tested-sanitizer

simplyblock_core/models/cluster.py

@@ -61,6 +61,13 @@ class Cluster(BaseModel):
     strict_node_anti_affinity: bool = False
     tls: bool = False
     is_re_balancing: bool = False
+    storage_tiering: bool = True
+    s3_access_key: str = "foobar"
+    s3_secret_key: str = "barfoobarfoo"
+    s3_endpoint: str = "http://192.168.10.146:9000"
+    s3_bucket: str = "mybucket"
+    s3_workerpool_mask: str = "0x00000000"
+    s3_workerpool_size: int = 32
     full_page_unmap: bool = True
 
     def get_status_code(self):

simplyblock_core/models/snapshot.py

@@ -22,3 +22,4 @@ class SnapShot(BaseModel):
     snap_ref_id: str = ""
     snap_uuid: str = ""
     vuid: int = 0
+    backedup_at: str = ""