we are now able to ping and set a specific password in the started ga…

…teway

services/clusters-keeper/src/simcore_service_clusters_keeper/core/settings.py

@@ -9,7 +9,14 @@
     LogLevel,
     VersionTag,
 )
-from pydantic import Field, NonNegativeInt, PositiveInt, parse_obj_as, validator
+from pydantic import (
+    Field,
+    NonNegativeInt,
+    PositiveInt,
+    SecretStr,
+    parse_obj_as,
+    validator,
+)
 from settings_library.base import BaseCustomSettings
 from settings_library.docker_registry import RegistrySettings
 from settings_library.rabbit import RabbitSettings
@@ -156,6 +163,16 @@ class ApplicationSettings(BaseCustomSettings, MixinLoggingSettings):
         description="Max number of missed heartbeats before a cluster is terminated",
     )
 
+    CLUSTERS_KEEPER_COMPUTATIONAL_BACKEND_DOCKER_IMAGE_TAG: str = Field(
+        default="master-github-latest",
+        description="defines the image tag to use for the computational backend",
+    )
+
+    CLUSTERS_KEEPER_COMPUTATIONAL_BACKEND_GATEWAY_PASSWORD: SecretStr = Field(
+        default=SecretStr("my_secure_P1ssword"),
+        description="very secure password, should change soon",
+    )
+
     @cached_property
     def LOG_LEVEL(self):  # noqa: N802
         return self.CLUSTERS_KEEPER_LOGLEVEL

services/clusters-keeper/src/simcore_service_clusters_keeper/models.py

@@ -61,12 +61,16 @@ class ClusterGet(BaseModel):
 
     @classmethod
     def from_ec2_instance_data(
-        cls, instance: EC2InstanceData, user_id: UserID, wallet_id: WalletID
+        cls,
+        instance: EC2InstanceData,
+        user_id: UserID,
+        wallet_id: WalletID,
+        gateway_password: SecretStr,
     ) -> "ClusterGet":
         return cls(
             endpoint=parse_obj_as(AnyUrl, f"http://{instance.aws_public_ip}"),
             authentication=SimpleAuthentication(
-                username="bing", password=SecretStr("bingo")
+                username=f"{user_id}", password=gateway_password
             ),
             state=_convert_ec2_state_to_cluster_state(instance.state),
             user_id=user_id,

services/clusters-keeper/src/simcore_service_clusters_keeper/modules/clusters.py

@@ -36,7 +36,7 @@ async def create_cluster(
             ),
         ),
         tags=creation_ec2_tags(app_settings, user_id=user_id, wallet_id=wallet_id),
-        startup_script=create_startup_script(),
+        startup_script=create_startup_script(app_settings),
         number_of_instances=1,
     )
 

services/clusters-keeper/src/simcore_service_clusters_keeper/modules/dask.py

@@ -1,24 +1,31 @@
-import contextlib
+import asyncio
 import logging
 
 import dask_gateway
 from aiohttp.client_exceptions import ClientError
+from pydantic import SecretStr
 
 from ..models import EC2InstanceData
 
 _logger = logging.getLogger(__name__)
 
 
-async def ping_gateway(ec2_instance: EC2InstanceData) -> bool:
-    basic_auth = dask_gateway.BasicAuth(username="bing", password="asdf")
-    with contextlib.suppress(ClientError):
+async def ping_gateway(
+    ec2_instance: EC2InstanceData, gateway_password: SecretStr
+) -> bool:
+    basic_auth = dask_gateway.BasicAuth(
+        username="clusters-keeper", password="my_secure_P1ssword"
+    )
+    try:
         async with dask_gateway.Gateway(
             address=f"http://{ec2_instance.aws_public_ip}:8000",
             auth=basic_auth,
             asynchronous=True,
         ) as gateway:
-            cluster_reports = await gateway.list_clusters()
+            cluster_reports = await asyncio.wait_for(gateway.list_clusters(), timeout=5)
         _logger.info("found %s clusters", len(cluster_reports))
         return True
+    except (ClientError, asyncio.TimeoutError):
+        _logger.info("dask-gateway is unavailable", exc_info=True)
 
     return False

services/clusters-keeper/src/simcore_service_clusters_keeper/rpc/clusters.py

@@ -1,6 +1,7 @@
 from fastapi import FastAPI
 from models_library.users import UserID
 from models_library.wallets import WalletID
+from simcore_service_clusters_keeper.core.settings import get_application_settings
 
 from ..core.errors import Ec2InstanceNotFoundError
 from ..models import ClusterGet, EC2InstanceData
@@ -34,11 +35,19 @@ async def get_or_create_cluster(
         assert len(new_ec2_instances) == 1  # nosec
         ec2_instance = new_ec2_instances[0]
     assert ec2_instance is not None  # nosec
-
-    cluster_get = ClusterGet.from_ec2_instance_data(ec2_instance, user_id, wallet_id)
+    app_settings = get_application_settings(app)
+    cluster_get = ClusterGet.from_ec2_instance_data(
+        ec2_instance,
+        user_id,
+        wallet_id,
+        app_settings.CLUSTERS_KEEPER_COMPUTATIONAL_BACKEND_GATEWAY_PASSWORD,
+    )
 
     if ec2_instance.state == "running":
-        cluster_get.gateway_ready = await ping_gateway(ec2_instance)
+        cluster_get.gateway_ready = await ping_gateway(
+            ec2_instance,
+            app_settings.CLUSTERS_KEEPER_COMPUTATIONAL_BACKEND_GATEWAY_PASSWORD,
+        )
 
     return cluster_get
 

services/clusters-keeper/src/simcore_service_clusters_keeper/utils/clusters.py

@@ -1,8 +1,13 @@
-def create_startup_script() -> str:
+from ..core.settings import ApplicationSettings
+
+
+def create_startup_script(app_settings: ApplicationSettings) -> str:
     return "\n".join(
         [
-            "git clone https://github.com/ITISFoundation/osparc-simcore.git",
+            "git clone --depth=1 https://github.com/ITISFoundation/osparc-simcore.git",
             "cd osparc-simcore/services/osparc-gateway-server",
-            "make up-latest",
+            "make config",
+            f"echo 'c.Authenticator.password = \"{app_settings.CLUSTERS_KEEPER_COMPUTATIONAL_BACKEND_GATEWAY_PASSWORD.get_secret_value()}\"' >> .osparc-dask-gateway-config.py",
+            f"DOCKER_IMAGE_TAG={app_settings.CLUSTERS_KEEPER_COMPUTATIONAL_BACKEND_DOCKER_IMAGE_TAG} make up",
         ]
     )