feathr-ai / feathr

Feathr – A scalable, unified data and AI engineering platform for enterprise
https://join.slack.com/t/feathrai/shared_invite/zt-1ffva5u6v-voq0Us7bbKAw873cEzHOSg
Apache License 2.0
1.99k stars 260 forks source link

Features Registry Error with mssql #1075

Open mansoorsyed11 opened 1 year ago

mansoorsyed11 commented 1 year ago

Hey team @xiaoyongzhu, I'm trying to register the features in mssql im getting the below error. I have tried different formats of paths as below.

Format 1.

api_version: 1 feature_registry: api_endpoint: mssql+pyodbc://userid:password@host:port/database?driver=ODBC+Driver+17+for+SQL+Server&autocommit=True

ERROR:

No connection adapters were found for 'mssql+pyodbc://userid:password@host:port/database?driver=ODBC+Driver+17+for+SQL+Server&autocommit=True/projects' .

Format 2.

api_version: 1 feature_registry: api_endpoint: https: //mssql+pyodbc://userid:password@host:port/database?driver=ODBC+Driver+17+for+SQL+Server&autocommit=True

ERROR 2:

HTTPSConnectionPool(host='mssql+pyodbc', port=443): Max retries exceeded with url: //userid:password@host:1433/sqld-mlops-dev?driver=ODBC+Driver+17+for+SQL+Server&autocommit=True/projects (Caused by NewConnectionError('<urllib3.connection.HTTPSConnection object at 0x7fea17f46640>: Failed to establish a new connection: [Errno -2] Name or service not known'))

Format 3 :

api_version: 1 feature_registry: api_endpoint: mssql://userid:password@host:port/database

Error 3:

HTTPSConnectionPool(host='mssql+pyodbc', port=443): Max retries exceeded with url: //userid:password@host:1433/sqld-mlops-dev/projects (Caused by NewConnectionError('<urllib3.connection.HTTPSConnection object at 0x7fea17f46640>: Failed to establish a new connection: [Errno -2] Name or service not known'))

Note: Like i tried different formats of api to connect sql its not working help me to solve this problem thanks @xiaoyongzhu .

blrchen commented 1 year ago

Add this week's oncall @enya-yx to help

enya-yx commented 1 year ago

@mansoorsyed11 Do you mean the 'api_endpoint' in configs? You should set an web service first and use the url (something like: ...webapp.azurewebsites.net) as this 'api_endpoint', You can find some examples in our samples: https://github.com/feathr-ai/feathr/tree/main/docs/samples

mansoorsyed11 commented 1 year ago

@enya-yx No im planning to use an SQL server other than azure preview there is no document on connecting sql server to register the features. correct me if I'm wrong azure purview is optional if we use sql server to register the features according to your documentation?

mansoorsyed11 commented 1 year ago

@enya-yx can please elaborate more on this problem and how to register the features in mssql.

enya-yx commented 1 year ago

@enya-yx can please elaborate more on this problem and how to register the features in mssql.

@mansoorsyed11 You need to setup a webapp and set some envs in its configuration so that it can connect to your sql server. Here's a doc that might be helpful: https://github.com/feathr-ai/feathr/blob/main/docs/dev_guide/build-and-push-feathr-registry-docker-image.md

Currently, feathr only supports registering features through webapp instead of connecting to the sql server directly.

enya-yx commented 1 year ago

add this week's on call @windoze who is expertise in register by sql

windoze commented 1 year ago

@mansoorsyed11 Feathr doesn't support writing features into MSSQL database directly, a registry service needs to be deployed separately. Please refer to https://feathr-ai.github.io/feathr/concepts/feature-registry.html for more information about the registry service.