Deployment method of intelligent recommendation training service based on Dolphincheduler

Abstract

The invention discloses an intelligent recommendation training service deployment method based on Dolphinschema. The intelligent recommendation training service deployment method comprises the steps that S1, a Dolphinschema scheduling system is deployed, basic software is installed, the front end and the rear end of the Dolphinschema scheduling system are deployed, basic software configuration associated with the Dolphinschema scheduling system is configured, and the Dolphinschema scheduling system is switched on; s2, packaging the python virtual environment and the intelligent recommendation training service, packaging a python dependency package on which the intelligent recommendation training service depends into the python virtual environment so as to be called by the intelligent recommendation training service during operation, and packaging a code file, a script file and related files of the intelligent recommendation training service so as to be operated by the intelligent recommendation training service; and S3, operating the Dolphinschner scheduling system, configuring an intelligent recommendation training service operation scheme at the front end of the Dolphinschner scheduling system, operating the intelligent recommendation training service after the configuration is completed, and monitoring the execution condition of the intelligent recommendation training service. The method is simple, high in efficiency, reliable, feasible and high in practicability, and the project deployment time is greatly shortened after the method is used.

Description

technical field [0001] The invention belongs to the technical field of machine learning, and more specifically relates to a Dolphinscheduler-based deployment method for an intelligent recommendation training service. Background technique [0002] At present, the existing intelligent recommendation training service is mainly deployed on a desktop cluster environment without a scheduling system, and the training environment needs to be manually configured. However, such an environment requires clustered deployment, which is very complicated and prone to single point of failure. It does not support multi-tenancy. It is not flexible enough to use in the field of machine learning and big data platform business. You can see the status of the task, but you cannot visualize and easily monitor the indicators of the service and the cluster environment. Contents of the invention [0003] In order to solve the above problems, the present invention provides a Dolphinscheduler-based de...

AI Technical Summary

Problems solved by technology

However, such an environment requires clustered deployment, which is very complicated and prone to single point of failure. It does not support multi-tenancy. It is not flexible enough to use in the field of machine learning and big data platform business. Can see task status, cannot visualize and easily monitor various indicators of services and cluster environments

Images