Skip to content

Orchestrator - Research a non-intrusive way to retry transactions when SQL server connection is lost #201

@Ark-kun

Description

@Ark-kun
sqlalchemy.exc.OperationalError: (pymysql.err.OperationalError) (2013, 'Lost connection to MySQL server during query')
[SQL: SELECT output_artifact_link.output_name, artifact_node.id, artifact_node.had_data_in_past, artifact_node.may_have_data_in_future, artifact_node.type_name, artifact_node.type_properties, artifact_node.producer_execution_id, artifact_node.producer_output_name, artifact_node.artifact_data_id, artifact_node.extra_data 
FROM output_artifact_link INNER JOIN artifact_node ON artifact_node.id = output_artifact_link.artifact_id 
WHERE output_artifact_link.execution_id = %(execution_id_1)s]
[parameters: {'execution_id_1': '019d44d481dc3efa2d3a'}]
(Background on this error at: https://sqlalche.me/e/20/e3q8)

From the first look, this seems non-trivial. Some SqlAlchemy docs say that in this case transaction needs to be rolled back and re-created from scratch.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions