Skip to content

dev/mip-275/nodes dbs' cleanup after algorithm execution

Kostas FILIPPOPOLITIS requested to merge dev/MIP-275/cleanup_after_algorithm_execution into master

Created by: apmariglis

https://team-1617704806227.atlassian.net/browse/MIP-275

  • Clean up procedure removed form the AlgorithmExecutor class
  • Instead an asyncio method cleanup_loop(self) in the Controller class checks every CLEANUP_INTERVAL if there are leftover tables to be cleaned up

The way this works is that, inside the Controller, after an algorithm finishes executing (successfully or not) the node_ids of all the participated nodes are stored in the self._nodes_for_cleanup dictionary along with the context_id of the execution and are then processed for dropping in the cleanup_loop method

Merge request reports