You could clear cache with the endpoint `/cache/clear`
- make sure core servers are down. Run the stop command “service paxata-server stop”
- Run command “curl -X POST http://<WORKER_NODE>/cache/clear?pretty=true”
- restart the pipeline service so it will build out the filesystem structure and start writing there again
If the /cache/clear doesn't work, you can run “rm -rf /u01/paxata/cache/*” on the worker.