Large disk space usage due to pgsql_tmp files

High space usage on a query engine pod in the /var/lib/immuta/postgresql/data directory

On kubernetes or other self-managed installations that still operate with the Query Engine pods deployed it is possible that disk usage can reach full capacity and cause an outage. This is rare but can happen if there is usually an issue with the Query Engine deployment itself.

You can dig into the Query Engine pod itself with the following command:

kubectl exec -n <NAMESAPCE> -it <prefix>-immuta-query-engine-0 -- bash
df -h

This will show you the disk usage. The usual volume is /var/lib/immuta/postgresql/data and will contain numerous files named pgsql_tmpXXXX.XX. DO NOT delete these files, doing so might corrupt the database.

Simply perform a rollout restart with the following command to allow the Query Engine to consume the files and restore itself properly:

kubectl rollout restart deployment/<prefix>-immuta-query-engine

If the directory does not reduce in size following this then please open a support case with the Immuta Support Team.