> > > FWIW, it can be useful to configure the OS with strict memory overcommit. > That > causes postgres to fail more gracefully, because the OOM killer won't be > invoked. >
In the current setup the database is used as an embedded db, with the application sharing the same host as the database. Setting the memory overcommit affects the other application, but configuring LimitAS for the postgresql systemd service should work. Does this mean that the fact that this query uses more than 100x the configured work_mem is not considered a bug? Should I create a bug report?
