I can't tell where you're getting this message from, but kernel logs would mention "OOMKill", so I'm assuming this isn't from the kernel itself, but another monitor logging it and reporting on it perhaps.
Create extra swap and that will help prevent this.
