I'm running a low-RAM VPS with CentOS 8. I've noticed that firewalld
service uses way too much RAM (up to 20%). So I guess it may be better to switch to use only built-in nftables
.
I'm quite familiar with old iptables
as well as firewalld
syntax.
However, I'm not familiar with nftables
syntax and configuration (and currently I cannot afford learning it and manual rewrite due to time restrictions).
Is it possible to export existing firewalld
rules so nftables
would be using them "directly" and it will be possible to completely disable firewalld
service?
P.S. Will it be possible for fail2ban
to work with nftables
directly in such configuration?
EDIT: It's not a question of budget/RAM increase. This VM is running on existing on-premise infrastructure, without spare resources left, so its RAM is limited to 256MB (OpenVZ), and, unfortunately, cannot be easily enlarged - without updating the infrastructure, but it's out of my sphere of responsibility (and rebuilding the infrastructure will require way more time than simply exporting configuration from firewalld to nftables).
Currently it's running OK, but uses almost all RAM, so I'm looking for ways to optimize it. AFAIK, firewalld
is just a front-end to nftables
, so my idea is to use it only for configuration, and leave the main job to nftables
- if that's possible.