What do you run; Opnsense, pfsense, Smoothwall, maybe a WAF like wazuh?
Today was update/audit firewall day. I’m running a standalone instance of pFsense on a Protectli Vault FW4B - 4 Port - Intel Quad Core - 8GB RAM - 120GB mSATA SSD with unbound, pfBlockerNG, Suricata, ntopng, and heavily filtered. I did bump the swap to 8 GB as I’ve previously noticed a few ‘out of swap’ errors under load.
Before I signed off, I ran it through a couple porn sites to see if my adblocking strategy was working. Not one intrusive ad. Sweet!
Show me what you got.
Opnsense on a thin client, riser with a quad port Intel NIC.
Nothing fancy, old ubiquiti gateway with a dedicated pihole server for my DNS.
Same. What’s the deal with having elaborate firewall stuff for a normal family home anyway?
If the built in stuff isn’t good enough then 99.9% of households would be compromised a long time ago already.
The last stats I remember reading cited some 1.5 million home networks are compromised on a daily basis. Some people, such as myself, run more complex services on their local servers that are perhaps tied into remotes such as VPS. You’ll see a lot of selfhosters with rather elaborate firewall defenses set up. I self host a lot of services I use that the ‘normal family home’ would outsource to public entities. I have a rack in the closet and several VPS, so I need something more than just Windows Firewall, or similar, that I can dial in to my unique environment.
Also, because I can.
Some of it is for fun and testing, learning. Which I used to do. I used to have an old watchdog that I put pfsense on, just don’t need it nowadays.
Once i learn how it works and have run through the setup, I move on. Just need to spend my time in other areas, but now I have an understanding of it and can apply that logic or idea to other things and troubleshooting.
Pfsense guy here, and professionally Palo alto guy. Can someone tl;dr the purpose of blockerng and suricata? I thought I remember the Lawrence systems folks mentioning using it for IPS but with segmentation at home “human” IPS seems more relevant than digital
OpenWRT on a Linksys router, with adguard home for DNS blocking.
I used to run OPNSense on some older x86 hardware, but wanted to move to something simpler and less power hungry.
I use firewalld with a script that automatically updates a blocklist of known shady IPs.
I think I have the same protectli as you and it is awesome. Need it for my 2.5gb uplink. I use openwrt on it… Didn’t really like opnsense. I am more used to linux than bsd.
I host lots of services and get bombarded by scrapers, scanners, and skids both at home and on my VPSs.
I use ipset for the usual blocklists which I download regularly. I also have tarpits on 22/tcp (endlessh). I pipe the IPs from the endlessh logs into fail2ban which feeds the ipsets. I have ipset blocks and fail2ban on my home firewall and all VPSs and coordinate over mqtt. So
any fail2ban trigger > mqtt > every ipset block. Touch my 22/tcp anywhere and you get banned instantly everywhere. The program I use for this is called vallumd and it runs on openwrt.I also put maltrail everywhere but I’m not totally sure how to interpret and respond to the results. Probably will implement a pipe from maltrail to my mqtt > blocklist setup.
I don’t do any network-level adblocking… Might be a future project.
I’ve been using Ufw but airvpn’s kill switch seems to override it, should i be using something else?
I have found that a lot of VPN kill switches interfere with other security measures. For instance, I use tailscale on my VPS. I also run a local VPN. If I have the kill switch on the local VPN engaged, it interferes with tailscale and I cannot ssh in to my VPS. So, a not so elegant solution for me is to disengage the local VPN’s kill switch for that session, and then re-enable it after I am finished administering my VPS. After which I will do a DNS leak check to make sure everything is as it was. Takes a couple of quick steps, but it seems to work.
Show me what you got.
you’re doing the same thing i am, so there’s not point. lol
Yeah, but you got charts n’ graphs and a big writeup. Nice job.
pfSense on this:
https://a.co/d/6WpafWQI also block outgoing port 53 only allowing my Pihole through.
I use Tailscale to access the network while away.
Do you run unbound on pFsense?
No my pfSense setup is fairly minimal
I run IPFire on a PC Engines apu4d4 (https://pcengines.ch/apu4d4.htm). I use dynDNS, WireGuard and set up a DMZ with it. I also have a WiFi card installed und use hostAPD to run that.
I think they stopped producing them because the AMD SOC they used is EOL. I was a big fan of their open platform.
OpenWRT.
I’ve always wondered about OpenWRT. In my uneducated thinking, running an access point/wifi, firewall, router, etc, all in the same package would create a bottleneck right at the point you wouldn’t want it. What has been your experience?
Everything works fine. It’s super handy having such fine control over my router.
Opnsense with unbound DNS here. Running on an old PC that got converted to dedicated firewall (with added NIC card for ports). Nothing crazy, just enough to control what communicates out of my network.
Used to do the same thing with an old PC. Hell, at one time I was running one off a laptop with USB to RJ45 adapters for the WAN/LAN ports.
Hiding behind my firewalls. Shhhhh.
Sitting in my bunker
Hid behind my wall…
In perfect isolation here behind my wall
Waiting for the worms to come
Nock nock, someone’s home?
RST
Ubiquiti DM pro with its built in suricata. Honeypots, no remote mgmt, ACLs to minimum need, HA networks in isolation. DPI, multiple pi-holes. Phone alerts on intrusion wazuh just for node security compliance. ManageEngine for patches. NTFY alerts on console access.
It’s not perfect
nftables. Deny all, start adding stuff until þings work.
My firewalls are simple, b/c I run a private VPN and just shut off all traffic except over WG. I’ve got one exposed VPS reverse proxying services from oþer VPSes over WG.
But: nftables, and only nftables. I’m a big believer in understanding how stuff works, and þe rulesets created by firewalld and ilk are convoluted - complexity adds risk.
Haha, I thought that said “until pings work”
Also an accurate reading.
Rock on!
