Taming the beasts… aka suricata blueprint

Ramosel

I am in the process of installing a new pfsense firewall and v2.0 of the infamous guide would come just in handy :)

Agreed :D

I have a matching hardware spare so I've started a 2.2.1 build and am just going to hold tight until the guide comes out. I'm venturing into new territory with Suricata and would rather follow the knowledge. Until then, my 2.1.5 with Snort is running just fine.

Curious though, is there any "school of thought" as to order of loading Squid3, PfBlockerNG and Suricata?

Rick

Cino

@jflsakfja any progress? I'm pretty sure you're probably done with the write up, but still waiting on the pfSense team to give the OK?

A Former User

Negative on progress, since I still haven't got the OK. Patience is a virtue we all need ;)

Ramosel

@jflsakfja:

Patience is a virtue we all need ;)

I'll second that one too!

Rick

n3by

Hello,

On my home I use pfsense 2.1.5 and now I switched from Snort to Suricata, set it as recommended by jflsakfja instructions in this thread….

Looking at alert logs yesterday I found that China people try to probe/hack my home network ( probably they found that Lenovo tablet can't report home and want to see whats wrong... ) so I put them in an alias blocked and set them in firewall as permanent blocked traffic In ( WAN ) and Out ( LAN ) but I still get alert in Suricata from there IP.

And by the way I still have Pfblocker set up to block all incoming traffic from Asia, and other sources.

Isn't pfsense firewall blocking traffic before it arriving at Suricata ?

Thank you.

edit:
security by obscurity ( pictures removed )

Cino

@jflsakfja:

Patience is a virtue we all need ;)

After reading https://forum.pfsense.org/index.php?topic=88244 , I think we will need more then patience. I understand what you're doing/asking for.. I've seen many pfsense guides on the internet, and none of them have gotten in trouble.. They just put the standard trademark disclaimer.

I'll wait and see

A Former User

@n3by:

Hello,

On my home I use pfsense 2.1.5 and now I switched from Snort to Suricata, set it as recommended by jflsakfja instructions in this thread….

Looking at alert logs yesterday I found that China people try to probe/hack my home network ( probably they found that Lenovo tablet can't report home and want to see whats wrong... ) so I put them in an alias blocked and set them in firewall as permanent blocked traffic In ( WAN ) and Out ( LAN ) but I still get alert in Suricata from there IP.

And by the way I still have Pfblocker set up to block all incoming traffic from Asia, and other sources.

Isn't pfsense firewall blocking traffic before it arriving at Suricata ?

Thank you.

edit:
security by obscurity ( pictures removed )

Short answer is nope :).
Long answer is: Think of the way packets are processed as a realtime copy of the packet stream. pf processes the actual packets, while at the same time copying them and sending them to suricata (actually logging the stream as is, and suricata sniffing that copy, but that's contrary to my stupidly simple explanations, so ignore it).

Here's what happens and why you still get alerts:
A packet arrives from IP 1.1.1.1. That packet is copied and processed, sent on its way to your computer. At the exact time (relative, play along) that the packet is copied, suricata picks it up and starts processing the copy. The actual packet has long reached your computer, suricata is munching on a copy. After it decides that IP 1.1.1.1 is bad, it adds it to the blocked hosts. The next time a packet from that IP arrives, both pf and suricata will pick it up (original+copy). While pf is still deciding what to do with the packet (it will drop it, since the IP is known bad), suricata will still process a copy of it, and generate an alert.

The downside: wasted processing. I tried working around with it with suricata's BPF (an exercise for the extremely stubborn out there) but could not, not without significant performance penalties. Basically BPF says "ignore packets from these IPs". There is no reason to analyze a packet that's known to be blocked anyway.

The upside: each time the bad IP generates an alert, the alert timestamp is updated. If it keeps doing it while being inside the ban timeframe, then the IP is perpetually kept on blocked hosts (well, until something like a reboot kicks it (temporarily) out).

All in all, don't worry if you are still seeing alerts from blocked IPs. As long as your rules are set up correctly (easy to test) then you are set :)

fsansfil

@n3by:

On my home I use pfsense 2.1.5 and now I switched from Snort to Suricata, set it as recommended by jflsakfja instructions in this thread….

Unless you are writing your own rules or you are an ISP dealing with 30k users, right now you are better with Snort than Suricata. Using ET open and Snort VRT rulesets, youll get more coverage of threat protection with Snort (All VRT rules compatible with the engine)

I personnaly prefer Suricata for a few reasons. IP protocol detection is better tuned…it will actually alert on some IPv6, Hop-by-Hop, etc... And when writing rules and outputting to syslog, I get more info on why the rules didnt load... Plus, the loggin is more detailed with Suricata.

But the Snort OpenAppID is promising and give you a good overlook of apps on your network/ports -do you know whats on your port TCP 443 ?- And the IP reputation with Snort is alot more user friendly than Suricata. Also, the host attribute table feature of Snort can be usefull depending on the size of your network and uniqueness of your users. Again, I find Snort better at detecting packet overlapping, while you could spend days fine tuning Surita Stream engine and end up just disabling the Stream rules ;)

30$/year VRT with Snort is a great deal of protection for the money. ET Pro is more expensive…which they had a home user pricing...

Then again, if you have time up your sleeve, try them both Suricata/Snort and find out for yourself their strengths. But honestly, on pfSense, for protecting a home or SOHO network w/o getting too much into technical stuff; go with the Snort package.

F.

chriva

@jflsakfja:

…
The upside: each time the bad IP generates an alert, the alert timestamp is updated. If it keeps doing it while being inside the ban timeframe, then the IP is perpetually kept on blocked hosts (well, until something like a reboot kicks it (temporarily) out).

All in all, don't worry if you are still seeing alerts from blocked IPs. As long as your rules are set up correctly (easy to test) then you are set :)

Hi to all and thanks for creating this thread !
I'm following your suggestion to set up a suricata ids,
Can someone explain how can I check the list of "blocked hosts" and their timers?

Thanks,
Chris

–-
I figured it out: services -> suricata -> blocks gives me all the info I need.

Ramosel

@fsansfil:

Unless you are writing your own rules or you are an ISP dealing with 30k users, right now you are better with Snort than Suricata. Using ET open and Snort VRT rulesets, youll get more coverage of threat protection with Snort (All VRT rules compatible with the engine)

There was talk (here and elsewhere) many months back that Snort as a long term product was dead… or dying... in the wake of Suricata and its resources. Has there been a change to that thought? New people involved?

Just curious?

Rick

bmeeks

@Ramosel:

There was talk (here and elsewhere) many months back that Snort as a long term product was dead… or dying... in the wake of Suricata and its resources. Has there been a change to that thought? New people involved?

Just curious?

Rick

I think a couple of things have happened. First, thus far the Cisco purchase of Snort has not resulted in the open source project side being squashed. That was a fear early on after the purchase. Second, Snort 3.0-BETA supports multi-threading. So once v3.0 goes from BETA to RELEASE, the argument about Suricata's performance advantages will lose some steam.

I think both systems are fine. Each has its own unique features. Suricata can grab and log a lot more information than Snort can at the moment (all the JSON stuff, TLS cert exchanges, etc.), but Snort sports the new OpenAppID functionality.

Bill

jim82

New user here, bear with me.

Following the first post and reading it over and over, I don't understand the part about floating rules.

Here's what I did(also see screenshots)

Created new interface called DMZ(did this to test on my current system)
Created Floating Rule, as described, but ONLY for the interface DMZ
Created allow rule for everything on the interface tab for DMZ(started out with DNS only, but nothing went through, so I changed it to any)

Testing with ping = failed
Testing NSLOOKUP = failed

When disabling the floating rule, all traffic pass, as expected.

I'm sure it's me messing this up in some way, but I don't see how/why.

Assistance greatly appreciated.
BR Jim

interface_tab.png_thumb

floating.png_thumb

host_timeout.png_thumb

floating_DISABLED.png_thumb

logs.png_thumb

A Former User

Can I see a screenshot of the floating rule in question?

If you are talking about the "block all" floating rule, it should only apply to traffic destined for pfsense's ports (that's why there is a giant red warning under it).

awsiemieniec

Thank you for the post - I've been walking through it and adjusting as necessary. I have a server at a colo accessed via a tunnel so some adjustment is necessary.

About that floating rule, the first one your mention where you write in large red "DON'T CHANGE DESTINATION PORT RANGE!!!". If I follow that example EXACTLY as you write it, rule #1 :P, then I end up blocking all outgoing traffic. Here is the float rule: (attached) Do you really intend to block ALL? I'm corn-fused?! Maybe I missed a step?

Thx.

Floating.png_thumb

A Former User

Most common question I had to answer so far ;D

The rule you show will block all traffic.

The rule you want will block all traffic destined for pfsense's ports!. That's where the "don't change ports" part comes in.

Adjust that rule to destination pfsense's ports and it will be OK ;)

jim82

@jflsakfja:

Can I see a screenshot of the floating rule in question?

If you are talking about the "block all" floating rule, it should only apply to traffic destined for pfsense's ports (that's why there is a giant red warning under it).

Thanks for your reply. I guess the post above concerns the same confusion. Don't get me wrong, but you write the following:

Next up Floating tab:
Set up a rule but make these changes:
Action Block
Quick TICKED!!!
Interface Hold CTRL and click on all interfaces EXCEPT LAN(admin) and SYNC
Direction any
Source any
Destination any

If you read this directly(as I did, since I'm absolute beginner), your rule will block everything in/out on all interfaces, except "LAN".

I did this, and got confused. I could not wrap my head around, how on earth a Floating block ANY ANY ANY to all interfaces would possibly allow any traffic to pass through.

My suggestion is to clarify(maybe more red big letters) that this floating block rule is ONLY for the ports you specify as being web interface and SSH(which makes good sense).

Thanks for your guide, I'm looking forward to following the next steps.

BR Jim

A Former User

I agree, the text is a bit confusing, but it was meant to say "you created the rule, now head over to the floating tab and set up an identical rule to it, making these changes".

It's getting changed in the next version anyways (since I can't edit old posts) when I finally find the time to finish it. Caught up with work (a LOT of it) these days, and the guide is pushed back on my priorities list.

jim82

Sounds great, thanks for the clarification. Couldn't wrap my head around that weird floating rule :)

Looking forward to Version2 8)

lrosenman

Ok – reading through 28(!) pages is not my idea of fun. Is there a good summary for current (May, 2015) setups from scratch on 2.2.2 of PFSense, and using Suricata and any other helpful stuff for a colo'd LAN offering services to folks on the Internet?

A Former User

Not currently, something is being worked on :)