Quick Snort Setup Instructions for New Users
-
I have the community rules box checked, but when I go to the tab to update, the button is greyed out. Am I doing something wrong? If I select the emerging threat box, I am able to click the update button, but with just community rules checked, I am not.
-
Thank you beemks ! Yes, I think it worked for me. Does it means what Snort consider Autogenerated Supress list as "top list" ?
-
Thank you beemks ! Yes, I think it worked for me. Does it means what Snort consider Autogenerated Supress list as "top list" ?
The auto-generated list will have the name of the interface followed by a random number (UUID).
Bill
-
I don't want to discount Bill's efforts on this thread. It is absolutely the best place to start.
That said, I've recently introduced pfSense and the Snort package to a few friends who are long time, big time, professional security hawks looking for a solution at home a bit more elegant than running generations old (but affordable) dedicated firewall hardware. I believe the best "find" I have come across and directed my friends to is the "fine tuning" post started by user "jflsakfja" as this thread:
https://forum.pfsense.org/index.php/topic,64674.0.html
This user requested the ability some time ago at the start of that thread to be able to edit his post in this sticky rather than having to continuously add to an existing thread. I, for one, would like this be reconsidered by the mods as the above thread is slowly being buried as time passes. I can only surmise the lack of updates as anticipated by "jflsakfja" could be because of a lack of response (evidenced by lack of edits here as of this date) to that request. Or perhaps because I've pushed him/her for more information…? If not at least maybe this post can serve as a jump point for folks looking for or could benefit from that information.
I'd like to see his/her updates continue as the schema introduced by this user may not be the absolute best way of setting up Snort and pfBlocker but its the best I've come across and certainly has made my system more efficient and less troublesome. Judging by recent posts in the Packages area, it seems many others could benefit from this schema as well.... if they knew about it.
Respectfully submitted,
Rick -
Is there any way to log the packets that trigger a snort alert? Mainly I want to see HTTP header and request associated with the alert.
-
Is there any way to log the packets that trigger a snort alert? Mainly I want to see HTTP header and request associated with the alert.
The new 2.9.6.0 version of Snort offers increased packet/file capture abilities according to the post on the Snort.org web site. I am working now on readying that version for the next Snort package update. I will investigate what is offered in the new binary and see what I can reasonably incorporate into the GUI.
I had a similar request from some folks in the new Suricata BETA package thread where the wish was a clickable link from the alert on the ALERTS tab to a view of the packets that triggered the alert. I am mulling over in my head the best way to accomplish that without bogging down the firewall CPU searching through hundreds of megabytes of packet capture files. In general its better to offload such tasks to an external system.
Bill
-
@jflsakfja:
The Missing Part to Quick Snort Setup Instructions for New Users
In tab "Rules", under "Category" select:
(–- means blank table at time of writing)Hello,
does this mean I have to enable ALL when it's written all? Jesus, that would take ages as you can only enable one at a time … Especially GPLv2 has like 85% disabled by default ... Holy, I have 3x WAN, I'd never find the time for that.
I don't really understand enabled/disabled here. For your example with "2000419" (which occurs right after I started snort with default settings) the rule is faded/greyed out - means it is disabled by default, right? So why does it generate an alert?
Edit: Whoa I couldn't use that whole package for 3x WAN, it uses 42% of 2027 MB for 1 WAN already o.O
-
@jflsakfja:
The Missing Part to Quick Snort Setup Instructions for New Users
In tab "Rules", under "Category" select:
(–- means blank table at time of writing)Hello,
does this mean I have to enable ALL when it's written all? Jesus, that would take ages as you can only enable one at a time … Especially GPLv2 has like 85% disabled by default ... Holy, I have 3x WAN, I'd never find the time for that.
I don't really understand enabled/disabled here. For your example with "2000419" (which occurs right after I started snort with default settings) the rule is faded/greyed out - means it is disabled by default, right? So why does it generate an alert?
Edit: Whoa I couldn't use that whole package for 3x WAN, it uses 42% of 2027 MB for 1 WAN already o.O
There is an "Enable All Rules in the Current Category" button on the RULES tab for a Snort interface. So select the GPLv2 Community Rules in the drop-down, then click the "Enable All Rules in the Current Category" button (it is a plus "+" icon).
The answer to your second question about a "disabled rule" apparently causing an alert is as follows. Many Snort rules either fire, or look for previously fired, internal triggers called flowbits. Google the terms "snort flowbits" to find some web sites with more detailed explanations. In order for some rules to fire an alert, they need to see flowbits set by some other rules. The Auto-Flowbits feature of Snort will walk through all of your enabled rules looking any flowbit dependencies. If it finds a needed flowbit is only set by a currently disabled rule, then it will auto-enable that rule (unless you have explicitly manually disabled it). So the short answer is the alert is coming from a rule that the auto-flowbits process toggled from "default disabled" to "default enabled" behind the scenes. If you really do not want an alert for that rule, simply add it's GID:SID to the Suppression List.
Bill
-
Hi Bill,
You said in a couple of your posts that to work around NAT and knowing the origin of an internal threat you'd also enable snort on the LAN interface.
You suggested leaving snort also looking at the WAN interface but with a lot less rules than the LAN and handling everything else on the LAN (I'm assuming you also mean setting both WAN and LAN snort configurations to look at both source and destination).
I can understand the logic of moving it to the LAN since otherwise in a NAT environment you can't see which client is generating threat alerts, but what is the reason behind leaving some of the rules on the WAN and not moving everything to the LAN?
Also another question: my pfSense setup consists of 3 LANs, if I were to setup snort with the vast majority/all the rules on the LAN, would I need to have one setup for each interface? Which in turn would mean 3x as much RAM usage, and manually setting each of the interfaces manually? Is there one way to apply the same config to all LANs through without the increase in memory?
EDIT
Forgot to say my pfSense is a VM running on ESXi 5.5 Update 1, so is this something I can change by configuring promiscuous mode on the VM network?
Thank you.
-
Hi Bill,
You said in a couple of your posts that to work around NAT and knowing the origin of an internal threat you'd also enable snort on the LAN interface.
You suggested leaving snort also looking at the WAN interface but with a lot less rules than the LAN and handling everything else on the LAN (I'm assuming you also mean setting both WAN and LAN snort configurations to look at both source and destination).
I can understand the logic of moving it to the LAN since otherwise in a NAT environment you can't see which client is generating threat alerts, but what is the reason behind leaving some of the rules on the WAN and not moving everything to the LAN?
Also another question: my pfSense setup consists of 3 LANs, if I were to setup snort with the vast majority/all the rules on the LAN, would I need to have one setup for each interface? Which in turn would mean 3x as much RAM usage, and manually setting each of the interfaces manually? Is there one way to apply the same config to all LANs through without the increase in memory?
EDIT
Forgot to say my pfSense is a VM running on ESXi 5.5 Update 1, so is this something I can change by configuring promiscuous mode on the VM network?
Thank you.
It is true there is no necessary technical advantage running Snort with split rules (some on WAN and most on LAN, for example). It is more of a preference thing. Depends primarily on the network design you have and what you want to protect in what ways.
As for a VMware environment with multiple LANs, I don't know of a way short of running Snort on each LAN interface. I have not, though, played with promiscuous mode in VMware. I don't know if that would be a solution or not.
Bill
-
Well I'm torn between the two setups in my case.
If I set it up on the WAN, I only need to set it up once and it will work for all interfaces. However I lose the ability to see which client generated the alert or, in my case, which LAN it came from which makes it even harder. Also I lose the ability to have different per LAN settings
If I set it up on the LAN I can solve all of the problems listed above, but I will have to configure it separately on each of the LANs and it will be a RAM hungry beast.
I was looking at the different performance methods and based on the description decided to try AC-SPLIT. For a VMware environment, would this be preferable to AC-BNFA when trying to optimise for performance while attempting to keep the memory down?
-
Well I'm torn between the two setups in my case.
If I set it up on the WAN, I only need to set it up once and it will work for all interfaces. However I lose the ability to see which client generated the alert or, in my case, which LAN it came from which makes it even harder. Also I lose the ability to have different per LAN settings
If I set it up on the LAN I can solve all of the problems listed above, but I will have to configure it separately on each of the LANs and it will be a RAM hungry beast.
I was looking at the different performance methods and based on the description decided to try AC-SPLIT. For a VMware environment, would this be preferable to AC-BNFA when trying to optimise for performance while attempting to keep the memory down?
Everything I've ever seen posted from the Snort VRT guys seems to point to AC-BNFA being the best choice 99% of the time. It you are trying to keep decent performance but optimize RAM usage, I would stick with AC-BNFA.
Depending on how many rules you choose, Snort can run in 2GB of RAM. Most folks find it works better with 4GB, though, with the more typical choice of enabled rules.
Bill
-
I read somewhere that AC-BNFA-NQ should provide better performance than AC-BNFA and it also appears to be the default in Snort now, as mentioned here: http://manual.snort.org/node16.html
Currently I have Snort running on a single interface as AC-BNFA-NQ with:
- Snort VRT free Registered User
- Snort Community Ruleset
- ETOpen
with 90% of all rules enabled, on a box with 2GB of RAM and the RAM usage never goes above 35% (it is also running pfBlocker, OpenVPN client + server, Squid + LightSquid, HAVP)…
Are you saying AC-BNFA would be better even?
-
I read somewhere that AC-BNFA-NQ should provide better performance than AC-BNFA and it also appears to be the default in Snort now, as mentioned here: http://manual.snort.org/node16.html
Currently I have Snort running on a single interface as AC-BNFA-NQ with:
- Snort VRT free Registered User
- Snort Community Ruleset
- ETOpen
with 90% of all rules enabled, on a box with 2GB of RAM and the RAM usage never goes above 35% (it is also running pfBlocker, OpenVPN client + server, Squid + LightSquid, HAVP)…
Are you saying AC-BNFA would be better even?
No, it is no better in terms of memory consumption. The NQ means "not queued" if I am recalling correctly, and has some relation to overall performance. The NQ settings are popular now.
Bill
-
Is there any way to log the packets that trigger a snort alert? Mainly I want to see HTTP header and request associated with the alert.
The new 2.9.6.0 version of Snort offers increased packet/file capture abilities according to the post on the Snort.org web site. I am working now on readying that version for the next Snort package update. I will investigate what is offered in the new binary and see what I can reasonably incorporate into the GUI.
I had a similar request from some folks in the new Suricata BETA package thread where the wish was a clickable link from the alert on the ALERTS tab to a view of the packets that triggered the alert. I am mulling over in my head the best way to accomplish that without bogging down the firewall CPU searching through hundreds of megabytes of packet capture files. In general its better to offload such tasks to an external system.
Bill
I really hope you'll be able to incorporate full packet capture offloaded to an external system. That would be ideal. I see your reservations and challenges about doing it all on the fw box. Although I believe any of us wanting to do full packet capture are already using barnyard2 sending to an external system. Thanks for all your hard work - I really appreciate it, as I'm sure others do as well :)
-
I am having an issue trying to install snort on my new box. The PBI seems to be missing. Here is what I am seeing…
Beginning package installation for snort .
Downloading package configuration file... done.
Saving updated package information... done.
Downloading snort and its dependencies...
Checking for package installation...
Downloading https://files.pfsense.org/packages/amd64/8/All/snort-2.9.6.0-amd64.pbi ... could not download from there or http://files.pfsense.org/packages/amd64/8/All//snort-2.9.6.0-amd64.pbi.
of snort-2.9.6.0-amd64 failed!I peeked in https://files.pfsense.org/packages/amd64/8/All/ and didn't see the package, just older installs. Was there a reason 2.9.6.0 was pulled?
-
I am having an issue trying to install snort on my new box.
And you are having issues with the search box on the forum as well?
-
I am having an issue trying to install snort on my new box.
And you are having issues with the search box on the forum as well?
Search was used many times and all I came across was that it is available on the servers via IPv6 and not IPv4. If you knew a quick fix, you could have posted it thus having a thread to show in search when others like myself run in to this. It's my understanding that is how the search feature works…
-
Sigh, how about you stop flooding this thread with completely off-topic junk.
-
Someone found a workaround for it in this thread at the end. https://forum.pfsense.org/index.php?topic=74486.15
