Snort 2.9.4.1 pkg v.2.5.8
-
Hi Bill.
Can you test a client IP header rule that triggers in this scenario?
The way NAT reflection works is that the packets go to the firewall, and are then sent from the firewall to the NAT target, changed to appear as if they came from the firewall, so that replies go back through the firewall.
There is no security issue there, unless you count the loss of the client IP from the perspective of the NAT target/server. The server sees the firewall itself as the "client", not the actual IP of the client.
Supermule:
I'm not sure I understand what you are asking me to do. If in reference to my previous post, the point of that post to Shinzo was simply to show a mechanism to see the pre-NAT LAN IP addresses in the Snort logs. If you run Snort on just the WAN (and you are using NAT), then Snort logs all "local" traffic as originating from the WAN IP. Here is an example. Suppose host 192.168.0.1 on the LAN communicates with a BOT C-n-C server somewhere on the web. Snort will see and log that traffic, but what you will see in the ALERT log for the Source IP is your WAN IP and then the Destination IP of the BOT C-n-C server. So while you will know that you have an infected host on your LAN, you won't know from the Snort logs which one it is.
However, with my scheme (Snort on LAN), then you will see in the ALERT log the local pre-NAT LAN IP of the infected host (192.168.0.1 in my example) as well as the Destination IP of the C-n-C server. You can then tell which LAN host is the problem.
Bill
-
I understand :)
-
P.S. – while you can do it, it would be sort of pointless and a waste of resources to run a duplicate set of rules on both the WAN and LAN interfaces in a scenario like I described. I prefer the setup where known bad hosts are blocked right away at the perimeter (the WAN). I get this from those ET-CIARMY and RBN rules. Then my LAN side interface is the next layer of protection and contains the richer rule set.
Bill
Whats your recommendation then for this setup in regards to duplicate settings on the Preprocessors tab? Are they unnecessarily redundant?
-
Thanks for the new snort interface.
Ok , my problem is…..
Base on my understanding with snort blocking capabilities
LAN setting - I can block both LAN@internal and External IP with selection of new whitelist setting which i have uncheck the "Add firewall Local Networks to the list (excluding WAN)." option. Block setting "both"
WAN setting - I can block external setting , wth default whitelist setting. Block setting "both"Both LAN and WAN setting will have different categories selection base on my need and requirement.
For the above setting , I manage to
WAN Setting - block external IP
LAN Setting - block only external IP (supposely to block both internal@LAN ip and external IP)When I view the whitelist IP with "view list" function , in the new whitelist setting, the local IP still there , sort nothing happened with the uncheck action I have done?
Bug? or I missed something?
4 am here, since yesterday morning, I will need to have some rest first. Really appreciate if any one can advice me the next action.
-
2.1-BETA1 (amd64)
built on Fri May 17 16:45:31 EDT 2013
FreeBSD 8.3-RELEASE-p8
Snort v2.9.4.1 pkg v. 2.5.8Hello,
Ever since I enabled Snort on my LAN interface, my firewall logs have been flooded with the alerts on the attached screen shot. Is there any way to get rid of these alerts? I've been running snort just on the WAN interface for a few good months now and never had these alerts on the firewall logs before. I started to test Snort on the LAN interface a couple of days ago and noticed my firewall log was full of those alerts. I do have IPv6 disabled on the WAN and LAN interfaces. BTW, thanks for this package bmeeks you’ve done a great job with it.
-
Whats your recommendation then for this setup in regards to duplicate settings on the Preprocessors tab? Are they unnecessarily redundant?
The Preprocessors are unique to each instance of Snort. When you run Snort on more than one interface, you actually have two completely separate and independent processes running – totally separate running copies of the binary. You may already have known that, so forgive me for repeating it if you did. Not trying to insult anyone's skills, but just making sure we start from the same baseline for my explanation below.
So you could certainly- depending on how you have the rule sets selected- use different Preprocessor settings on the different interfaces. It's just my view that on today's reasonably fast hardware the preprocessor overhead is not that big of a deal. I don't think it's nearly as taxing on CPU and RAM as say having duplicate sets of ALL the rules running on the LAN and WAN interfaces. The rules are where all the CPU and RAM resources get used (of course the HTTP_INSPECT, Frag3 and Stream5 can use some extra RAM depending on how they are configured). But in high traffic situations, different preprocessors settings would certainly be the way to go. You just have to be careful that you don't disable a preprocessor that an enabled rule depends on. Doing so will cause those "FATAL ERROR" messages on startup.
So to sum up my answer to your question, if had 2GB or more of RAM and not a lot of load (say less than 50 Mbits/sec) with a reasonably modern CPU; then I would just run the default preprocessor setup on both interfaces.
Bill
-
Thanks for the new snort interface.
Ok , my problem is…..
Base on my understanding with snort blocking capabilities
LAN setting - I can block both LAN@internal and External IP with selection of new whitelist setting which i have uncheck the "Add firewall Local Networks to the list (excluding WAN)." option. Block setting "both"
WAN setting - I can block external setting , wth default whitelist setting. Block setting "both"Both LAN and WAN setting will have different categories selection base on my need and requirement.
For the above setting , I manage to
WAN Setting - block external IP
LAN Setting - block only external IP (supposely to block both internal@LAN ip and external IP)When I view the whitelist IP with "view list" function , in the new whitelist setting, the local IP still there , sort nothing happened with the uncheck action I have done?
Bug? or I missed something?
4 am here, since yesterday morning, I will need to have some rest first. Really appreciate if any one can advice me the next action.
I went back and looked at the code. Currently it is by design that the interface Snort is running on (except when it's the WAN) is always added to HOME_NET and the WHITELIST. Do you really want to block your local LAN IP addresses? If so, why? You can easily lock yourself out of the firewall this way (at least lock out a workstation). It has always been the behavior of Snort to add the interface running Snort to HOME_NET and the WHITELIST. What changed in this last update was instead of whitelisting the entire WAN subnet, only the WAN IP is whitelisted. The checkbox just adds all the other locally attached firewall networks to HOME_NET and the WHITELIST. Formerly these were not added by default.
Bill
-
2.1-BETA1 (amd64)
built on Fri May 17 16:45:31 EDT 2013
FreeBSD 8.3-RELEASE-p8
Snort v2.9.4.1 pkg v. 2.5.8Hello,
Ever since I enabled Snort on my LAN interface, my firewall logs have been flooded with the alerts on the attached screen shot. Is there any way to get rid of these alerts? I've been running snort just on the WAN interface for a few good months now and never had these alerts on the firewall logs before. I started to test Snort on the LAN interface a couple of days ago and noticed my firewall log was full of those alerts. I do have IPv6 disabled on the WAN and LAN interfaces. BTW, thanks for this package bmeeks you’ve done a great job with it.
Traffic on port 1900 to that destination IPv6 address is SSDP (Simple Service Discovery Protocol) traffic. Snort does not, to the best of my knowledge, generate such traffic and has no need for it. I would say this is not related to enabling Snort on the LAN.
Bill
-
thanks Bill for your answer…but its weird, if I disable Snort on the LAN interface, those log entries stop and only return if I enable Snort on the LAN interface again.
-
So to sum up my answer to your question, if had 2GB or more of RAM and not a lot of load (say less than 50 Mbits/sec) with a reasonably modern CPU; then I would just run the default preprocessor setup on both interfaces.
Thank you Bill for your clear and concise reply. Snort has jumped in leaps and bounds since you started contributing and adding to Ermal's package. ;D
-
I have no choice with referring to torrent or p2p blocking 2with snort. With current version, without blocking the client who is downloading using torrent software. Not much can be done as the client kept continue downloading the file without failure. I have try it with my notebook , even though snort manage to block some external IP (using both LAN and WAN setting) , my torrent software keep continue downloading as nothing blocking it at all.
I do understand that with the setting I intended to used , my client or some workstation will be accidentally block(if they forgot to shut off their torrent software. But it sort like reminder to the client not to use torrent software in my network. After all with cron , it will only block for 5 min. So the client must make sure their pc is according to my requirement within 5 minute.
Really appreciate if the whitelist issue can be settle.
Thanks for the new snort interface.
Ok , my problem is…..
Base on my understanding with snort blocking capabilities
LAN setting - I can block both LAN@internal and External IP with selection of new whitelist setting which i have uncheck the "Add firewall Local Networks to the list (excluding WAN)." option. Block setting "both"
WAN setting - I can block external setting , wth default whitelist setting. Block setting "both"Both LAN and WAN setting will have different categories selection base on my need and requirement.
For the above setting , I manage to
WAN Setting - block external IP
LAN Setting - block only external IP (supposely to block both internal@LAN ip and external IP)When I view the whitelist IP with "view list" function , in the new whitelist setting, the local IP still there , sort nothing happened with the uncheck action I have done?
Bug? or I missed something?
4 am here, since yesterday morning, I will need to have some rest first. Really appreciate if any one can advice me the next action.
I went back and looked at the code. Currently it is by design that the interface Snort is running on (except when it's the WAN) is always added to HOME_NET and the WHITELIST. Do you really want to block your local LAN IP addresses? If so, why? You can easily lock yourself out of the firewall this way (at least lock out a workstation). It has always been the behavior of Snort to add the interface running Snort to HOME_NET and the WHITELIST. What changed in this last update was instead of whitelisting the entire WAN subnet, only the WAN IP is whitelisted. The checkbox just adds all the other locally attached firewall networks to HOME_NET and the WHITELIST. Formerly these were not added by default.
Bill
-
I have no choice with referring to torrent or p2p blocking 2with snort. With current version, without blocking the client who is downloading using torrent software. Not much can be done as the client kept continue downloading the file without failure. I have try it with my notebook , even though snort manage to block some external IP (using both LAN and WAN setting) , my torrent software keep continue downloading as nothing blocking it at all.
I do understand that with the setting I intended to used , my client or some workstation will be accidentally block(if they forgot to shut off their torrent software. But it sort like reminder to the client not to use torrent software in my network. After all with cron , it will only block for 5 min. So the client must make sure their pc is according to my requirement within 5 minute.
Really appreciate if the whitelist issue can be settle.
I can make the change, but I have to figure out a way to incorporate it so it does not break existing settings for folks who do not want to use it in the same manner. I will put this on the To-Do list for the next release. I will alter the WHITELIST so the Local Networks (including the interface Snort is running on) are not automatically included when the new Checkbox is cleared.
In the meantime, I can show you how to modify the default behavior on your system if you will PM (Private Message) me.
Bill
-
thanks Bill for your answer…but its weird, if I disable Snort on the LAN interface, those log entries stop and only return if I enable Snort on the LAN interface again.
OK, I will do a little research and see if something pops up. I don't use IPv6 on my systems yet, so I have never seen this traffic.
Bill
-
I have no choice with referring to torrent or p2p blocking 2with snort. With current version, without blocking the client who is downloading using torrent software. Not much can be done as the client kept continue downloading the file without failure. I have try it with my notebook , even though snort manage to block some external IP (using both LAN and WAN setting) , my torrent software keep continue downloading as nothing blocking it at all.
I do understand that with the setting I intended to used , my client or some workstation will be accidentally block(if they forgot to shut off their torrent software. But it sort like reminder to the client not to use torrent software in my network. After all with cron , it will only block for 5 min. So the client must make sure their pc is according to my requirement within 5 minute.
Really appreciate if the whitelist issue can be settle.
I can make the change, but I have to figure out a way to incorporate it so it does not break existing settings for folks who do not want to use it in the same manner. I will put this on the To-Do list for the next release. I will alter the WHITELIST so the Local Networks (including the interface Snort is running on) are not automatically included when the new Checkbox is cleared.
In the meantime, I can show you how to modify the default behavior on your system if you will PM (Private Message) me.
Bill
Thanks will PM you now
-
thanks Bill for your answer…but its weird, if I disable Snort on the LAN interface, those log entries stop and only return if I enable Snort on the LAN interface again.
OK, I will do a little research and see if something pops up. I don't use IPv6 on my systems yet, so I have never seen this traffic.
Bill
No worries Bill, I got the firewall to stop logging IPv6 multicast, no need to look into this, thanks
-
No worries Bill, I got the firewall to stop logging IPv6 multicast, no need to look into this, thanks
OK. I never did find any connection in my Google research between SSDP and Snort. There is one possibility, though, that just occurred to me. Snort puts the interface it is running on in promiscuous mode. Perhaps that is why you suddenly started seeing the traffic when Snort was enabled on the interface.
Just not logging it should be fine.
Bill
-
Just posting to confirm that the blank page problem has been resolved.
Thank you ;D
-
The new gui is very nice. The viewing of the whitelist defaults was very helpful in making sure it was getting the list that I wanted to make sure something important didn't get blocked.
Nice job!
-
I am getting a bunch of unimportant alerts on the LAN from an ipv6 address (it is actually a dhcp solicit request to port 547). This is the first time putting an ipv6 address in the suppression list so I don't know if this is new behavior or not. I have already upgraded all my installations to 2.5.8 package but I have a feeling older packages have the same issue though.
06/14/13 15:15:37 2 Attempted Information Leak fe80::344b:2f8a:dc36:80e5 ff02::c 122:23 (portscan) UDP Filtered Portsweep
I tried adding the ipv6 address to suppress fe80::344b:2f8a:dc36:80e5 ff02::c but it gets written to the interface suppression file incorrectly which causes snort to not start up…
#(portscan) UDP Filtered Portsweep
suppress gen_id 122, sig_id 23, track by_src, ip fe80::344b:2f8a:dc36:80e5I really don't want to ignore the scan completely. I would just modify the suppression file directly but it will get overwritten on the next time the suppression file gets written to disk.
UPDATE: Even if I manually edit the suppresion file for the interface the gui seems to overwrite it with the corrupt line when I try even starting the snort instance on that interface. It was worth a try anyway.
-
I am getting a bunch of unimportant alerts on the LAN from an ipv6 address (it is actually a dhcp solicit request to port 547). This is the first time putting an ipv6 address in the suppression list so I don't know if this is new behavior or not. I have already upgraded all my installations to 2.5.8 package but I have a feeling older packages have the same issue though.
06/14/13 15:15:37 2 Attempted Information Leak fe80::344b:2f8a:dc36:80e5 ff02::c 122:23 (portscan) UDP Filtered Portsweep
I tried adding the ipv6 address to suppress fe80::344b:2f8a:dc36:80e5 ff02::c but it gets written to the interface suppression file incorrectly which causes snort to not start up…
#(portscan) UDP Filtered Portsweep
suppress gen_id 122, sig_id 23, track by_src, ip fe80::344b:2f8a:dc36:80e5I really don't want to ignore the scan completely. I would just modify the suppression file directly but it will get overwritten on the next time the suppression file gets written to disk.
UPDATE: Even if I manually edit the suppresion file for the interface the gui seems to overwrite it with the corrupt line when I try even starting the snort instance on that interface. It was worth a try anyway.
Thanks for reporting. That definitely looks like a bug. I will get it fixed and included in the new 2.5.9 update I am working on. This update will include the 2.9.4.6 binary upgrade of Snort itself, and the 2.5.9 GUI package update will include some new features. I will figure out this Suppress List bug with IPv6 and include that fix as well.
Update – after looking at what is written to the Suppress List, it dawned on me what the problem is. I added a "fix" to allow wrapping of the long IPv6 addresses in the Alerts tab columns. What I had to do is provide a zero-length space in the IPv6 string to give the browser a line break opportunity for word wrapping. I put one of those codes at each colon in the address. That's what the code is. So I inadvertently introduced this problem. I will get it fixed, but would rather just include it in the 2.5.9 update mentioned above. That will be coming out before July.
Until then you should be able to manually edit the list. First, open the list on the Suppress List tab, click the edit icon, then delete any bogus entries. Save the list. Then manually add the information. It should not get overwritten again unless you click the plus ( + ) icon on the Alerts tab to automatically add another entry.
Bill
