Hi Spiros
please read my answers below.

> On 3 Dec 2017, at 08:15, Spiros Papageorgiou <[email protected]> wrote:
> 
> HI all,
> 
> I have an nscrub setup on an LTS16. The config is routing/assymetric mode. I 
> have a few problems and questions:
> - When I stop nscrub the nscrub-VM is left at a cripppled state where it 
> can't even ping IPs that are on connected interfaces (ex the gateway). Why is 
> that? How can i avoid this?
> 
Please provide your nscrub configuration file (or cli), ifconfig, and cat 
/proc/net/pf_ring/dev/ethX/info for the interface you are using in nscrub.
(feel free to write to my email address directly if you don’t want to share 
your data on the ml)
> - The white/black/gray dynamic lists are always empty when mitigating even 
> when nscrub drops attack packets. I'm reading with
> attackers?target_id=pc\&action=list\&profile=black\&list=dynamic
> 
Please send me your target configuration, you can dump it with nscrub-export
> - When pinging from the internet a host defined as a target in scrub, I can 
> see many packets are delayed.
> 64 bytes from x.y.z.130: icmp_seq=1 ttl=125 time=2.42 ms
> 64 bytes from x.y.z.130: icmp_seq=2 ttl=125 time=3002 ms
> 64 bytes from x.y.z.130: icmp_seq=3 ttl=125 time=2002 ms
> 64 bytes from x.y.z.130: icmp_seq=4 ttl=125 time=1002 ms
> 64 bytes from x.y.z.130: icmp_seq=5 ttl=125 time=3.09 ms
> This also happens when the target is in bypass enabled mode. Why this happens 
> and how can i avoid this?
> 
I need to see the nscrub configuration as above.
> - UDP packets are dropped even when I have default action "drop disable". Is 
> this a bug? See the below snippet, where I try to disable udp/src/53/drop. It 
> accepts the command but it there is not result.
> root@nscrub:~# nscrub-export all
> 
> target pc profile DEFAULT udp src 53 drop enable
> 
> root@nscrub:~# curl -u admin:admin 
> http://127.0.0.1:8880/profile/udp/src/53/accept?target_id=pc\&profile=default\&action=disable
>  
> <http://127.0.0.1:8880/profile/udp/src/53/accept?target_id=pc\&profile=default\&action=disable>
> { "envelope_ver": "1.0", "hostname": "katharistis", "epoch": 1512284852, 
> "status": 200, "description": "OK", "data": { "function": 
> "\/profile\/udp\/src\/53\/accept", "return": "success" } }root@nscrub:~#
> root@nscrub:~# nscrub-export all
> target pc profile DEFAULT udp src 53 drop enable
> 
http://127.0.0.1:8880/profile/udp/src/53/accept?target_id=pc\&profile=default\&action=disable
 
<http://127.0.0.1:8880/profile/udp/src/53/accept?target_id=pc%5C&profile=default%5C&action=disable>

Please note that you are using “accept" instead of “drop” in the url, I 
recommend you using nscrub-cli which is more clear.
> - What is the suggested config for mitigating DNS attacks? The victim still 
> needs to be able to do DNS requests and get the answers. Keep in mind that 
> nscrub does not see the DNS requests from the victim (assym mode).
> 
There are a few settings to mitigate DNS attacks that apply to requests:
dns request check_method <method>
dns request rate src [PPS]
dns request rate transaction_id [PPS]
dns request threshold [PPS]
dns request type NUM drop [enable|disable]

As of answers, all you can do is to configure UDP rating:

udp rate src [PPS]
udp rate dst [PPS]
> - Is the mitigation capabilities of nscrub efficient when I redirect an 
> attacked IP, through nscrub in realtime or nscrub needs time to profile a 
> "first seen IP" before mitigating attacks?
> 
With the current algorithms, you can redirect the IP on demand.
> - As far as i understand, nscrub tests IPs using some algorithms and 
> classifies the IPs to the white/black/grey list. Is that right?
> 
Yes, some of the algorithms work this way.

Alfredo
> Sp
> _______________________________________________
> Ntop-misc mailing list
> [email protected]
> http://listgateway.unipi.it/mailman/listinfo/ntop-misc

Attachment: signature.asc
Description: Message signed with OpenPGP

_______________________________________________
Ntop-misc mailing list
[email protected]
http://listgateway.unipi.it/mailman/listinfo/ntop-misc

Reply via email to