Bro not producing a notice.log

I have a Bro cluster setup in the AWS cloud, currently just with one node. My problem is that Bro is not producing the notice.log, it should just log successful SSH logins but it doesn’t. I have tried SSH and FTP bruteforcing the worker node and exceeding the limit of failed connections, again no notice.log. I can see the detect-bruteforcing.bro scripts loaded in the loaded_scripts.log. I am pretty new to Bro, so I am not sure what I am doing wrong.

Regards,

I want to say that’s likely because AWS disables promiscuous mode so getting Bro to work requires some additional tricks. Can anyone verify?

I don't know if anything has changed in the last few years, but I know it
used to be the case that you could not put an AWS interface into
promiscuous mode. To get around this, you had to use a tool like
daemonlogger to dump packets from the external interface to a tap, tun, or
bridge interface and monitor that. For larger scale implementations, you
can use openvpn internally to route all the traffic back to your sensor. I
set up a PoC doing similar with Snort a few years back. (
https://github.com/jeffgeiger/CloudSnort) Hopefully that helps, if this is
still the case.

Best,

Jeff Geiger

Jeff Geiger <jeff.geiger <at> gmail.com> writes:

I don't know if anything has changed in the last few years, but I know it

used to be the case that you could not put an AWS interface into promiscuous
mode. To get around this, you had to use a tool like daemonlogger to dump
packets from the external interface to a tap, tun, or bridge interface and
monitor that. For larger scale implementations, you can use openvpn
internally to route all the traffic back to your sensor. I set up a PoC
doing similar with Snort a few years back.
(https://github.com/jeffgeiger/CloudSnort) Hopefully that helps, if this is
still the case.

Best,

Jeff Geiger

On Thu, Apr 7, 2016 at 6:04 PM, Mike Dopheide <dopheide <at> gmail.com>

wrote:I want to say that's likely because AWS disables promiscuous mode so
getting Bro to work requires some additional tricks. Can anyone verify?

I have a Bro cluster setup in the AWS cloud, currently just with one node.

My problem is that Bro is not producing the notice.log, it should just log
successful SSH logins but it doesn't. I have tried SSH and FTP bruteforcing
the worker node and exceeding the limit of failed connections, again no
notice.log. I can see the detect-bruteforcing.bro scripts loaded in the
loaded_scripts.log. I am pretty new to Bro, so I am not sure what I am doing
wrong.

Regards,

_______________________________________________
Bro mailing listbro <at>

bro-ids.orghttp://mailman.ICSI.Berkeley.EDU/mailman/listinfo/bro

_______________________________________________
Bro mailing list
bro <at> bro-ids.org
http://mailman.ICSI.Berkeley.EDU/mailman/listinfo/bro

So lets say I bridge an interface and start to monitor br0 instead of eth0,
the br0 interface will not be in promiscuous mode and will allow to produce
the notice.log? I am asking because you said you have done it few years ago,
therefore wondering whether anything has changed in AWS.

I still can't understand it why is it not producing the notice.log. I have
all the other logs (conn, http, ssl, x509 etc) and they are working fine. I
am not sure how Bro exactly works but if there is more than 30 rejected FTP
requests in the conn.log why doesn't it raise a notice if that's what the
FTP-bruteforce is looking for.

Regards

Are you running Bro on the machine that is running the ssh and ftp server?

Your logs are likely broken, but you haven't looked closely enough at them. Just because a log exists doesn't mean bro is seeing both sides of the connection.

https://www.bro.org/documentation/faq.html#why-isn-t-bro-producing-the-logs-i-expect-a-note-about-checksums

That worked perfectly, thanks !!

However, now I am seeing a lot of possible_split_routing and
data_before_established.
Any ideas?