Maybe I’m crazy, but I think OSSEC is like a log daemon +…
It’s cross platform, it includes encryption, it has built in filtering and can 
do active response. Why would it make sense to duplicate log shipping if you 
need it to do the security stuff? I.e. OSSEC ought to be a good log aggregator 
to serve it’s primary security goal IMO.

--
James Pulver
CLASSE Computer Group
Cornell University

From: [email protected] [mailto:[email protected]] On 
Behalf Of Jeremy Rossi
Sent: Wednesday, June 18, 2014 7:50 AM
To: [email protected]
Subject: Re: [ossec-list] logging all, but not commands?

We would Very much welcome it. Some suggestions, but nothing more for the 
branch :).

Agent -> master:

json and use first char of { to pick new code path for processing the messages. 
This will allow master to work with legacy agents and new agents cleanly.

Master->agent:

This is harder but something I am working now as part of the work on actice 
response. Reasons it is harder is that unless we change the method if 
encryption/communication at the same time we have no concept of agent version 
so no idea what formats of messages are acceptable.  I still don't know the 
best method for dealing with this and love ideas.

Logall:

Side note this log all feature comes up all the time and is confusing I think 
and maybe something we should solve better.  But I am worried about turning 
ossec from security to a log daemon as other tools have solved that problem.





On Jun 18, 2014, at 12:30 AM, "Janelle" 
<[email protected]<mailto:[email protected]>> wrote:
I guess I need to start a new brach and work on a way to do this :-)

On Tuesday, June 17, 2014 2:57:34 PM UTC-7, Michael Starks wrote:
On 2014-06-17 16:31, Janelle wrote:
> Trying to send "archives" to a syslog server for archival, and it
> can't handle all the extraneous code.

Ah, yes. I have done this as well and had this problem with keepalives
and such. Another issue is that the ossec log format isn't syslog. It
looks like syslog but it's not. Then there's this:

Agentless:
yyyy mmm dd hh:mm:ss (script_name) username->agentless ossec: agentless:
Change detected

Command output:
yyyy mmm dd hh:mm:ss agentname->command

Syscheck:
yyyy mmm dd hh:mm:ss (agent_name) agent_ip->syscheck

Windows:
yyyy mmm dd hh:mm:ss (agent_name) agent_ip->log_name log_name_again

Agent IP could be the IP or it could be "any."

What I wish was that:

1. This was syslog formatted (or something else nice like JSON)
2. In the syslog header, the hostname could be either the agent name or
the actual hostname
3. Newlines were removed or otherwise handled gracefully

And/or that raw logs could be sent over to a syslog server like alerts
are now, in addition to being analyzed. You could even strip off the
ossec header at that point and the syslog server wouldn't know the
difference.
--

---
You received this message because you are subscribed to the Google Groups 
"ossec-list" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to 
[email protected]<mailto:[email protected]>.
For more options, visit https://groups.google.com/d/optout.
--

---
You received this message because you are subscribed to the Google Groups 
"ossec-list" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to 
[email protected]<mailto:[email protected]>.
For more options, visit https://groups.google.com/d/optout.

-- 

--- 
You received this message because you are subscribed to the Google Groups 
"ossec-list" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to