Snort 2.6.0 High Memory Usage on FreeBSD

I've been working with Snort 2.6.0 on FreeBSD.

When you look at the snort.conf you'll see a bunch of rules commented out.

# include $RULE_PATH/web-attacks.rules
# include $RULE_PATH/backdoor.rules
# include $RULE_PATH/shellcode.rules
# include $RULE_PATH/policy.rules
# include $RULE_PATH/porn.rules
# include $RULE_PATH/info.rules
# include $RULE_PATH/icmp-info.rules
# include $RULE_PATH/virus.rules
# include $RULE_PATH/chat.rules
# include $RULE_PATH/multimedia.rules
# include $RULE_PATH/p2p.rules
# include $RULE_PATH/spyware-put.rules

When you start Snort you'll see it uses much more memory compared to earlier versions.

  654 root        1 -58    0   248M   247M bpf    0   0:01  3.30% snort

If this is too much, and you are willing to sacrifice Snort performance, you can enable the following in snort.conf:

config detection: search-method lowmem

This results in less memory usage.

  656 root        1 -58    0 39800K 39128K bpf    0   0:01  0.00% snort

With this option enabled you can even uncomment all rules.

  661 root        1 -58    0 59224K 58580K bpf    0   0:01  1.95% snort

What if you want the best performance and all rules? Well, on FreeBSD you are going to encounter a 512 MB default RAM limitation that will prevent Snort from running. 
 
# limit
cputime      unlimited
filesize     unlimited
datasize     524288 kbytes
stacksize    65536 kbytes
coredumpsize unlimited
memoryuse    unlimited
vmemoryuse   unlimited
descriptors  11095 
memorylocked unlimited
maxproc      5547 
sbsize       unlimited

You can change this by making the following entries in /boot/loader.conf:

kern.dfldsiz="1G"                # Set the initial data size limit
kern.maxdsiz="1G"                # Set the max data size

Reboot when done. Here is the result.

# limit
cputime      unlimited
filesize     unlimited
datasize     1048576 kbytes
stacksize    65536 kbytes
coredumpsize unlimited
memoryuse    unlimited
vmemoryuse   unlimited
descriptors  11095 
memorylocked unlimited
maxproc      5547 
sbsize       unlimited


Now you can run Snort with all rules and best performance, and see it occupies over 900 MB.

  645 root        1 -58    0   925M   926M bpf    0   0:04  0.00% snort

I have not tested Snort to see the effect of various options, although I prefer to run as many rule sets as makes sense for my environment. Note I did not add Bleeding or Community rule sets for this example.

Comments

Anonymous said…
Richard, there are other options available, please see the manual.

http://www.snort.org/docs/snort_htmanuals/htmanual_260/node10.html

Also, you should see that the memory consumption decreases once Snort has started and settled down.

Also, take a look at this post on snort-dev:

http://marc.theaimsgroup.com/?l=snort-devel&m=115627559410302&w=2
12:32 PM
Post a Comment

Popular posts from this blog

Zeek in Action Videos

Image
This is a quick note to point blog readers to my Zeek in Action YouTube video series for the Zeek network security monitoring project .  Each video addresses a topic that I think might be of interest to people trying to understand their network using Zeek and adjacent tools and approaches, like Suricata, Wireshark, and so on.  I am especially pleased with Video 6 on monitoring wireless networks . It took me several weeks to research material for this video. I had to buy new hardware and experiment with a Linux distro that I had not used before -- Parrot .  Please like and subscribe, and let me know if there is a topic you think might make a good video.
Read more

MITRE ATT&CK Tactics Are Not Tactics

Image
Just what are "tactics"? Introduction MITRE ATT&CK  is a great resource, but something about it has bothered me since I first heard about it several years ago. It's a minor point, but I wanted to document it in case it confuses anyone else. The MITRE ATT&CK Design and Philosophy document from March 2020 says the following: At a high-level, ATT&CK is a behavioral model that consists of the following core components: • Tactics, denoting short-term, tactical adversary goals during an attack; • Techniques, describing the means by which adversaries achieve tactical goals; • Sub-techniques, describing more specific means by which adversaries achieve tactical goals at a lower level than techniques; and • Documented adversary usage of techniques, their procedures, and other metadata. My concern is with MITRE's definition of "tactics" as "short-term, tactical adversary goals during an attack," which is oddly recursive. The key word in the tacti
Read more

New Book! The Best of TaoSecurity Blog, Volume 4

Image
  I've completed the TaoSecurity Blog book series . The new book is  The Best of TaoSecurity Blog, Volume 4: Beyond the Blog with Articles, Testimony, and Scholarship .  It's available now for Kindle , and I'm working on the print edition.  I'm running a 50% off promo on Volumes 1-3 on Kindle through midnight 20 April. Take advantage before the prices go back up. I described the new title thus: Go beyond TaoSecurity Blog with this new volume from author Richard Bejtlich. In the first three volumes of the series, Mr. Bejtlich selected and republished the very best entries from 18 years of writing and over 18 million blog views, along with commentaries and additional material.  In this title, Mr. Bejtlich collects material that has not been published elsewhere, including articles that are no longer available or are stored in assorted digital or physical archives. Volume 4 offers early white papers that Mr. Bejtlich wrote as a network defender, either for technical or pol
Read more