- master: The latest code from the git master branch.
- latest: The latest release version (currently 6.0)
- 6.0: The latest 6.0 patch release.
- 5.0: The latest 5.0 patch release.
Specific version tags also exist for versions 4.1.5 and newer.
Examples:
docker pull jasonish/suricata:latest
docker pull jasonish/suricata:6.0.4
The images are rebuilt and pushed to Docker Hub daily to ensure they
are using the most up to date packages of the base OS, and in the case
of the master
tag, to use the latest Suricata code for the git
master branch.
Tags without an architecture like amd64
or arm64v8
are multi-architecture
image manifests. For the most part Docker will do the right thing, however if
you need to pull the image for a specific architecture you can do so by
selecting a tag with an architecture in the name, for example:
docker pull jasonish/suricata:latest-amd64
docker pull jasonish/suricata:6.0.4-arm64v8
In addition to Docker Hub, these containers are also pushed to quay.io and can be pulled like:
docker pull quay.io/jasonish/suricata:latest
You will most likely want to run Suricata on a network interface on your host machine rather than the network interfaces normally provided inside a container:
docker run --rm -it --net=host \
--cap-add=net_admin --cap-add=sys_nice \
jasonish/suricata:latest -i <interface>
But you will probably want to see what Suricata logs, so you may want to start it like:
docker run --rm -it --net=host \
--cap-add=net_admin --cap-add=sys_nice \
-v $(pwd)/logs:/var/log/suricata \
jasonish/suricata:latest -i <interface>
which will map the logs directory (in your current directory) to the Suricata log directory in the container so you can view the Suricata logs from outside the container.
This container will attempt to run Suricata as a non-root user
provided the containers has the capabilities to do so. In order to
monitor a network interface, and drop root privileges the container
must have the sys_nice
and net_admin
capabilities. If the
container detects that it does not have these capabilities, Suricata
will be run as root.
Docker example:
docker run --rm -it --net=host \
--cap-add=net_admin --cap-add=sys_nice \
jasonish/suricata:latest -i eth0
Podman example:
sudo podman run --rm -it --net=host \
--cap-add=net_admin --cap-add=sys_nice \
jasonish/suricata:latest -i eth0
The directory /var/log/suricata
is exposed as a volume. Another
container can attach it by using the --volumes-from
Docker option.
For example:
-
Start the Suricata container with a name:
docker run -it --net=host --name=suricata jasonish/suricata -i enp3s0
-
Start a second container with
--volumes-from
:docker run -it --net=host --volumes-from=suricata logstash /bin/bash
This will expose /var/log/suricata
from the Suricata container as
/var/log/suricata
in the Logstash container.
The Suricata container exposes the following volumes:
/var/log/suricata
- The Suricata log directory./var/lib/suricata
- Rules, Suricata-Update cache and other runtime data that may be useful to retain between runs./etc/suricata
- The configuration directory.
Note: If
/etc/suricata
is a volume, it will be populated with a default configuration from the container.
If doing bind mounts you may want to have the Suricata user within the container match the UID and GID of a user on the host system. This can be done by setting the PUID and PGID environment variables. For example:
docker run -e PUID=$(id -u) -e PGID=$(id -g)
which will result in the bind mounts being owned by the user starting the Docker container.
The easiest way to provide Suricata a custom configuration is to use a
host bind mount for the configuration directory, /etc/suricata
. It
will be populated on the first run of the container. For example:
mkdir ./etc
docker run --rm -it -v $(pwd)/etc:/etc/suricata jasonish/suricata:latest -V
When the container exits, ./etc
will be populated with the default
configuration files normally found in /etc/suricata
.
Note: The files created in this directory will likely not be owned by the same uid as your host user, so you may need to use sudo to edit this files, or change their permissions.
Hopefully this can be fixed.
In this directory the Suricata configuration can be modified, and Suricata-Update files may be placed. It just needs to be provided as a volume in subsequent runs of Suricata. For example:
docker run --rm -it --net=host \
-v $(pwd)/etc:/etc/suricata \
--cap-add=net_admin --cap-add=sys_nice \
jasonish/suricata:latest -i eth0
The SURICATA_OPTIONS
environment variable can be used to pass command line
options to Suricata. For example:
docker run --net=host -e SURICATA_OPTIONS="-i eno1 -vvv" jasonish/suricata:latest
The easiest way to run Suricata-Update is to run it while the container is running. For example:
In one terminal, start Suricata:
docker run --name=suricata --rm -it --net=host \
--cap-add=net_admin --cap-add=sys_nice \
jasonish/suricata:latest -i eth0
Then in another terminal:
docker exec -it --user suricata suricata suricata-update -f
The will execute suricata-update
in the same container that is
running Suricata (note --name=suricata
), then signal Suricata to
reload its rules with suricatasc -c reload-rules
.
This image is useable on the Raspberry Pi OS, however due to an incompatibility between Raspberry Pi OS and Docker, the timestamps in the logs will be wrong. There are 2 possible fixes to this issue:
- Use the
--privileged
option to Docker - Upgrade the libseccomp2 package on Raspberry Pi OS to a newer version from the backports repo.
wrapper.py
is a script that wraps Suricata-in-Docker for use on the command
line as if it was installed locally.
wrapper.py
has its own arguments that can be seen by running wrapper.py -h
.
Arguments that occur after --
are Suricata arguments and are passed directly
to Suricata (most of the time). The Suricata arguments are preprocessed to setup
any required volumes to provide the appearance that Suricata is running
directly.
./wrapper.py -- -r /path/to/filename.pcap
./wrapper.py --tag 5.0.4 -- -i eno1 -l .
Note that this tool is a work in process and may change, including a change of name.
The build scripts, Dockerfiles and any other files in this repo are MIT licensed.