aiden
/
suricata
mirror of https://github.com/OISF/suricata
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

docs: refactor DPDK docs and add performance tuning section

Browse Source 
Ticket: #5857
Ticket: #5858
pull/8861/head
Lukas Sismis 2 years ago committed by Victor Julien
parent 03319263db
commit 1c3cb1e8cc
1 changed files with 92 additions and 73 deletions
Show Stats Download Patch File Download Diff File

165
doc/userguide/configuration/suricata-yaml.rst
Unescape Escape View File

@ -1927,16 +1927,16 @@ Data Plane Development Kit (DPDK)
`Data Plane Development Kit <https://www.dpdk.org/>`_ is a framework for fast
packet processing in data plane applications running on a wide variety of CPU
architectures. DPDK `Environment Abstraction Layer (EAL)
architectures. DPDK's `Environment Abstraction Layer (EAL)
<https://doc.dpdk.org/guides/prog_guide/env_abstraction_layer.html>`_
provides a generic interface to low-level resources. It is a unique way how
DPDK libraries access NICs. EAL creates an API for application to access NIC
DPDK libraries access NICs. EAL creates an API for an application to access NIC
resources from the userspace level. In DPDK, packets are not retrieved via
interrupt handling. Instead, the application `polls
<https://doc.dpdk.org/guides/prog_guide/poll_mode_drv.html>`_ NIC for newly
<https://doc.dpdk.org/guides/prog_guide/poll_mode_drv.html>`_ the NIC for newly
received packets.
DPDK allows the user space application to directly access memory where NIC
DPDK allows the user space application to directly access memory where the NIC
stores the packets. As a result, neither DPDK nor the application copies the
packets for the inspection. The application directly processes packets via
passed packet descriptors.
@ -1958,7 +1958,7 @@ Support for DPDK can be enabled in configure step of the build process such as:
Suricata makes use of DPDK for packet acquisition in workers runmode.
The whole DPDK configuration resides in the `dpdk:` node. This node encapsulates
2 main subnodes and those are eal-params and interfaces.
2 main subnodes, and those are eal-params and interfaces.
::
@ -1981,43 +1981,47 @@ The whole DPDK configuration resides in the `dpdk:` node. This node encapsulates
copy-iface: none # or PCIe address of the second interface
The node `dpdk.eal-params` consists of `DPDK arguments
<https://doc.dpdk.org/guides/linux_gsg/linux_eal_parameters.html>`_
that are usually passed through command line. These arguments are used to
initialize and configure EAL. Arguments can be specified in either long or
short forms. When specifying the arguments, the dashes are omitted.
Among other settings, this configuration node is able to configure available
NICs to Suricata, memory settings or other parameters related to EAL.
The node `dpdk.interfaces` wraps a list of interface configurations. Items of
the list follows the structure that can be found in other capture interfaces.
The `DPDK arguments
<https://doc.dpdk.org/guides/linux_gsg/linux_eal_parameters.html>`_, which
are typically provided through the command line, are contained in the node
`dpdk.eal-params`. EAL is configured and initialized using these
parameters. There are two ways to specify arguments: lengthy and short.
Dashes are omitted when describing the arguments. This setup node can be
used to set up the memory configuration, accessible NICs, and other EAL-related
parameters, among other things. The definition of lcore affinity as an EAL
parameter is a standard practice. However, lcore parameters like `-l`, `-c`,
and `--lcores`` are specified within the `suricata-yaml-threading`_ section
to prevent configuration overlap.
The node `dpdk.interfaces` wraps a list of interface configurations. Items on
the list follow the structure that can be found in other capture interfaces.
The individual items contain the usual configuration options
such as `threads`/`copy-mode`/`checksum-checks` settings. Other capture
interfaces, such as AF_PACKET, rely on the user that NICs are appropriately
configured.
Configuration through kernel does not apply to applications running under DPDK.
The application is solely responsible for the initialization of NICs it is
using. So, before the start of Suricata, NICs that Suricata uses, must undergo
the process of initialization.
As a result, there are extra extra configuration options (how NICs can be
interfaces, such as AF_PACKET, rely on the user to ensure that NICs are
appropriately configured.
Configuration through the kernel does not apply to applications running under
DPDK. The application is solely responsible for the initialization of the NICs
it is using. So, before the start of Suricata, the NICs that Suricata uses,
must undergo the process of initialization.
As a result, there are extra configuration options (how NICs can be
configured) in the items (interfaces) of the `dpdk.interfaces` list.
At the start of the configuration process, all NIC offloads are disabled to
prevent any packet modification. According to the configuration, checksum
validation offload can be enabled to drop invalid packets. Other offloads can
not be currently enabled.
Additionally, the list items of `dpdk.interfaces` contains DPDK specific
not currently be enabled.
Additionally, the list items in `dpdk.interfaces` contain DPDK specific
settings such as `mempool-size` or `rx-descriptors`. These settings adjust
individual parameters of EAL. One of the entries of the `dpdk.interfaces` is
individual parameters of EAL. One of the entries in `dpdk.interfaces` is
the `default` interface. When loading interface configuration and some entry is
missing, the corresponding value of the `default` interface is used.
The worker threads must be assigned to a specific cores. The configuration
module `threading` can be used to set threads affinity.
The worker threads must be assigned to specific cores. The configuration
module `threading` must be used to set thread affinity.
Worker threads can be pinned to cores in the array configured in
`threading.cpu-affinity["worker-cpu-set"]`. Performance-oriented setups have
everything (the NIC, memory and CPU cores interacting with the NIC) based on
everything (the NIC, memory, and CPU cores interacting with the NIC) based on
one NUMA node.
It is therefore required to know layout of the server architecture to get the
It is therefore required to know the layout of the server architecture to get the
best results. The CPU core ids and NUMA locations can be determined for example
from the output of `/proc/cpuinfo` where `physical id` described the NUMA
number. The NUMA node to which the NIC is connected to can be determined from
@ -2032,34 +2036,37 @@ the file `/sys/class/net/<KERNEL NAME OF THE NIC>/device/numa_node`.
## cat /sys/class/net/<KERNEL NAME OF THE NIC>/device/numa_node e.g.
cat /sys/class/net/eth1/device/numa_node
If Suricata has enabled at least 2 (or more) workers, the incoming traffic is
load balanced across the worker threads by Receive Side Scaling (RSS).
Suricata operates in workers runmode. Packet distribution relies on Receive
Side Scaling (RSS), which distributes packets across the NIC queues.
Individual Suricata workers then poll packets from the NIC queues.
Internally, DPDK runmode uses a `symmetric hash (0x6d5a)
<https://www.ran-lifshitz.com/2014/08/28/symmetric-rss-receive-side-scaling/>`_
that redirects bi-flows to specific workers.
Before Suricata can be run, it is required to allocate sufficient number of
hugepages. Suricata allocates continuous block of memory.
For efficiency, CPU allocates memory in RAM in chunks. These chunks are usually
in size of 4096 bytes. DPDK and other memory intensive applications makes use
of hugepages. Hugepages start at the size of 2MB but they can be as large as
1GB. Lower count of pages (memory chunks) allows faster lookup of page entries.
The hugepages need to be allocated on the NUMA node where the NIC and CPU
resides. Otherwise, if the hugepages are allocated only on NUMA node 0 and
the NIC is connected to NUMA node 1, then the application will fail to start.
Therefore, it is recommended to first find out to which NUMA node the NIC is
connected to and only then allocate hugepages and set CPU cores affinity to
the given NUMA node. If the Suricata deployment is using multiple NICs on
different NUMA nodes then hugepages must be allocated on all of those NUMA
nodes.
Before Suricata can be run, it is required to allocate a sufficient number of
hugepages. For efficiency, hugepages are continuous chunks of memory (pages)
that are larger (2 MB+) than what is typically used in the operating systems
(4 KB). A lower count of pages allows faster lookup of page entries. The
hugepages need to be allocated on the NUMA node where the NIC and affiniated
CPU cores reside. For example, if the hugepages are allocated only on NUMA
node 0 and the NIC is connected to NUMA node 1, then the application will fail
to start. As a result, it is advised to identify the NUMA node to which the
NIC is attached before allocating hugepages and setting CPU core affinity to
that node. In case Suricata deployment uses multiple NICs, hugepages must be
allocated on each of the NUMA nodes used by the Suricata deployment.
::
## To check number of allocated hugepages:
sudo dpdk-hugepages.py -s
# alternative (older) way
grep Huge /proc/meminfo
## Allocate hugepages on NUMA node 0:
echo 8192 | sudo tee \
## Allocate 2 GB in hugepages on all available NUMA nodes:
# (number of hugepages depend on the default size of hugepages 2 MB / 1 GB)
sudo dpdk-hugepages.py --setup 2G
# alternative (older) way allocates 1024 2 MB hugepages but only on NUMA 0
echo 1024 | sudo tee \
/sys/devices/system/node/node0/hugepages/hugepages-2048kB/nr_hugepages
@ -2067,37 +2074,49 @@ DPDK memory pools hold packets received from NICs. These memory pools are
allocated in hugepages. One memory pool is allocated per interface. The size
of each memory pool can be individual and is set with the `mempool-size`.
Memory (in bytes) for one memory pool is calculated as: `mempool-size` * `mtu`.
Sum of memory pool requirements divided by the size of one hugepage results in
the number of required hugepages. It causes no problem to allocate more memory
than required but it is vital for Suricata to not run out of hugepages.
Mempool cache is local to the individual CPU cores and holds packets that were
recently processed. As the mempool is shared among all cores, cache tries to
minimize the required inter-process synchronization. Recommended size of the
cache is covered in the YAML file.
There has been an ongoing effort to add a DPDK support into Suricata. While the
capture interface is continually evolving, there has been certain areas with
an increased focus. The current version of the DPDK capture interface provides
support for physical NICs and for running on physical machines in workers
runmode. The work has not been tested neither with the virtual interfaces nor
in the virtual environments like VMs, Docker or similar.
Although the capture interface uses DPDK library, there is no need to configure
any lcores. The capture interface uses the standard Suricata threading module.
Additionally, Suricata is intended to run as a primary process only.
The sum of memory pool requirements divided by the size of one hugepage results
in the number of required hugepages. It causes no problem to allocate more
memory than required, but it is vital for Suricata to not run out of hugepages.
The mempool cache is local to the individual CPU cores and holds packets that
were recently processed. As the mempool is shared among all cores, the cache
tries to minimize the required inter-process synchronization. The recommended
size of the cache is covered in the YAML file.
To be able to run DPDK on Intel cards, it is required to change the default
Intel driver to either `vfio-pci` or `igb_uio` driver. The process is
described in `DPDK manual page regarding Linux drivers
<https://doc.dpdk.org/guides/linux_gsg/linux_drivers.html>`_.
DPDK is natively supported by Mellanox and thus their NICs should work
"out of the box".
**Current DPDK support** involves Suricata running on:
* a physical machine with a physical NICs such as:
* mlx5 (ConnectX-4/ConnectX-5/ConnectX-6)
* ixgbe
* i40e
* ice
* a virtual machine with virtual interfaces such as:
* e1000
* VMXNET3
* virtio-net
Other NICs using the same driver as mentioned above should work as well.
The DPDK capture interface has not been tested neither with the virtual
interfaces nor in the virtual environments like VMs, Docker or similar.
The minimal supported DPDK is version 19.11 which should be available in most
repositories of major distributions. Alternatively, it is also possible to use
`meson` and `ninja` to build and install DPDK from scratch.
repositories of major distributions.
Alternatively, it is also possible to use `meson` and `ninja` to build and
install DPDK from source files.
It is required to have correctly configured tool `pkg-config` as it is used to
load libraries and CFLAGS during the Suricata configuration and compilation.
This can be tested by querying DPDK version as:
To be able to run DPDK on Intel cards, it is required to change the default
Intel driver to either `vfio-pci` or `igb_uio` driver. The process is described
in `DPDK manual page regarding Linux drivers
<https://doc.dpdk.org/guides/linux_gsg/linux_drivers.html>`_. DPDK is natively
supported by Mellanox and thus their NICs should work "out of the box".
::
pkg-config --modversion libdpdk
Pf-ring
~~~~~~~

Write Preview
Loading…
Cancel
Save