> Source: [sk153373](https://support.checkpoint.com/results/sk/sk153373)
# sk153373 - Multi-Queue on Check Point Security Gateways
| Property | Value |
|----------|-------|
| Solution ID | sk153373 |
| Date Created | 2019-05-12 |
| Last Modified | 2025-05-04 |
| Technical Level | General |
| Products | Security Gateway, Scalable Platforms |
| Versions | R82.10, R82, R81.20, R81.10 (EOS), R81 (EOS) |
| OS | Gaia |
## Solution
**Important** **:** This article also applies to R80.30 3.10, but **not** to R80.30. **Table of Contents:**
* Introduction to Multi-Queue
* What's New
* Multi-Queue Configuration in Gaia Clish
* Multi-Queue Configuration Examples
* Notes
Introduction to Multi-Queue {#TARGET_INTRO}
-------------------------------------------
The Multi-Queue approach scales network performance as the number of CPUs increases by allowing them to transfer packets through more than one network queue at a time.
Today's high-end servers have more processors and more CPU cores. In single-queue networking, the scale of the protocol stack is restricted, as the network performance does not scale when the number of CPUs increases. Network interfaces cannot transmit or retrieve packets in parallel because each interface has only one TX and RX queue.
Multi-Queue support removes these bottlenecks by allowing paralleled packet processing.
Multi-queue highlights {#TARGET_WHATS_NEW}
------------------------------------------
* Multi-Queue is fully automated:
* Multi-Queue is enabled by default on all supported interfaces.
* The number of queues on each interface is determined automatically, based on the number of available CPUs (SNDs) and the NIC/driver limitations.
* Queues are automatically affined to the SND cores.
* Multi-Queue configuration does not require a reboot in order to be applied.
* Multi-Queue is now managed by the Clish command line.
* Multi-Queue is now managed by the Out-of-the-box experience performance tool.
Multi-Queue drivers and limitations
-----------------------------------
Multi-Queue is supported on these interface drivers:
|------------|------------------|--------------------------------------------------------------------|---------------------------------|
| Driver | Max Speed (Gbps) | Description | Maximum Number of RX Queues |
| igb | 1 | Intel� Network Adapter Driver for PCIe 1 Gigabit Ethernet Network | 2-16 (depends on the interface) |
| ixgbe | 10 | Intel� Network Adapter Driver for PCIe 10 Gigabit Ethernet Network | 16 |
| i40e | 40 | Intel� Network Adapter Driver for PCIe 40 Gigabit Ethernet Network | 64 |
| i40evf | 40 | Intel� i40e driver for Virtual Function Network Devices | 4 |
| mlx5_core | 100G | Mellanox� ConnectX� mlx5 core driver | 60 |
| mlx5_core | 40 | Mellanox� ConnectX� mlx5 core driver | 60 |
| ena | 20 | Elastic Network Adapter in Amazon� EC2 | Configured automatically |
| mana | 200+ | Microsoft Azure Network Adapter Driver | Configured automatically |
| virtio_net | 10 | VirtIO paravirtualized device driver from KVM� | Configured automatically |
| vmxnet3 | 8 | VMXNET Generation 3 driver from VMware� | 8 |
Multi-Queue Configuration in Gaia Clish {#TARGET_CLISH}
-------------------------------------------------------
> The applicable Gaia Clish commands show and configure Multi-Queue on the supported interfaces.
>
> ### Syntax
>
> To show the existing Multi-Queue configuration:
>
> *show interface \ multi-queue \[verbose\]*
>
> To configure Multi-Queue:
>
> *set interface \ multi-queue {off \| auto \| manual core \}
> save config*
>
> ### Parameters
>
> |-------------------------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
> | Parameter | Description |
> | {off \| auto \| manual} | Multi-Queue operational modes: * **Off**: Multi-Queue is disabled * **Auto**: Multi-Queue is automatically configured using the maximum number of SNDs. Interface queues and IRQ affinity are configured accordingly. * **Manual**: Multi-Queue is configured manually, according to the user. The user decides which CPU cores will be used for Multi-Queue. Interface queues and IRQ affinity are configured accordingly. |
> | core | One or more CPU core numbers on which to apply the command. The list delimiter is a comma. * This is relevant for 'manual' mode only. * **Note**: Whitespaces are prohibited. Examples: "core 0,1,4,5", "core 24". |
> | verbose | Verbosity, shows additional information: * IRQ numbers * RX/TX queue counter |
Multi-Queue Configuration Examples {#TARGET_CLISH_EXAMPLES}
-----------------------------------------------------------
Show / Hide this section
> **Note**: Configuring the interface Multi-Queue settings may cause a temporary packet loss due to the reset of the network interface.
>
> **Set automatic Multi-Queue mode on interface eth2:**
>
> ```
> set interface eth2 multi-queue auto
> ```
>
> **Set manual Multi-Queue mode on eth1 and eth2. Use 6 CPU cores 0, 1, 2, 4, 5, 6 (no whitespaces):**
>
> ```
> set interface eth2 multi-queue manual core 0,1,2,4,5,6
> ```
>
> #### Show interface eth2 Multi-Queue configuration:
>
> ```
> show interface eth2 multi-queue
> Total 8 cores. Multiqueue 2 cores
> i/f type state config cores
> --------------------------------------------------------------------------
> eth2 igb Up Auto 4,0
> ```
>
> ```
> Note:�The output does not include network interfaces that are currently in the down state.
> ```
>
> #### Show the current Multi-Queue configuration, including IRQ numbers:
>
> ```
> show interface eth2 multi-queue verbose
> ```
>
> ```
> Total 8 cores. Multiqueue 2 cores: 0,4
> i/f type state config cores
> --------------------------------------------------------------------------
> eth2 igb Up Auto 4(62),0(79)
> ```
>
> ```
> core interfaces queue irq rx packets tx packets
> -------------------------------------------------------------------------------------------
> 0 eth2 eth2-TxRx-1 79 212 80
> 4 eth2 eth2-TxRx-0 62 16232 18901
> ```
Notes {#TARGET_NOTES}
---------------------
* **NIC reset during Multi-Queue configuration** :
When a new Multi-Queue configuration is applied, the NIC is reset and there is a momentary loss of packets.
* **New NIC installed/enabled** :
When a new NIC is installed or enabled, it needs to be configured using the above commands
* **VSX Gateways / VSX Clusters** :
In a VSX environment, the Multi-Queue configuration is scoped to the Virtual System, in which the tool is being executed. As a result, only the interfaces specific to the Virtual System can be configured/showed. In order to manage interfaces from a particular Virtual System, use the '*set virtual-system* ' command to switch between Virtual Systems.
* In VSX, when all VLANs are configured on a bond interface, and the bond interface is shared between Virtual Systems, you must configure Multi-Queue in the context of VS0.
* **Clusters** :
Multi-Queue should be configured in the same way on all cluster members.
* **Out-of-the-Box Tool:**When the out-of-the-box performance tool is on, Multi-Queue cannot be manually configured. Specifically, the number of Multi-Queue queues cannot be changed. Changes to CPU core affinity will cause the Out-of-the-Box tool to stop its work.
* **Upgrade between different Gaia OS kernel versions** :
Multi-Queue settings are NOT preserved when you upgrade from a system based on 2.6.18 kernel to 3.10 kernel.
---
# Agent Instructions
This content is from the Check Point Support Center (https://support.checkpoint.com), the official knowledge base for Check Point cybersecurity products.
## Navigating This Knowledge Base
- **Complete index**: [llms.txt](https://support.checkpoint.com/llms.txt)
- **All SK articles**: [SecureKnowledge Sitemap](https://support.checkpoint.com/sitemaps/secureknowledge-sitemap-index.xml)
- **SK article URL pattern**: `https://support.checkpoint.com/results/sk/{skId}`
- **Markdown responses**: AI bot User-Agents automatically receive `text/markdown` content