• Blog


How to setup BGP for Exchange Connectivity

This is an example configuration of MetaMux on an Arista 7130 device being used with a typical exchange connection. BGP / PIM peering is done via internal routing daemons. ContentsOverview:Device Configuration:Enabling the MetaMux Application:Ingress traffic:Egress traffic:Routing (BGP/PIM) Configuration:Complete Sample ConfigurationServer side configurationAdjusting the MAC addressIP and Routing Configuration Overview: This example shows 2 servers connected to one exchange gateway but may easily be modified for additional servers Market data and order responses from the Exchange feed will be delivered to the server on NIC 1 The Layer 3 processes (such as BGP / PIM peering) is done via internal routing...
Continue reading →

How to install an application

Applications may be installed on your Arista 7130 device. Application installs will persist across reboots. Applications are installed in /opt/apps when using the install process. Presently, MOS ships with the latest stable version of each of the relevant applications, however if there’s a need to install a specific version of an application, instructions are below. Installing an application from the CLI Download the desired software application image as an RPM from the Software Download page. The file will be named something like “application-version.x86_64.rpm” i.e. metamux-0.14.3-1967.x86_64.rpm Copy the file to your Arista 7130 device. This can be done in a number...
Continue reading →

Automating MOS upgrade

Ways to mass-deploy MOS updates. ContentsBackgroundUsageCaveatsRunning the script on a 7130 deviceFeedback Background While upgrading MOS is generally fairly straightforward via the web UI or using the CLI, it can be time-consuming for mass-deployment. Below is a automation script that network administrators can use to bulk-upgrade Arista 7130 devices. The script is written in Python using the pexpect module. It requires the ssh and scp executable on your host. You can download the script from the Arista releases page. Usage python mos-upgrade-ssh.py <options>. Where options are: -u admin, --user admin -p PASSWORD, --password PASSWORD you will be prompted if none...
Continue reading →

How to remove MOS image with older releases

Step-by-step guide Check what you have installed. For example: C16-B1-10381-3 login: admin C16-B1-10381-3>enable C16-B1-10381-3#show boot system mos_0.7.0 mos_0.7.1 mos_0.7.2 mos_0.8.0 [BOOT] Remove the one(s) you no longer need. Note you cannot remove currently active volumes (marked as [BOOT]). Also note that you can nuke multiple volumes at once with wildcards – for example: C16-B1-10381-3#bash sudo lvremove /dev/vg1/mos_0.7* Do you really want to remove active logical volume mos_0.7.2? [y/n]: y Logical volume "mos_0.7.2" successfully removed Do you really want to remove active logical volume mos_0.7.1? [y/n]: y Logical volume "mos_0.7.1" successfully removed Do you really want to remove active logical volume...
Continue reading →

What is MOS core?

MOS core is a regular MOS which includes all the features without the weight of any bundled FPGA Apps. Ideal for customers using 7130 Connect devices (7130-16, 7130-48, 7130-96), as well as custom application developers. MOS core installations are ideal for devices that are getting “insufficient free space” error while doing an upgrade. testing:~$ sudo mos-install mos-0.16.0.iso Volume group "vg1" has insufficient free space (160 extents): 164 required. Volume group "vg1" has insufficient free space (11 extents): 125 required. % Command 'sudo /usr/sbin/mos-install mos-0.14.1.iso' returned non-zero exit status 5 test(config)#boot system file:mos-0.14.1.iso Volume group "vg1" has insufficient free space (99...
Continue reading →

CVP to K8s: full-scale production network simulation

Intro Network simulation environments have always been limited to a single compute node, which made the labbing of a full-scale production network an exercise in compromise and trade-offs. At the same time compute resources are cheap and abundant and modern application designs are making use of them by adopting meshed scale-out architectures, treating multiple hosts as a single pool of resources. In this post, we’ll see how (with just a few clicks*) we can build a replica of a real production network, orchestrated by Kubernetes based on information extracted from Arista’s CloudVision Portal (CVP). * Assuming all the prerequisites are met

What to do when you are getting a “meltdown”, “powerdown” or “system powered down” message

This is step-by-step guide on what actions to take if you are getting messages similar to below. Jan 17 02:08:34 C16-B1-10381-3 daemon.emerg /usr/bin/meltdownd: Temperature 35000 for /sys/class/hwmon/hwmon3/temp1_input is over critical value of 70000. or 17 02:08:34 C16-B1-10381-3 daemon.emerg /usr/bin/meltdownd: over temperature, system powerdown or Jan 17 02:08:50 C16-B1-10381-3 daemon.emerg /usr/bin/meltdownd: system powered down Step-by-step guide Upgrade to MOS-0.16.0 or later version. This can be downloaded from our releases page. Please refer to our FAQ page on how to update your Arista 7130 device. For details about the fix refer to MOSH-1223 within the MOS-0.16.0 release notes. Power cycle the device....
Continue reading →

CVP TIP – Encrypt local username passwords with Configlet Builder

Introduction   Arista CloudVision Portal (CVP) uses configlets to create configuration snippets for individual or groups of switches based on user selection. These configlets can be either static or dynamic. Static configlets include static EOS CLI configuration statements as if they were right on the switch configuration file. These configlets are used to create the full configuration for the network switches. An example of a static configlet in Arista CloudVision Portal: Below example is a static Arista EOS CLI configuration. These small pieces of configuration snippets can be applied as configlets in Network Provisioning view to selected device or devices....
Continue reading →

Troubleshooting a suspected defective fan

Problem: The “show environment cooling” command shows a fan as “NOT WORKING”. Example: M32-A2-06308-5>show env cooling System Cooling Status: OK Ambient temperature: 41C Airflow: BACK-TO-FRONT (REV) Fan Status Speed ---- ------------------------ -------- 1 OK 51% 2 OK 51% 3 NOT WORKING 0% 4 OK 50% Solution: A 7130 device uses two field replaceable fan modules with two fans each providing a total of four fans. The fans are numbered 1-4 from left to right as you look at the rear of the device. The fans should fit snuggly so that they snap into place. Fans are removed by pinching the...
Continue reading →

How to fix high CPU usage caused by the logrotate process

There have been reports of devices experiencing high cpu usage due to the logrotate process consuming CPU resources on Arista 7130 devices. This can be confirmed by dropping down to bash and running the “top” command as shown on the following images. The primary cause of the problem was a logrotate bug introduced in MOS-0.16.0 (fixed in MOS-0.16.7). An incorrect regex pattern lead to the compressed quagga log files being rotated infinitely along with the normal log files leading to chains like this: /mnt/flash/log/quagga/quagga.log-20180513.gz-20180520.gz-20180527.gz-20180603.gz-20180610.gz-20180617.gz-20180624.gz-20180701.gz-20180708.gz-20180715.gz-20180722.gz This leads to the device eventually either running out of disk space or inodes or both....
Continue reading →

Blinking green/amber lights on ports

Below is a short description of what the blinking lights on the SFP ports of an Arista 7130 mean. Green – an SFP is plugged in and is not reporting loss of signal from the SFP. Red – possible link detection and flickers with traffic flow. Amber – link has been detected but there is no traffic.  

At what point-of-time does MetaWatch timestamp?

The term packet capture is used widely to refer to capturing network traffic. It is however, worth understanding precisely what the term packet refers to. The IEEE sets the standards for Ethernet and define an Ethernet packet at the physical layer as a unit of data wrapping a MAC frame at the data link layer (IEEE Std 802.3-2015 Chapter 3.1.1). An Ethernet packet is therefore an Ethernet frame, prefixed by preamble and start frame delimiter (SFD) bytes and optionally suffixed by an extension field (only used for half-duplex GbE networks). Moving up to the network layer, the IETF set the...
Continue reading →

Why do interface statistics differ between the CLI and SNMP

As of MOS release 0.8.0 support for all interfaces has been added to SNMP, reporting interface status (up or down), carrier status (RX link state), interface alias (CLI interface description, as of 0.8.1), and statistics. However, SNMP statistics reported are those provided by the standard Net-SNMP implementation and differ from those reported via the CLI in two important ways: SNMP statistics reported by the interface and if-mib MIBs are a subset of those reported by the CLI. SNMP unicast frames reported by the interface MIB are actually total frame counts. These differences will be resolved in later releases as we...
Continue reading →

What happened to the CLI “phy” commands?

The ‘phy’ command has been deprecated and split into two commands to separately change the ports speed and signalling mode. Speed is set and reported as an integer in Mb/sec. While the ‘phy’ commands will continue to work their use should be replaced with the newer commands, particularly in saved configs. For example setting a port to 1G is now done via the following: hostname>enable hostname#configure hostname(config)#interface et1 hostname(config-if-et1)#speed 1000 To show the speed of a group of ports the ‘show speed’ command should now be used: hostname>enable hostname#configure hostname(config)#interface et5-8 hostname(config-if-et5-8)#show speed Port Speed ---- ------ et5 1000 et6...
Continue reading →

How to install Xilinx or Altera development tools on an Arista 7130

In order to perform development work on an Arista 7130 device with an FPGA, it’s necessary to install the tools from Xilinx on that device. This allows remote access to the FPGA via the network. Step-by-step guide for Xilinx tools Download the Vivado WebInstall for Linux 64 package from the Xilinx web site – http://www.xilinx.com/support/download.html Run the installer to install the tools on an x86_64 linux machine. We regularly use CentOS for this purpose. [root@dory ~]# chmod a+x /tmp/Xilinx_Vivado_SDK_2014.4_1119_1_Lin64.bin [root@dory ~]# /tmp/Xilinx_Vivado_SDK_2014.4_1119_1_Lin64.bin Use the graphical interface to install the “Hardware Server (Standalone)” package. The installation location is not important. Copy...
Continue reading →

Troubleshooting Metamux 3.0.0 beta version packet counts

Problem The beta version of MetaMux 3.0.0 does not currently adhere to Inter-Frame-Gap requirements, forcing queued packets out the egress application (ap) port hard back-to-back. This is a known issue in MetaMux 3.0.0beta versions that will be fixed in a future version. This is documented in the release notes. While all packets are transmitted out the front panel (no drops), some devices cannot handle this properly. An example of this is our et packet counters. The et packet counters do not count back-to-back packets correctly. Solution There is no solution for this at this time. Running “show etX counters verbose...
Continue reading →

How to evaluate Arista 7130 Layer 1 features

This guide helps you to get the most out of your Arista 7130 evaluation. It provides examples of how to investigate specific aspects of Arista 7130. Support site Play with it Latency Signal integrity Packet counters Media conversion Extract packets with tcpdump Web User Interface Remote upgradeability Market data broadcast and tapping Some nice UI features Access to Linux and management platform ContentsSupport sitePlay with itLatencySignal integrityPacket countersMedia conversionExtract packets with tcpdumpWeb User InterfaceRemote upgradeabilityMarket data broadcast and tappingSome nice UI featuresAccess to Linux and management platformhelloworld.pyhelloworld.cTalk to us Support site Visit Software Downloads and download the latest user guide....
Continue reading →

How to test SNMP traps on an Arista 7130 device

This is an experiment that you can perform using only the tools available on an Arista 7130 device. The purpose of this experiment is to demonstrate the behavior of the link up/down traps and explicitly show what information is contained in the notification. Hopefully, you can use this to bootstrap the correct config for whatever other tools you are using for SNMP. Config: ! command: show running-config ! time: Tue 27 Mar 2018 03:08:47 UTC ! device: A32-A1-24519-6 (A32-A1, MOS-0.16.8) interface et1 source mac interface et2 source mac management snmp snmp-server community public ro snmp-server host version 2c public...
Continue reading →


Get every new post on this blog delivered to your Inbox.

Join other followers: