Common Troubleshooting Issues

This section contains some common problems and how to fix them.

General Issues

Member unreachable

Check if the member’s details in IXP Manager matches with what the member actually has.

The most important ones are:

  • MAC Address

  • IPv4/6 Address

  • Port connection matches where the member’s connection is

A common scenario is a member performed maintenance and their MAC changed without notification.

Members lost traffic right after deployment

In the unlikely case that members lose traffic right after a deployment, make use of the rollback function. To start the rollback process, click File -> Rollback Cerberus config. This restores the configuration on the switches to the last working config. By default, the failed configuration will be saved to /etc/cerberus/failed/. This makes it easier to troubleshoot and debug issues at a later time.

Issues in Miru

Miru not showing up on the sidebar

If Miru is not on the sidebar, you most likely missed a step in the Miru installation process. Follow the below checklist to ensure everything is configured correctly. If things are missing, please follow the install instructions for Miru.

  • The belthazaar folder exists in ${IXPROOT}/vendor/ - If not, Miru was not installed. Follow all the instructions for the install.

  • That the miru folder exists in ${IXPROOT}/resources/skins/

  • Ensure that $IXPROOT/.env has the following config set in it VIEW_SKIN=miru

  • Ensure that ${IXPROOT}/public/mxgraph exists

  • Ensure that ${IXPROOT}/config/custom.php``has ``Athos set

  • Ensure that ${IXPROOT}/config/custom.php``has ``Cerberus set

Switches not showing up in Miru

  • Ensure that there are switches configured within IXP Manager

  • Ensure that the switches are configured as active

“error saving config”

Permission error for saving the configuration. Follow the steps in Athos troubleshoot to ensure permissions are configured correctly.

This error is only when storing the config directly and you are not using the athos api

No output when running tests

Check Athos logs, the issue is most likely due to a failure detected in testing that is taking too long to report back to Miru.

Clicking Start Tests does nothing

This indicates that there is an issue between Miru and Athos. Check the Athos troubleshoot section for more details.

Topology not saving

This indicates that there is an issue between Miru and Athos. Check the Athos troubleshoot section for more details.

No option to deploy

  • Ensure that ${IXPROOT}/config/custom.php``has ``cerberus and its api_url is configured properly

Athos

The best way to troubleshoot athos is by running the athos docker container outside of IXP Manager.

First ensure that you have the docker image for athos by running:

docker images belthazaar/athos

The latest image can also be pulled with docker image pull belthazaar/athos.

Within $ATHOSROOT run runDocker.sh as root.

This will start the athos docker container and run the tests of the last given topology.

If you are using the athos wrapper api, ensure that it is running with systemctl status athos_api.service. If it is not running, run systemctl start athos_api.service.

If you are not using the Athos wrapper API, ensure that your $MY_WWW_USER has permission to run the runDocker.sh script within $ATHOSROOT and has read and write permissions at minimum within the following locations:

  • $ATHOSROOT/etc

  • $ATHOSROOT/ixpman_files

Athos API wrapper

First thing to check would be that the wrapper is running. If you are running it as a service as described in the setup, you can check it’s status by running systemctl status athos_api.service. If not, you can always check the status with ps aux | grep athos to find all instances of athos and athos api.

Wrapper Talking to Athos

The API wrapper will need to be run as either root, or a user that has docker permissions with networking capabilities.

The quickest way to ensure that the wrapper API is working as intended is by running the following:

#Get the xml file for the topology
curl -X GET http://127.0.0.1:8989/get_xml
#Start and run an instance of Athos with output
curl -X GET http://127.0.0.1:8989/run_athos

If this does not work check to ensure that the athos directory used in the API wrapper is set to the directory where you have extracted Athos. The default location for both of these is /athos.

IXP Manager Configured correctly

Next would be to ensure that the configuration in ${IXPROOT}/config/custom.php matches how the wrapper is configured to run.

Below is a sample config looking for the default address of the wrapper api at http://127.0.0.1:8989.

return [
    'athos' => [
        // The directory where your athos is stored
        'dir' => '/athos',
        // The URL for the athos wrapper API
        'api_url' => 'http://localhost:8989'
    ],
    'cerberus' => [
        'api_url' => 'http://localhost:8080/api',
    ]
];

Cerberus

  • First step is to ensure that Cerberus is running and using the latest version. Check the version with cerberus-controller --version and check it matches the latest version.

  • Ensure that the Cerberus controller is running, if you have it setup as service check its status with sudo systemctl status cerberus.service.

  • Check configurations are set properly. The default configurations are: * /etc/cerberus/topology.json * /etc/cerberus/failed/ - Failed configurations location * /etc/cerberus/rollback/ - Rollback configurations location

  • Check the logs in /var/log/cerberus/cerberus.log