Frequently Asked Questions (FAQ)

A local OS error occurred while connecting to a device

A local OS error occurred while connecting to a device

When running ANTA, you can receive A local OS error occurred while connecting to <device> errors. The underlying OSError exception can have various reasons: [Errno 24] Too many open files or [Errno 16] Device or resource busy.

This usually means that the operating system refused to open a new file descriptor (or socket) for the ANTA process. This might be due to the hard limit for open file descriptors currently set for the ANTA process.

At startup, ANTA sets the soft limit of its process to the hard limit up to 16384. This is because the soft limit is usually 1024 and the hard limit is usually higher (depends on the system). If the hard limit of the ANTA process is still lower than the potential connections of all devices, the ANTA process may request to the operating system too many file descriptors and get an error, a WARNING is displayed at startup if this is the case.

Solution

One solution could be to raise the hard limit for the user starting the ANTA process. You can get the current hard limit for a user using the command ulimit -n -H while logged in. Create the file /etc/security/limits.d/10-anta.conf with the following content:

<user> hard nofile <value>

The user is the one with which the ANTA process is started. The value is the new hard limit. The maximum value depends on the system. A hard limit of 16384 should be sufficient for ANTA to run in most high scale scenarios. After creating this file, log out the current session and log in again.

Tests throttling WARNING in the logs

Tests throttling WARNING in the logs

ANTA is designed to execute many tests concurrently while ensuring system stability. If the total test count exceeds the maximum concurrency limit, tests are throttled to avoid overwhelming the asyncio event loop and exhausting system resources. A WARNING message is logged at startup when this occurs.

By default, ANTA schedules up to 50000 tests concurrently. This should be sufficient for most use cases, but it may not be optimal for every system. If the number of tests exceeds this value, ANTA executes the first 50000 tests and waits for some tests to complete before executing more.

Solution

You can adjust the maximum concurrency limit using the ANTA_MAX_CONCURRENCY environment variable. The optimal value depends on your system CPU usage, memory consumption, and file descriptor limits.

Warning

Increasing the maximum concurrency limit can lead to system instability if the system is not able to handle the increased load. Monitor system resources and adjust the limit accordingly.

Device Connection Limits

Each EOS device is limited to a maximum of 100 concurrent connections. This means that, even if ANTA schedules a high number of tests, it will only attempt to open up to 100 connections at a time towards each device.

Tip

If you run ANTA on a large fabric or encounter issues related to resource limits, consider tuning ANTA_MAX_CONCURRENCY. Test different values to find the optimal setting for your environment.

Timeout error in the logs

Timeout error in the logs

When running ANTA, you can receive <Foo>Timeout errors in the logs (could be ReadTimeout, WriteTimeout, ConnectTimeout or PoolTimeout). More details on the timeouts of the underlying library are available here: https://www.python-httpx.org/advanced/timeouts.

This might be due to the time the host on which ANTA is run takes to reach the target devices (for instance if going through firewalls, NATs, …) or when a lot of tests are being run at the same time on a device (eAPI has a queue mechanism to avoid exhausting EOS resources because of a high number of simultaneous eAPI requests).

Solution

Use the timeout option. As an example for the nrfu command:

anta nrfu --enable --username username --password arista --inventory inventory.yml -c nrfu.yml --timeout 50 text

In this command, ANTA NRFU is configured with several options. Notably, the --timeout parameter is set to 50 seconds (instead of the default 30 seconds) to allow extra time for API calls to complete.

ImportError related to urllib3

ImportError related to urllib3 when running ANTA

When running the anta --help command, some users might encounter the following error:

ImportError: urllib3 v2.0 only supports OpenSSL 1.1.1+, currently the 'ssl' module is compiled with 'OpenSSL 1.0.2k-fips  26 Jan 2017'. See: https://github.com/urllib3/urllib3/issues/2168

This error arises due to a compatibility issue between urllib3 v2.0 and older versions of OpenSSL.

Solution

1. Workaround: Downgrade urllib3

   If you need a quick fix, you can temporarily downgrade the urllib3 package:

   pip3 uninstall urllib3
   
   pip3 install urllib3==1.26.15
   

2. Recommended: Upgrade System or Libraries:

   As per the [urllib3 v2 migration guide](https://urllib3.readthedocs.io/en/latest/v2-migration-guide.html), the root cause of this error is an incompatibility with older OpenSSL versions. For example, users on RHEL7 might consider upgrading to RHEL8, which supports the required OpenSSL version.
   

AttributeError: module 'lib' has no attribute 'OpenSSL_add_all_algorithms'

AttributeError: module 'lib' has no attribute 'OpenSSL_add_all_algorithms' when running ANTA

When running the anta commands after installation, some users might encounter the following error:

AttributeError: module 'lib' has no attribute 'OpenSSL_add_all_algorithms'

The error is a result of incompatibility between cryptography and pyopenssl when installing asyncssh which is a requirement of ANTA.

Solution

1. Upgrade pyopenssl

   pip install -U pyopenssl>22.0
   

Caveat running on non-POSIX platforms (e.g. Windows)

Caveat running on non-POSIX platforms (e.g. Windows)

While ANTA should in general work on non-POSIX platforms (e.g. Windows), there are some known limitations:

• On non-Posix platforms, ANTA is not able to check and/or adjust the system limit of file descriptors.

ANTA test suite is being run in the CI on a Windows runner.

__NSCFConstantString initialize error on OSX

__NSCFConstantString initialize error on OSX

This error occurs because of added security to restrict multithreading in macOS High Sierra and later versions of macOS. https://www.wefearchange.org/2018/11/forkmacos.rst.html

Solution

1. Set the following environment variable

   export OBJC_DISABLE_INITIALIZE_FORK_SAFETY=YES
   

EOS AAA configuration for an ANTA-only user

EOS AAA configuration for an ANTA-only user

Here is a starting guide to configure an ANTA-only user to run ANTA tests on a device.

Warning

This example is not using TACACS / RADIUS but only local AAA

1. Configure the following role.

   role anta-users
      10 permit command show
      20 deny command .*
   

   You can then add other commands if they are required for your test catalog (ping for example) and then tighten down the show commands to only those required for your tests. To figure out the full list of commands used by your catalog or ANTA in general, you can use anta get commands

2. Configure the following authorization (You may need to adapt depending on your AAA setup).

   aaa authorization commands all default local
   

3. Configure a user for the role.

   user anta role anta-users secret <secret>
   

4. You can then use the credentials anta / <secret> to run ANTA against the device and adjust the role as required.

Still facing issues?

If you’ve tried the above solutions and continue to experience problems, please follow the troubleshooting instructions and report the issue in our GitHub repository.