Changes in Release 1.5

New Functionality 

Release 1.5.4 

Juniper vPTX/vEVO support by Stefano Sasso (initial configuration, IPv4+IPv6, OSPFv2/v3, IS-IS, BGP, BFD, VLANs, VRFs, MPLS, SR-MPLS)
Juniper vSRX can be run as a vrnetlab-packaged container (by Stefano Sasso)
Documentation migrated to netlab.tools

Release 1.5.3 

‘unknown’ device type can be used to deploy devices not yet supported by netlab
Public libvirt networks can be used to connect labs to outside world
netlab can use (and preserve) existing libvirt networks
Maximum node name length is configurable.
Add MPLS VPN support for SR OS
Improved virtualization provider installation testing with (hopefully) useful error messages
Support for containerlab 0.41.0

Breaking changes:

containerlab 0.41.0 is the minimum version of containerlab that works with release 1.5.3
Strict validation of provider- and tool attributes might break some advanced topologies

Release 1.5.2 

External network management tools
Aruba AOS-CX Support
Reusable topology components
Tunnel interfaces
Add --force flag to netlab down command
Use Ansible filters in netlab templating system
Deploy per-node custom configuration templates
Dry run mode for netlab up and netlab down commands.

Release 1.5.1 

Simplified definition of VLAN access links
Simplified definition of VRF links
Link groups
Output module for D2 diagram scripting language
Stricter type checks for node names, group names, VLAN names, VRF names, link roles, address pools, device names, and Linux bridges. Names must be identifiers: alphanumeric strings up to 16 characters long.
netlab status file tracks active lab instances
netlab create command refuses to create provider configuration files in a directory with a running lab instance.
netlab up command checks the status file to prevent multiple copies of the same lab instance.

Breaking changes:

Identifiers (names) in existing lab topologies might be too long to meet the identifier requirements. You will have to shorten those names.
Default ASAv password was changed. You might have to rebuild your ASAv boxes.

Release 1.5.0 

Multi-provider topologies
Running multiple lab instances on Linux Servers
Split large libvirt topology into smaller vagrant up batches
Use custom network/bridge name and IP subnet for vagrant-libvirt management network
Create DHCP bindings for the whole vagrant-libvirt management subnet
‘loopback’ link type can be used to create additional loopback interfaces
netlab show devices command displays valid device types
Split topology-defaults.yml into numerous smaller files
Strict validation of core and module attributes

Breaking changes 

Release 1.5.3 

Containerlab changed topology attribute names in release 0.41.0, forcing us to change topology file template and completely breaking backward compatibility.
Release 1.5.3 implements strict validation of provider- and tool attributes, which might trigger errors in some topologies. Please open an issue if we missed a provider attribute you need.

Release 1.5.2 

We switched to Python Box 7.0.x and replaced tons of calls to get_from_box (internal function to get a value from a hierarchical key) with calls to Box .get method. That migration might have broken some obscure code path that is not covered with our CI/CD tests.
Loopback interface names for loopback links and VRF loopbacks are sequential within a single node, starting with one. Previously, loopback interfaces created from links used linkindex (a low value based on links list), and VRF loopbacks used vrfidx (a value starting with 100).

Release 1.5.1 added stricter checks of node names, group names, VLAN names, VRF names, link roles, addressing pools, devices, and Linux bridges. These names must be identifiers: alphanumeric strings (plus hyphen and underscore) up to 16 characters long. You might have to shorten too-verbose names in your topology files.

ASAv Default Password

Newer ASAv versions enforce a minimum password length of 8. The default password (vagrant) was changed to vagrant1. You should rebuild your ASAv Vagrant boxes to use the new password (using the new build recipe), or change the defaults through user default file containing:

devices.asav.group_vars.ansible_ssh_pass: vagrant
devices.asav.group_vars.ansible_become_password: vagrant

Release 1.5.0 

Release 1.5.0 introduced strict validation of topology attributes. We ran extensive tests, including all the integration tests, so we’re pretty sure that whatever we ever wanted to use still works. Obviously it’s possible you used a combination of attributes that we did not test, or that we did not check attribute types in the past and unexpected types just worked.

We had to change the underlying data structure from lists of attributes to hierarchical dictionaries, we expect most plugins that added new core- or module attributes to break. Please take a look at the updated plugins in netsim/extra directory to see how your plugin should add new attributes (or open a GitHub case and we’ll help you).

Bug Fixes 

Bug Fixes in Release 1.5.4-post1 

Implement FRR on libvirt/Virtualbox as Ubuntu VM + FRR package
Make packet forwarding on Linux devices configurable
Control installation of Ubuntu packages with group variables
Make the list of ignored attributes configurable and allow ‘netlab_’ attributes in nodes
FRR BGP template fixes/enhancements
ebgp.utils implementation for FRR
Create empty links list if needed when using VLAN module
Merge global VRF data with node VRF data in post-transform hook
Add missing variable to ‘include_task’ module to display context-dependent task names during initial configuration
Mark ‘foreign’ nodes in multi-provider topology as ‘unmanaged’
Allow clab file binds templates in subdirectories
Consider node-specific provider when deploying device configurations

Bug Fixes in Release 1.5.4 

Fix Vagrant version in installation script and document supported versions for Vagrant and vagrant-libvirt plugin
Accept any Python expression as a format modifier for YAML/JSON output modules
Build default node.af dictionary from loopback addresses
Skip provider-specific shutdown steps if the provider is not installed
Fixes for Nokia SR OS IXR ECMP profiles
Replace ‘_’ with ‘-’ in Cumulus/Linux hostname
Copy BGP interface attributes from EBGP neighbors into bgp.neighbors data structure
Convert ‘netlab up’ CLI settings into int/bool if possible
Strip spaces from node names used in A-B link format
Nokia SR/OS: configure static blackhole routes for originated BGP networks
Enable ports in Nokia SR/OS initial configuration
Change containerlab release to 0.41.2

Bug Fixes in Release 1.5.3-post1 

Allow srl-agents clab provider attribute
Pick specific containerlab version in netlab install containerlab installation script
Update containerlab topology template for IPv6 management

Bug Fixes in Release 1.5.3 

Structured virtualization provider probes result in better error messages during ‘netlab up’ and ‘netlab down’
Strict validation of provider- and tool attributes
Use libvirt P2P UDP tunnels only on links with two nodes and link.type == p2p
Link to IOS XR build recipe was missing
Add sphinx-rtd-theme to docs requirements
containerlab multi-line startup-config
Load host MPLS kernel module before configuring MPLS in FRR containers.
Support SR Linux r23.3.1
SR Linux: Check if a VLAN has EVPN enabled
Fix MPLS EVPN transport configuration for SR OS
Fix IPv6 LLA configuration on Nexus OS
Abort if python-box version is lower than 7.0

Bug Fixes in Release 1.5.2 

D2 output: add list of interfaces to nodes
Allow IP address allocations that barely fit into a prefix (like 2 nodes into /30)
Unify loopback and virtual interface processing
Ignore node provider equal to topology provider
Avoid dependency on Python for custom container configuration templates/scripts
SR Linux/SROS: OSPF related fixes
SR Linux: Use passive flag from interface
SR OS: Remove VRF dependency on BGP module
Deal with Python Box 7.0 expanding dots in keys read from YAML files