Skip to main content
  1. Blog
  2. Article

Canonical
on 11 July 2017

MAAS Development Summary – July 3rd – 14th


Hello MAASters!

The purpose of this update is to keep our community engaged and informed about the work the team is doing. We’ll cover important announcements, work-in-progress for the next release of MAAS and bugs fixes in release MAAS versions.

MAAS 2.3 (current development release)

  • Completed Django 1.11 transition
      • MAAS 2.3 snap will use Django 1.11 by default.
      • Ubuntu package will use Django 1.11 in Artful+
  • Network beaconing & better network discovery
      • MAAS now listens for [unicast and multicast] beacons on UDP port 5240. Beacons are encrypted and authenticated using a key derived from the MAAS shared secret. Upon receiving certain types of beacons, MAAS will reply, confirming the sender that existing MAAS on the network has the same shared key. In addition, records are kept about which interface each beacon was received on, and what VLAN tag (if any) was in use on that interface. This allows MAAS to determine which interfaces observed the same beacon (and thus must be on the same fabric). This information can also determine if [what would previously have been assumed to be] a separate fabric is actually an alternate VLAN in an existing fabric.
      • The maas-rack send-beacons command is now available to test the beacon protocol. (This command is intended for testing and support, not general use.) The MAAS shared secret must be installed before the command can be used. By default, it will send multicast beacons out all possible interfaces, but it can also be used in unicast mode.
      • Note that while IPv6 support is planned, support for receiving IPv6 beacons in MAAS is not yet available. The maas-rack send-beacons command, however, is already capable of sending IPv6 beacons. (Full IPv6 support is expected to make beacons more flexible, since IPv6 multicast can be sent out on interfaces without a specific IP address assignment, and without resorting to raw sockets.)
      • Improvements to rack registration are now under development, so that users will see a more accurate representation of fabrics upon initial installation or registration of a MAAS rack controller.
  • Bug fixes
    • LP: #1701056: Show correct information for a device details page as a normal user
    • LP: #1701052: Do not show the controllers tab as a normal user
    • LP: #1683765: Fix format when devices/controllers are selected to match those of machines
    • LP: #1684216 – Update button label from ‘Save selection’ to ‘Update selection’
    • LP: #1682489 – Fix Cancel button on add user dialog, which caused the user to be added anyway
    • LP: #1682387 – Unassigned should be (Unassigned)

MAAS 2.2.1

The past week the team was also focused on preparing and QA’ing the new MAAS 2.2.1 point release, which was released on Friday June the 30th. For more information about the bug fixes please visit the following https://launchpad.net/maas/+milestone/2.2.1.

MAAS 2.2.1 is available in:

  • ppa:maas/stable

Related posts


Canonical
16 March 2026

Meet Canonical at NVIDIA GTC 2026: NVIDIA CUDA and NVIDIA Vera Rubin NVL72 support in Ubuntu 26.04 LTS

Ubuntu Article

Previewing at NVIDIA GTC 2026: NVIDIA CUDA support in Ubuntu 26.04 LTS, NVIDIA Vera Rubin NVL72 architecture support in Ubuntu 26.04, Canonical’s official Ubuntu image for NVIDIA Jetson Thor, upcoming support for NVIDIA DGX Station and NVIDIA DOCA-OFED, and NVIDIA RTX PRO 4500 support. NVIDIA GTC 2026 is here, bringing together the techno ...


Luci Stanescu
12 March 2026

AppArmor vulnerability fixes available

Ubuntu Article

Qualys discovered several vulnerabilities in the AppArmor code of the Linux kernel. These are being referred to as CrackArmor, while CVE IDs have not been assigned yet. All of the vulnerabilities require unprivileged local user access. The impact of these vulnerabilities ranges from denial of service to kernel memory information leak, rem ...


David Beamonte
11 March 2026

The bare metal problem in AI Factories

MAAS MAAS

As AI platforms grow into large-scale “AI Factories,” the real bottleneck shifts from model design to operational complexity. With expensive GPU accelerators, hardware failures and inconsistent configurations lead directly to lost throughput and reduced return on investment. While Kubernetes orchestrates workloads, it cannot fix broken ph ...