Go to Table of Contents

回到概览

Incus 6.5 has been released

2024年9月6日

Introduction

The Incus team is pleased to announce the release of Incus 6.5!

A strong focus for this release was on performance. Costly internal calls like resolving large number of profiles and devices have been optimized significantly leading to up to a 20-30x performance improvement. Similarly, handling of systems with thousands of instances per server has also been greatly improved, cutting down startup checks from tens of minutes down to tens of seconds.

But this isn't just a bugfix release, Incus 6.5 also introduces quite a few new features and other improvements. From making our CLI experience more consistent, to making it easier to perform low-level actions on virtual machines, to improving the life of application container users and through a number of great new features for OVN users, this release should have something for everyone!

image|690x190

As usual, you can try it for yourself online: https://linuxcontainers.org/incus/try-it/

Enjoy!

New features

Instance auto-restart

A common request ever since we first rolled out application containers support in Incus, the ability to have instances automatically restart when they exit makes it easier to handle applications crashing or reloading.

This is controlled through a new boot.autorestart configuration key which when set to true will have Incus attempt to restart a given instance up to 10 times over a 1 minute time span.

User requested instance shutdown/stop do not trigger the auto-restart logic.

stgraber@castiana:~$ incus launch docker:nginx nginx -c boot.autorestart=true
Launching nginx
stgraber@castiana:~$ incus info nginx | grep PID
PID: 178789
stgraber@castiana:~$ sudo kill -9 178789
stgraber@castiana:~$ incus list nginx
+-------+---------+----------------------+-----------------------------------------------+-----------------+-----------+
| NAME  |  STATE  |         IPV4         |                     IPV6                      |      TYPE       | SNAPSHOTS |
+-------+---------+----------------------+-----------------------------------------------+-----------------+-----------+
| nginx | RUNNING | 10.178.240.76 (eth0) | fd42:8384:a6f8:63a0:216:3eff:fef4:5a27 (eth0) | CONTAINER (APP) | 0         |
+-------+---------+----------------------+-----------------------------------------------+-----------------+-----------+
stgraber@castiana:~$

Documentation: https://linuxcontainers.org/incus/docs/main/reference/instance_options/#boot-related-options

Column selection in all list commands

Over the past few releases, we've been working on improving the consistency of the incus CLI commands. This started with making sure that all our list commands support --format and now with this release, all list commands also now support --columns.

This allows for easily customizing the output you're getting, as well as making it much easier to script the incus command by combining both --format=csv with --column= to select just the relevant column(s).

stgraber@castiana:~$ incus snapshot list v1 --columns=nT --format=csv
snap0,2024/09/06 15:04 EDT
snap1,2024/09/06 15:04 EDT

QMP command hooks and scriptlet

Incus currently relies on QEMU to run its virtual machines.

The way Incus interacts with QEMU can be a bit complex at times as it's effectively done through three different mechanisms:

  • QEMU command line
  • QEMU configuration file
  • QEMU Machine Protocol (QMP)

We usually try to avoid polluting the command line as much as possible, so this is kept to a minimum, but we allow the user to pass in additional arguments through raw.qemu.

Our preference for any device which doesn't need live-updating or doesn't need to ever be hotplugged or hot removed is the use of the QEMU configuration file. This is easily templated and can pretty easily be tested. We have the raw.qemu.conf configuration option that can be used to extend or override the content of that configuration file.

And then we have QMP which we use for anything hotpluggable, so effectively all disks, network interfaces, USB devices or any other PCI devices. As the QEMU team is slowly trying to deprecate the configuration file, we expect to progressively be moving more and more of the VM configuration over to QMP.

The main issue with QMP so far has been that unlike the QEMU command line or a config file, it's very opaque. It's not possible to easily see what's been configured and because any of those objects will have been configured after QEMU started, it wasn't possible to override or re-configure them through the existing mechanisms.

But things are different now thanks to a few new configuration options:

  • raw.qemu.qmp.early
  • raw.qemu.qmp.pre-start
  • raw.qemu.qmp.post-start
  • raw.qemu.scriptlet

The first three take a JSON list of QMP commands. QMP commands are normally already all JSON encoded, so that makes it easy to add a number of custom commands to the instance configuration. The commands will be run in order at one of the specified times.

early runs prior to Incus having added anything through QMP, pre-start runs after Incus has added all its devices through QMP and post-start runs immediately after QEMU was instructed to start the VM.

raw.qemu.scriptlet is an even more flexible option as it takes a sriptlet (python-like syntax) which must define a function named qemu_hook and passes it the stage as an argument. That stage is one of early, pre-start or post-start. The different between that and the raw.qemu.qmp options is that a scriptlet can handle command responses and have logic to react to it.

That means that this QEMU scriptlet can call the run_qmp command, pass a custom QMP command, read through its return value and issue more commands if needed, allowing for dynamic re-configuration of the VM.

Note that this is a very low level mechanism which we only expect expert users to use in very specific cases. As with any raw configuration key, its use is effectively unsupported by the Incus team and it should also be kept disabled for any untrusted projects.

Live disk resize support in virtual machines

It's now possible to resize either the VM root disk or any attached disk and have the VM be notified of the change. This then causes the operating system to update the size of the disk and allows the user to immediately make use of the additional space without having to restart the VM.

stgraber@castiana:~$ incus exec v1 bash
root@v1:~# lsblk
NAME   MAJ:MIN RM  SIZE RO TYPE MOUNTPOINTS
sda      8:0    0   10G  0 disk
├─sda1   8:1    0  100M  0 part /boot/efi
└─sda2   8:2    0  9.9G  0 part /
root@v1:~#
exit

stgraber@castiana:~$ incus config device override v1 root size=20GiB
Device root overridden for v1

stgraber@castiana:~$ incus exec v1 bash
root@v1:~# lsblk
NAME   MAJ:MIN RM  SIZE RO TYPE MOUNTPOINTS
sda      8:0    0   20G  0 disk
├─sda1   8:1    0  100M  0 part /boot/efi
└─sda2   8:2    0  9.9G  0 part /
root@v1:~#

PCI devices hotplug

Adding and removing PCI devices on a VM can now be done live.
This now matches the behavior found in NIC, GPU and disk devices.

OVN load-balancer health checks

Incus' support for OVN load-balancers has so far been pretty basic, essentially being limited to just basic load-balancing of traffic on a set of target with no monitoring of the backend.

But this is now changing with initial support for OVN's load-balancer health checks.
This is configured through a set of configuration keys on the load-balancer:

  • healthcheck => Enables health checking
  • healthcheck.failure_count => Number of failed attempts to consider backend as failed
  • healthcheck.interval => How often to check the backends (in seconds)
  • healthcheck.success_count => Number of successful attempts to consider backend as online
  • healthcheck.timeout => How long to wait for a response before considering it failed

Only healthcheck is required, all the others have reasonable defaults.

root@server01:~# incus launch images:ubuntu/24.04 c1
Launching c1
root@server01:~# incus exec c1 -- apt-get install --yes nginx
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
The following additional packages will be installed:
  nginx-common
Suggested packages:
  fcgiwrap nginx-doc ssl-cert
The following NEW packages will be installed:
  nginx nginx-common
0 upgraded, 2 newly installed, 0 to remove and 0 not upgraded.
Need to get 552 kB of archives.
After this operation, 1596 kB of additional disk space will be used.
Get:1 http://archive.ubuntu.com/ubuntu noble/main amd64 nginx-common all 1.24.0-2ubuntu7 [31.2 kB]
Get:2 http://archive.ubuntu.com/ubuntu noble/main amd64 nginx amd64 1.24.0-2ubuntu7 [521 kB]
Fetched 552 kB in 1s (619 kB/s)
Preconfiguring packages ...
Selecting previously unselected package nginx-common.
(Reading database ... 16176 files and directories currently installed.)
Preparing to unpack .../nginx-common_1.24.0-2ubuntu7_all.deb ...
Unpacking nginx-common (1.24.0-2ubuntu7) ...
Selecting previously unselected package nginx.
Preparing to unpack .../nginx_1.24.0-2ubuntu7_amd64.deb ...
Unpacking nginx (1.24.0-2ubuntu7) ...
Setting up nginx (1.24.0-2ubuntu7) ...
Setting up nginx-common (1.24.0-2ubuntu7) ...
Created symlink /etc/systemd/system/multi-user.target.wants/nginx.service → /usr/lib/systemd/system/nginx.service.
root@server01:~# incus launch images:ubuntu/24.04 c2
Launching c2
root@server01:~# incus list
+------+---------+--------------------+-----------------------------------------------+-----------+-----------+----------+
| NAME |  STATE  |        IPV4        |                     IPV6                      |   TYPE    | SNAPSHOTS | LOCATION |
+------+---------+--------------------+-----------------------------------------------+-----------+-----------+----------+
| c1   | RUNNING | 10.104.61.2 (eth0) | fd42:73ae:9013:c530:216:3eff:feff:ddf2 (eth0) | CONTAINER | 0         | server01 |
+------+---------+--------------------+-----------------------------------------------+-----------+-----------+----------+
| c2   | RUNNING | 10.104.61.3 (eth0) | fd42:73ae:9013:c530:216:3eff:fec4:611 (eth0)  | CONTAINER | 0         | server02 |
+------+---------+--------------------+-----------------------------------------------+-----------+-----------+----------+

root@server01:~# incus network load-balancer create default 172.31.254.50
Network load balancer 172.31.254.50 created
root@server01:~# incus network load-balancer backend add default 172.31.254.50 c1 10.104.61.2
root@server01:~# incus network load-balancer backend add default 172.31.254.50 c2 10.104.61.3
root@server01:~# incus network load-balancer port add default 172.31.254.50 tcp 80 c1,c2

root@server01:~# incus launch images:ubuntu/24.04 t1
Launching t1
root@server01:~# incus exec t1 -- nc -v 172.31.254.50 80
nc: connect to 172.31.254.50 port 80 (tcp) failed: Connection refused
root@server01:~# incus exec t1 -- nc -v 172.31.254.50 80
nc: connect to 172.31.254.50 port 80 (tcp) failed: Connection refused
root@server01:~# incus exec t1 -- nc -v 172.31.254.50 80
Connection to 172.31.254.50 80 port [tcp/http] succeeded!

root@server01:~# incus network load-balancer set default 172.31.254.50 healthcheck=true

root@server01:~# incus exec t1 -- nc -v 172.31.254.50 80
Connection to 172.31.254.50 80 port [tcp/http] succeeded!
^Croot@server01:~# incus exec t1 -- nc -v 172.31.254.50 80
Connection to 172.31.254.50 80 port [tcp/http] succeeded!
^Croot@server01:~# incus exec t1 -- nc -v 172.31.254.50 80
Connection to 172.31.254.50 80 port [tcp/http] succeeded!
^Croot@server01:~# incus exec t1 -- nc -v 172.31.254.50 80
Connection to 172.31.254.50 80 port [tcp/http] succeeded!
^Croot@server01:~# incus exec t1 -- nc -v 172.31.254.50 80
Connection to 172.31.254.50 80 port [tcp/http] succeeded!

Documentation: https://linuxcontainers.org/incus/docs/main/howto/network_load_balancers/

ECMP support for OVN interconnect

The network integration support for OVN interconnect has been extended in a few small ways:

  • The ovn.transit.pattern configuration option now supports a new peerName variable
  • It's now possible to have multiple peers on a network targeting the same network integration
  • IP allocation on the transit switch is now recorded directly in the OVN database rather than relying on random subnets

The end result is that it's now possible to change the default core.transit.pattern to include peerName in the template and then add multiple peers to a network, all pointing to the same interconnection.

This internally will result in mutliple transit switches being created and so long as the peer names match on all participating systems, traffic will be balanced between those switches through ECMP.

Doing so allows for very effective load-balancing of interconnection traffic.

root@chulak:~# incus list ic
+---------+---------+--------------------+-----------------------------------------------+-----------+-----------+----------+
|  NAME   |  STATE  |        IPV4        |                     IPV6                      |   TYPE    | SNAPSHOTS | LOCATION |
+---------+---------+--------------------+-----------------------------------------------+-----------+-----------+----------+
| ic-test | RUNNING | 10.47.238.2 (eth0) | fd42:4a11:5600:6807:216:3eff:feb5:2c79 (eth0) | CONTAINER | 0         | chulak   |
+---------+---------+--------------------+-----------------------------------------------+-----------+-----------+----------+
root@chulak:~# incus exec ic-test bash
root@ic-test:~# ping 10.170.69.2
PING 10.170.69.2 (10.170.69.2) 56(84) bytes of data.
From 45.45.148.162 icmp_seq=1 Destination Net Unreachable
--- 10.170.69.2 ping statistics ---
1 packets transmitted, 0 received, +1 errors, 100% packet loss, time 0ms
root@ic-test:~#

root@chulak:~# incus network peer create ovn-ic-test peer1 dcmtl --type=remote
Network peer peer1 created
root@chulak:~# incus network peer create ovn-ic-test peer2 dcmtl --type=remote
Network peer peer2 created
root@chulak:~# incus network peer create ovn-ic-test peer3 dcmtl --type=remote
Network peer peer3 created
root@chulak:~# incus network peer create ovn-ic-test peer4 dcmtl --type=remote
Network peer peer4 created

root@chulak:~# incus exec ic-test bash
root@ic-test:~# ping 10.170.69.2
PING 10.170.69.2 (10.170.69.2) 56(84) bytes of data.
64 bytes from 10.170.69.2: icmp_seq=1 ttl=62 time=11.8 ms
64 bytes from 10.170.69.2: icmp_seq=2 ttl=62 time=6.01 ms
--- 10.170.69.2 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1002ms
rtt min/avg/max/mdev = 6.012/8.930/11.848/2.918 ms

Documentation: https://linuxcontainers.org/incus/docs/main/howto/network_integrations/

Promiscuous mode for OVN NICs

A new security.promiscuous configuration key is now available on OVN NICs.

When it's enabled, any OVN traffic that has an unknown MAC address as its destination will now be sent over to the OVN NIC.

The main use for this is for nested environments where you want to have some nested containers or VMs directly sit on the parent OVN network without having their own dedicated ports.
This is typically a development/testing use case as promiscuous mode causes a lot of unnecessary network traffic to hit the NIC.

root@server01:~# incus launch images:ubuntu/24.04 t1
Launching t1
root@server01:~# incus exec t1 bash
root@t1:~# ip l
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
48: eth0@if49: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1422 qdisc noqueue state UP mode DEFAULT group default qlen 1000
    link/ether 00:16:3e:f3:d4:3e brd ff:ff:ff:ff:ff:ff link-netnsid 0
root@t1:~# ip link set eth0 address 00:16:3e:f3:d4:30
root@t1:~# ip -4 a add dev eth0 10.104.61.100/24
root@t1:~# ping 10.104.61.1
PING 10.104.61.1 (10.104.61.1) 56(84) bytes of data.
^C
--- 10.104.61.1 ping statistics ---
2 packets transmitted, 0 received, 100% packet loss, time 1009ms

root@t1:~#
exit

root@server01:~# incus config device override t1 eth0 security.promiscuous=true
Device eth0 overridden for t1
root@server01:~# incus exec t1 bash
root@t1:~# ip link set eth0 address 00:16:3e:f3:d4:30
root@t1:~# ip -4 a add dev eth0 10.104.61.100/24
root@t1:~# ping 10.104.61.1
PING 10.104.61.1 (10.104.61.1) 56(84) bytes of data.
64 bytes from 10.104.61.1: icmp_seq=1 ttl=254 time=1.20 ms
^C
--- 10.104.61.1 ping statistics ---
1 packets transmitted, 1 received, 0% packet loss, time 0ms
rtt min/avg/max/mdev = 1.197/1.197/1.197/0.000 ms
root@t1:~#

Documentation: https://linuxcontainers.org/incus/docs/main/reference/devices_nic/#nictype-ovn

Ability to run off IP allocation on OVN NICs

Another new OVN NIC option is the ability to turn off IP allocation completely.

This is often related to the previous case where a promiscuous NIC typically doesn't need to have its own IPv4 and IPv6 address. To handle this, it's now possible to set both ipv4.address and ipv6.address to none, disabling allocations.

Note that OVN doesn't allow disabling just one protocol, so both keys must currently be set to none for this to work.

root@server01:~# incus config device set t1 eth0 ipv4.address=none ipv6.address=none
root@server01:~# incus start t1
root@server01:~# incus exec t1 bash
root@t1:~# ip -4 a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
root@t1:~#

Documentation: https://linuxcontainers.org/incus/docs/main/reference/devices_nic/#nictype-ovn

Customizable OIDC scope request

It's now possible to configure the list of OpenID Connect Scopes that are being requested.

Setting oidc.scopes in the server config will override the default of openid, offline_access and can be useful to pull in additional information through scopes like profile.

Documentation: https://linuxcontainers.org/incus/docs/main/server_config/#openid-connect-configuration

Configurable LVM PV metadata size

Very very large LVM volumes groups containing thousands of logical volumes may exceed the reserved metadata size.

This was already configurable on LVM thin provisioned pools (default), but for thick provisioning, there was no matching configuration.

Now the lvm.metadata_size configuration key can be set to override LVM's default.
Note that this can only be done at creation time.

stgraber@castiana:~$ incus storage create demo lvm lvm.use_thinpool=false
Storage pool demo created
stgraber@castiana:~$ sudo vgs -o name,mda_size
  VG   VMdaSize
  demo  1020.00k
stgraber@castiana:~$ incus storage delete demo
Storage pool demo deleted

stgraber@castiana:~$ incus storage create demo lvm lvm.use_thinpool=false lvm.metadata_size=100MiB
Storage pool demo created
stgraber@castiana:~$ sudo vgs -o name,mda_size
  VG   VMdaSize
  demo  <101.00m
stgraber@castiana:~$ incus storage delete demo
Storage pool demo deleted

Documentation: https://linuxcontainers.org/incus/docs/main/reference/storage_lvm/#configuration-options

Configurable OVS socket path

There are a few cases where OpenVSwitch doesn't run at the usual address.

The most common case of this would be MicroOVN users where the OpenVSwitch socket is instead stored within /var/snap/microovn/common/....

Until now, those users had to jump through some hoops to get a working OVS socket in /run so Incus would properly connect to it.

With this change, it's now possible to set the network.ovs.connection configuration key to a valid OVSDB connection string and have Incus reach OpenVSwitch through that. The default value is unix:/run/openvswitch/db.sock.

Documentation: https://linuxcontainers.org/incus/docs/main/server_config/#server-options-misc

Complete changelog

Here is a complete list of all changes in this release:

Full commit list
  • incus/remote/list: Add support for column selection
  • i18n: Update translation templates
  • incus/cluster/group/list: Add support for column selection
  • i18n: Update translation templates
  • Translated using Weblate (Chinese (Simplified))
  • Translated using Weblate (Chinese (Simplified))
  • client: import examples for docs
  • client: name var for docs
  • client: alias & server/procotol default for docs
  • incusd/storage: Fix UsedBy values for sub-directory volumes
  • incusd/instance: Fix backup file locking issue
  • incusd/projects: Don't fail project creation on missing pools
  • incusd/device/pci: Allow hotplug
  • incusd/instance/qmp: Add CheckPCIDevice
  • incusd/instance/qemu: Use monitor.CheckPCIDevice
  • incusd/instance/qemu: Tweak comments on deviceStart
  • incusd/instance/qemu: Add hotplug support for generic PCI
  • client: fix typo in example
  • Translated using Weblate (Chinese (Simplified))
  • incus/operation/list: Add support for column selection
  • i18n: Update translation templates
  • doc/firewalld: Update Docker link
  • incus/network/zone/list: Add support for column selection
  • i18n: Update translation templates
  • incusd/instance/drivers/qmp: Export RunJSON
  • api: qemu_raw_qmp
  • incusd/instance: Add raw QMP config options
  • doc: Add QMP to wordlist
  • doc: Update configs
  • incusd/instance/qemu: Add QMP hooks
  • incusd/project: Update low-level properties
  • incus/network/forward/list: Add support for column selection
  • i18n: Update translation templates
  • Translated using Weblate (Chinese (Simplified))
  • cmd/incusd: Add hostname to dhcp request
  • incus/network/list-leases: Add support for column selection
  • i18n: Update translation templates
  • Translated using Weblate (Chinese (Simplified))
  • doc: Update incus_alias.md
  • incus/network/list-allocations: Add support for column selection
  • i18n: Update translation templates
  • api: network_load_balancer_health_check
  • incusd/network/ovn: Simplify CreateLoadBalancer
  • incusd/network: Update for CreateLoadBalancer changes
  • incusd/network/ovn: Add healthcheck support in LoadBalancer
  • incusd/network: Add healthcheck config options
  • incusd/network/ovn: Add healthcheck options
  • incusd/network/ovn: Reserve the last IPv4 address
  • doc/network/load_balancer: Add configuration options
  • doc: Update configs
  • incus/admin/init: Prompt for dir storage location
  • tests: Update for extra step in init
  • i18n: Update translation templates
  • incus/network/integration/list: Add support for column selection
  • i18n: Update translation templates
  • incus/storage/bucket/list: Add support for column selection
  • i18n: Update translation templates
  • api: oidc_scopes
  • incusd/config: Add oidc.scopes
  • incusd/oidc: Add custom scopes support
  • doc: Update configs
  • incus/storage/bucket: Add support for column selection in key list
  • i18n: Update translation templates
  • incus/snapshot/list: Add support for column selection
  • i18n: Update translation templates
  • incusd/storage/lvm: Fix resize logic to conserve LV state
  • incusd/network/ovn: Set missing send_periodic field
  • incusd/profiles: Improve listing performance
  • incusd/server/db: Increase transaction deadline to 30s
  • incusd/db/profiles: Support device cache in ToAPI
  • incusd: Pass profile device cache to ToAPI calls when possible
  • incusd/db/instances: Support device cache to ToAPI
  • incusd: Pass instance device cache to ToAPI calls when possible
  • incusd/db/instances: Allow passing profile devices to instance ToAPI
  • incusd: Pass profile device cache to instance ToAPI calls when possible
  • incusd/instances: Remove old retry logic
  • incusd/network_integration: Fix typo in doc string
  • doc: Update configs
  • incusd/main_forknet: Tweak DHCP client to apply DNS first
  • incusd/network/ovn: Use stable random for IC gateway chassis priority
  • api: network_integrations_peer_name
  • incusd/network_integrations: Add peerName to ovn.transit.pattern
  • incusd/network/ovn: Expose peerName to ovn.transit.pattern
  • doc: Update configs
  • incus/cluster/list-tokens: Add support for column selection
  • i18n: Update translation templates
  • incusd/storage_volumes_state: Handle unsupported response from drivers
  • incusd/db/cluster: Remove network integration/peer unique index
  • incusd/db/cluster: Update schema
  • lxd-to-incus: Handle Incus socket in /run/incus/
  • incusd/network/ovn: Record transit subnets
  • incusd/network/ovn: Add transit switch addresss allocation functions
  • incusd/network/ovn: Setup transit switch allocations
  • incusd/auth/openfga: Avoid deprecated ApiSchema and ApiHost
  • incusd/auth: Re-organize entitlement list
  • incusd/auth/openfga: Sort entries in openfga model
  • incusd/auth/openfga: Add missing network integration permission
  • incusd/auth/openfga: Require admin level to create projects
  • incusd/auth/openfga: Rebuild model
  • incusd/auth: Fix network integration object
  • incus/config/trust/list-tokens: Add support for column selection
  • i18n: Update translation templates
  • incus/network/peer/list: Add support for column selection
  • i18n: Update translation templates
  • incus/network/load-balancer/list: Add support for column selection
  • i18n: Update translation templates
  • Translated using Weblate (Chinese (Simplified))
  • Change Cloud Init "user" to "users"
  • shared/api: Fix incorrect struct naming for volume backups
  • client: Update for fixed volume backup structs
  • incus: Update for fixed volume backup structs
  • incusd: Update for fixed volume backup structs
  • incusd/storage_volume_backup: Fix swagger references
  • incusd/storage_bucket_backup: Fix swagger references
  • doc/rest-api: Refresh swagger YAML
  • incusd/device/nic: Make burst rate dynamic for ingress traffic
  • incusd/storage/lvm: Allow live resize
  • incusd/storage/zfs: Allow online resize of ZFS block volumes
  • incusd/device/disk: Add callback on resize
  • incusd/instance/drivers/qmp: Add resize handling
  • incusd/instance/qemu: Add disk resize handling
  • incusd/node/config: Add network.ovs.connection
  • doc: Switch /var/run to just /run
  • incusd/cluster/config: Switch from /var/run to /run
  • incusd/instance/agent-loader: Don't hardcode path
  • incusd/syslog: Update OVS path
  • doc: Update configs
  • incusd/network/ovs: Make OVS database configurable
  • incusd/state: Add OVS function
  • incusd: Set OVS function on State
  • incusd: Port to state.OVS
  • incusd: Reset OVS as needed
  • incusd/network/ovn: Limit MAC_Binding explosion
  • incusd/network/ovn: Add ARP limits to updated routers
  • incusd/network/ovn: Wait a bit longer for northd to allocate addresses
  • incusd/apparmor: Don't constantly query the version and cache
  • incusd/storage/driver/dir: Don't needlessly re-apply project id on quota changes
  • incusd/storage/quota: Don't fail on missing paths
  • incusd/storage/lvm: Retry setactivation skip for busy environments
  • api: qemu_scriptlet
  • incusd/instance: Add qemu scriptlet config options
  • incusd: Move QEMU default values to a subpackage
  • incusd/scriptlet: Move the logger definition
  • incusd/scriptlet: Add helper functions
  • incusd/scriptlet: Add Unmarshal function
  • incusd/scriptlet: Add qemu scriptlet
  • incusd/project: Update low-level properties
  • doc: Update metadata
  • incusd/scriptlet: Remove deprecated starlark.SourceProgram
  • Makefile: Switch minimum Go to 1.22
  • gomod: Update dependencies
  • doc: Update requirements
  • incusd/instance/drivers/qemu: Fix node name overflow logic
  • incusd/instance/drivers/qemu: Add missing node name handling
  • incusd/api_internal: Add API to notify volume resizes
  • incusd/cluster: Fix redirect loop with shared volumes across multiple servers
  • incusd/storage/backend: Notify instances following block custom volume resize
  • api: instance_auto_restart
  • incusd/instance: Add boot.autorestart
  • doc: Update metadata
  • incusd/instance/drivers: Implement shouldAutoRestart
  • incusd/instance/drivers/lxc: Implement boot.autorestart
  • incusd/instance/drivers/qemu: Implement boot.autorestart
  • tests: Validate autorestart logic
  • client: Fix error handling in push mode copy
  • incusd/network/ovn: Fix send_periodic syntax
  • incusd/project: Validate group names
  • incusd/db: Confirm cluster group validity during placement
  • doc/cluster_group: Mention renaming groups
  • api: storage_lvm_metadatasize
  • doc/storage_lvm: Add lvm.metadata_size
  • incusd/storage/lvm: Add lvm.metadata_size
  • incusd/storage/zfs: Only attempt to load the module if the tools exist
  • incusd/instance/edk2: Add Void Linux x86_64 paths
  • incusd/profiles: Empty default profile on forced deletion
  • Revert "incusd/instance/agent-loader: Don't hardcode path"
  • incusd/device: Add new Register function
  • incusd/instance/drivers: Use Register function
  • incusd/device: Don't make Register depend on validate
  • incusd/storage/drivers: Add isDeleted flag
  • incusd/storage/drivers/ceph: Rework parseClone
  • incusd/storage/drivers/ceph: Rework parseParent
  • incusd/storage/drivers/ceph: Make use of isDeleted flag
  • incusd/instance/qemu: Allow setCPUs to re-use QMP
  • incusd/instance/qmp: Handle QMP occasionally returning multiple responses
  • incusd/seccomp: Update syscall numbers
  • incusd/instance/drivers/qemu: Double number of hotplug slots
  • incusd/instance/qemu: Rework PCI hotplug
  • incusd/instance/drivers/edk2: Limit calls to GetenvEdk2Path
  • incusd/instance/drivers/edk2: Actually check that the files exist
  • incusd/device/config: Fix comment
  • api: ovn_nic_promiscuous
  • doc/devices/nic_ovn: Add security.promiscuous
  • incusd/network/ovn: Only set DHCP options on LSP when not setting up a router interface
  • incusd/network/ovn: Add support for promiscuous Logical Switch Port
  • incusd/network/ovn: Wire in security.promiscuous
  • incusd/device/nic: Add security.promiscuous
  • api: ovn_nic_ip_address_none
  • doc/devices/nic_ovn: Add none for ipv4.address/ipv6.address
  • incusd/device/nic_ovn: Allow 'none' as value for ipv4.address/ipv6.address
  • incusd/network/ovn: Add support for disabling allocation on LSP
  • incusd/network/ovn: Wire in support for ipvX.address=none
  • incusd/network/ovn: Fix BGP advertisement of load balancers
  • incus-user: Handle deleted projects
  • Makefile: Set minimum Go to 1.22.0
  • Makefile: Remove deprecated flag
  • gomod: Update dependencies
  • incusd/auth: Update for openfga-go-sdk API breakage

Documentation

The Incus documentation can be found at:
https://linuxcontainers.org/incus/docs/main/

Packages

There are no official Incus packages as Incus upstream only releases regular release tarballs. Below are some available options to get Incus up and running.

Installing the Incus server on Linux

Incus is available for most common Linux distributions. You'll find detailed installation instructions in our documentation.

https://linuxcontainers.org/incus/docs/main/installing/

Homebrew package for the Incus client

The client tool is available through HomeBrew for both Linux and MacOS.

https://formulae.brew.sh/formula/incus

Chocolatey package for the Incus client

The client tool is available through Chocolatey for Windows users.

https://community.chocolatey.org/packages/incus/6.5.0

Winget package for the Incus client

The client tool is also available through Winget for Windows users.

https://winstall.app/apps/LinuxContainers.Incus

Support

Monthly feature releases are only supported up until the next release comes out. Users needing a longer support length and less frequent changes should consider using Incus 6.0 LTS instead.

Community support is provided at: https://discuss.linuxcontainers.org
Commercial support is available through: https://zabbly.com/incus
Bugs can be reported at: https://github.com/lxc/incus/issues

Contents