Go to Table of Contents

回到概览

Incus 6.4 has been released

2024年8月9日

Introduction

The Incus team is pleased to announce the release of Incus 6.4!

This is a very balanced release with something new for everyone!

It comes with a number of bugfixes and new features to help with the OCI support added in the previous release. It also brings in a number of new features for more complex shared/cluster environments. And it's jam packed with bugfixes, fixing a lot of annoyances around storage, clustering, OpenID authentication, auditing and more.

image|690x309

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

Enjoy!

New features

Cluster group configuration

Cluster groups now have a standard configuration table like most other Incus objects.

This means the usual set of commands and APIs:
- incus cluster group edit
- incus cluster group get
- incus cluster group set
- incus cluster group show
- incus cluster group unset

Per-cluster group CPU baseline and flags for VMs

Building on top of that support for cluster group configuration, we now have support for defining the VM CPU baseline on a per cluster group basis.

This makes it possible to have one cluster group per CPU model/generation and have Incus compute the common set of CPU flags for those servers.

For example, incus cluster group set foo instances.vm.cpu.x86_64.baseline=kvm64 instances.vm.cpu.x86_64.flags=auto will have Incus automatically go through the servers in the foo cluster group and then fill in the flags configuration key with the set of common CPU flags.

But this also allows setting up your own completely custom CPU defintion, for example, incus cluster group set foo instances.vm.cpu.x86_64.baseline=EPYCv2 instances.vm.cpu.x86_64.flags=-svm will expose a basic 2nd generation AMD EPYC CPU with the virtualization extension (svm) disabled.

Using a sub-path of a volume as a disk

It's now possible to use a path within an existing custom volume as the source for a disk entry.

stgraber@castiana:~$ incus launch images:ubuntu/24.04 demo
Launching demo
stgraber@castiana:~$ incus launch images:ubuntu/24.04 demo-sub
Launching demo-sub
stgraber@castiana:~$ incus storage volume create default demovol
Storage volume demovol created
stgraber@castiana:~$ incus config device add demo demovol disk pool=default source=demovol path=/mnt/demovol
Device demovol added to demo
stgraber@castiana:~$ incus exec demo bash
root@demo:~# mkdir -p /mnt/demovol/sub/path/
root@demo:~# echo world > /mnt/demovol/sub/path/hello
root@demo:~#·
exit
stgraber@castiana:~$ incus config device add demo-sub demovol disk pool=default source=demovol/sub/path path=/mnt/demovol
Device demovol added to demo-sub
stgraber@castiana:~$ incus exec demo-sub bash
root@demo-sub:~# cat /mnt/demovol/hello·
world

In this example, a demovol custom volume is created, then attached to the demo container, a sub-directory is created in that volume and that sub-directory is then attached to another container, demo-sub.

Per storage pool projects limits

Incus projects can have resource limits applied to them, ideal when providing access to a project to a third party. Up until now, it was possible to limit the total disk usage within a project, but that would apply to all storage pools.

As it's common to have different storage pools representing different storage characteristics (local vs remote) or class (ssd vs hdd), it's useful to have a way to provide limits per storage pool.

To do so, a new configuration key, limits.disk.pool.POOLNAME is now available in project configuration. Setting the limit to 0 fully disables that storage pool and causes it to disappear from the storage pool listing in that project.

stgraber@dakara:~$ incus project info test-limits
+------------------+-----------+-------+
|     RESOURCE     |   LIMIT   | USAGE |
+------------------+-----------+-------+
| CONTAINERS       | UNLIMITED | 0     |
+------------------+-----------+-------+
| CPU              | UNLIMITED | 0     |
+------------------+-----------+-------+
| DISK             | UNLIMITED | 0B    |
+------------------+-----------+-------+
| INSTANCES        | UNLIMITED | 0     |
+------------------+-----------+-------+
| MEMORY           | UNLIMITED | 0B    |
+------------------+-----------+-------+
| NETWORKS         | UNLIMITED | 0     |
+------------------+-----------+-------+
| PROCESSES        | UNLIMITED | 0     |
+------------------+-----------+-------+
| VIRTUAL-MACHINES | UNLIMITED | 0     |
+------------------+-----------+-------+
stgraber@dakara:~$ incus storage list
+---------+--------+-------------+---------+---------+
|  NAME   | DRIVER | DESCRIPTION | USED BY |  STATE  |
+---------+--------+-------------+---------+---------+
| default | zfs    |             | 45      | CREATED |
+---------+--------+-------------+---------+---------+
| foo     | dir    |             | 0       | CREATED |
+---------+--------+-------------+---------+---------+
stgraber@dakara:~$ incus project set test-limits limits.disk.pool.foo=0 limits.disk.pool.default=5GiB limits.disk=10GiB
stgraber@dakara:~$ incus project info test-limits
+------------------+-----------+-------+
|     RESOURCE     |   LIMIT   | USAGE |
+------------------+-----------+-------+
| CONTAINERS       | UNLIMITED | 0     |
+------------------+-----------+-------+
| CPU              | UNLIMITED | 0     |
+------------------+-----------+-------+
| DISK             | 10.00GiB  | 0B    |
+------------------+-----------+-------+
| DISK (DEFAULT)   | 5.00GiB   | 0B    |
+------------------+-----------+-------+
| INSTANCES        | UNLIMITED | 0     |
+------------------+-----------+-------+
| MEMORY           | UNLIMITED | 0B    |
+------------------+-----------+-------+
| NETWORKS         | UNLIMITED | 0     |
+------------------+-----------+-------+
| PROCESSES        | UNLIMITED | 0     |
+------------------+-----------+-------+
| VIRTUAL-MACHINES | UNLIMITED | 0     |
+------------------+-----------+-------+
stgraber@dakara:~$ incus storage list
+---------+--------+-------------+---------+---------+
|  NAME   | DRIVER | DESCRIPTION | USED BY |  STATE  |
+---------+--------+-------------+---------+---------+
| default | zfs    |             | 45      | CREATED |
+---------+--------+-------------+---------+---------+
stgraber@dakara:~$ incus create images:ubuntu/24.04 c1 --storage default -d root,size=5GiB
Creating c1

The instance you are starting doesn't have any network attached to it.
  To create a new network, use: incus network create
  To attach a network to an instance, use: incus network attach

stgraber@dakara:~$ incus create images:ubuntu/24.04 c2 --storage default -d root,size=5GiB
Creating c2
Error: Failed instance creation: Failed checking if instance creation allowed: Reached maximum aggregate value "5GiB" for "limits.disk.pool.default" in project "test-limits"
stgraber@dakara:~$ incus project set test-limits limits.disk.pool.foo=5GiB
stgraber@dakara:~$ incus create images:ubuntu/24.04 c2 --storage foo -d root,size=5GiB
Creating c2

The instance you are starting doesn't have any network attached to it.
  To create a new network, use: incus network create
  To attach a network to an instance, use: incus network attach

stgraber@dakara:~$ incus project info test-limits
+------------------+-----------+----------+
|     RESOURCE     |   LIMIT   |  USAGE   |
+------------------+-----------+----------+
| CONTAINERS       | UNLIMITED | 2        |
+------------------+-----------+----------+
| CPU              | UNLIMITED | 0        |
+------------------+-----------+----------+
| DISK             | 10.00GiB  | 10.00GiB |
+------------------+-----------+----------+
| DISK (DEFAULT)   | 5.00GiB   | 5.00GiB  |
+------------------+-----------+----------+
| DISK (FOO)       | 5.00GiB   | 5.00GiB  |
+------------------+-----------+----------+
| INSTANCES        | UNLIMITED | 2        |
+------------------+-----------+----------+
| MEMORY           | UNLIMITED | 0B       |
+------------------+-----------+----------+
| NETWORKS         | UNLIMITED | 0        |
+------------------+-----------+----------+
| PROCESSES        | UNLIMITED | 0        |
+------------------+-----------+----------+
| VIRTUAL-MACHINES | UNLIMITED | 0        |
+------------------+-----------+----------+
stgraber@dakara:~$

Here we can see a project get set up with a disk limit, first hiding one of the pools, then filling the other before setting a limit on the previously hidden pool.

Up until now, all OVN networks have had a uplink network set (network property).
That's the network on which the external facing router port will sit and through which all ingress/egress into/out-of the OVN network will happen.

Incus picks an IPv4 (and/or IPv6) address on that uplink network and then uses it to route all the traffic out of the virtual network and onto the physical network.

Now a special value of none for that network property will instruct Incus to create an OVN network which is not connected to any uplink and is threfore fully isolated.

root@server01:~# incus network create ovn-isolated network=none --type=ovn
Network ovn-isolated created
root@server01:~# incus launch images:ubuntu/24.04 c1 --network ovn-isolated
Launching c1
root@server01:~# incus list
+------+---------+--------------------+-----------------------------------------------+-----------+-----------+----------+
| NAME |  STATE  |        IPV4        |                     IPV6                      |   TYPE    | SNAPSHOTS | LOCATION |
+------+---------+--------------------+-----------------------------------------------+-----------+-----------+----------+
| c1   | RUNNING | 10.248.34.2 (eth0) | fd42:669c:8431:b3cc:216:3eff:fef3:fdb2 (eth0) | CONTAINER | 0         | server01 |
+------+---------+--------------------+-----------------------------------------------+-----------+-----------+----------+
root@server01:~# incus exec c1 bash
root@c1:~# ping 1.1.1.1
PING 1.1.1.1 (1.1.1.1) 56(84) bytes of data.
^C
--- 1.1.1.1 ping statistics ---
2 packets transmitted, 0 received, 100% packet loss, time 1031ms

Here we can see an isolated network being created and a container being placed on it.
The network provides IPv4 and IPv6 addresses as usual, but no traffic can come out.

Per-instance LXCFS

A new server configuration key, instances.lxcfs.per_instance can now be enabled to have Incus start a dedicated LXCFS instance for every container.

This is in contrast to the default of having a single LXCFS instance run for the entire system.

Enabling this comes at a slightly higher resource usage per container, but reduces the risk of one container flooding the shared LXCFS instance as well as make it so a LXCFS crash only affects one container.

stgraber@castiana:~$ pgrep -a lxcfs
1101 /opt/incus/bin/lxcfs /var/lib/incus-lxcfs
stgraber@castiana:~$ incus config set instances.lxcfs.per_instance=true
stgraber@castiana:~$ incus restart demo
stgraber@castiana:~$ pgrep -a lxcfs
1101 /opt/incus/bin/lxcfs /var/lib/incus-lxcfs
962122 lxcfs -f -p /run/incus/demo/lxcfs.pid --runtime-dir /run/incus/demo/lxcfs /var/lib/incus/devices/demo/lxcfs

Support for environment file at create/launch time

To make it easier to run OCI containers, it's now possible to specify environment variables through an environment variable file which gets read at creation time and converted to Incus configuration options.

stgraber@castiana:~$ cat mysql.env
MYSQL_DATABASE=wordpress
MYSQL_USER=wordpress
MYSQL_PASSWORD=wordpress
MYSQL_RANDOM_ROOT_PASSWORD=1

stgraber@castiana:~$ incus launch docker:mysql mysql --environment-file mysql.env 
Launching mysql

stgraber@castiana:~$ incus config show mysql
architecture: x86_64
config:
  environment.GOSU_VERSION: "1.17"
  environment.HOME: /root
  environment.MYSQL_DATABASE: wordpress
  environment.MYSQL_MAJOR: innovation
  environment.MYSQL_PASSWORD: wordpress
  environment.MYSQL_RANDOM_ROOT_PASSWORD: "1"
  environment.MYSQL_SHELL_VERSION: 9.0.1-1.el9
  environment.MYSQL_USER: wordpress
  environment.MYSQL_VERSION: 9.0.1-1.el9
  environment.PATH: /usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
  environment.TERM: xterm
  image.architecture: x86_64
  image.description: docker.io/library/mysql (OCI)
  image.type: oci
  volatile.base_image: d8df069848906979fd7511db00dc22efeb0a33a990d87c3c6d3fcdafd6fc6123
  volatile.cloud-init.instance-id: f12e3ddb-ac93-4942-b3e1-dcd560893140
  volatile.container.oci: "true"
  volatile.eth0.host_name: vethac8631aa
  volatile.eth0.hwaddr: 00:16:3e:20:32:87
  volatile.idmap.base: "0"
  volatile.idmap.current: '[{"Isuid":true,"Isgid":false,"Hostid":1000000,"Nsid":0,"Maprange":1000000000},{"Isuid":false,"Isgid":true,"Hostid":1000000,"Nsid":0,"Maprange":1000000000}]'
  volatile.idmap.next: '[{"Isuid":true,"Isgid":false,"Hostid":1000000,"Nsid":0,"Maprange":1000000000},{"Isuid":false,"Isgid":true,"Hostid":1000000,"Nsid":0,"Maprange":1000000000}]'
  volatile.last_state.idmap: '[]'
  volatile.last_state.power: RUNNING
  volatile.uuid: 5ed7f63b-5b6c-4c89-9dfa-117c2b785370
  volatile.uuid.generation: 5ed7f63b-5b6c-4c89-9dfa-117c2b785370
devices: {}
ephemeral: false
profiles:
- default
stateful: false
description: ""

In this example, we see a mysql container created from an OCI image and using environment variables defined in mysql.env.

Complete changelog

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

Full commit list
  • incus-simplestreams: Fix split images
  • [lxd-import] github: stop purging core20
  • [lxd-import] github: purge disabled/superseded snaps
  • [lxd-import] github: move snap removal to "Reclaim some space" step
  • [lxd-import] github: put docker removal its own step
  • [lxd-import] github: mask lxc{,-net}.service in a singe command
  • [lxd-import] lxd/instance/drivers/driver/qemu: Don't leak file descriptor when probing for Direct I/O support
  • [lxd-import] lxd/network/acl: Change protocol field for ovn logs
  • [lxd-import] lxd/instance: Reject limits.kernel config for VMs
  • [lxd-import] doc: limits.kernel only applies to containers (see #12874)
  • [lxd-import] lxd/storage: Fix resize for pools with custom zfs.pool_name
  • [lxd-import] lxd/storage/drivers/driver_zfs_utils: fix typos
  • [lxd-import] lxd/storage/drivers/driver_zfs_utils: make it explicit that blocksize is in bytes
  • [lxd-import] lxd/task/group: Make cancel type of context.CancelFunc for clarity
  • [lxd-import] doc: Add paragraph on how to delete images
  • [lxd-import] test: Add exec exit code test
  • [lxd-import] lxd/apparmor: allow confined services to receive required signals
  • [lxd-import] lxd/rsync: Consistently compare files on nanosecond basis
  • [lxd-import] test/suites/migration: Check for file contents after refresh
  • [lxd-import] test/suites/migration: Check local and remote instance refreshes based on nanoseconds
  • [lxd-import] doc: add paragraph on how to delete images
  • [lxd-import] doc: enable multiprocessing for pyspelling
  • [lxd-import] Makefile: have run-parts report which script it runs
  • [lxd-import] lxd/storage/drivers/ceph: Disable filesystem config keys on block volumes
  • [lxd-import] lxd/storage/drivers/lvm: Disable filesystem config keys on block volumes
  • [lxd-import] test: Add check to restore custom volumes of type block
  • [lxd-import] lxd/storage/drivers/ceph: Update UnmountVolumeSnapshot docstring
  • [lxd-import] lxd: Improve error check for existing certificates
  • [lxd-import] shared/api: Updates swagger description for certificate field.
  • [lxd-import] shared/api: Fix lint errors (receiver-naming).
  • [lxd-import] lxd/db/cluster: Fix lint error (revive: var-naming).
  • [lxd-import] lxd-migrate: Ignore lint error (revive: deep-exit).
  • [lxd-import] lxc/remote: Fix lint errors (revive: exported).
  • [lxd-import] lxd/storage/backend: Don't validate custom storage volumes twice
  • [lxd-import] lxd-generate: Return helpful error instead of panicking.
  • [lxd-import] lxd/storage/backend: Use quotes consistently for error messages
  • [lxd-import] lxd/project: Don't panic on StorageVolumeParts
  • [lxd-import] github: don't abort on remount failures
  • [lxd-import] test/main: add log grouping (GHA)
  • [lxd-import] test/main: show dmesg on failure
  • [lxd-import] lxd/api/internal: Use correct quoting for error in internalImportFromBackup
  • [lxd-import] lxd/db/cluster/devices: Use correct string quoting of device type for error in NewDeviceType
  • [lxd-import] lxd/instances/post: Improve error in createFromBackup
  • [lxd-import] lxd/storage/backend/lxd: Update backup.yaml after instance and volume DB records have been generated in CreateInstanceFromBackup
  • [lxd-import] test/main: don't wrap tests logs in log groups
  • [lxd-import] lxd/device/disk: Remove config.iso file when the cloud-init:config disk device is removed
  • [lxd-import] lxd/images: Add project to error in autoSyncImages
  • [lxd-import] lxd/project/project: Remove optimisation from StorageVolumeProject
  • [lxd-import] lxd/storage/volumes: Remove unnecessary 2 line variable definition in doCustomVolumeRefresh
  • [lxd-import] lxd/storage/volumes: Remove unnecessary 2 line variable definition in doVolumeCreateOrCopy
  • [lxd-import] lxd/storage/volumes: Validate source project in doCustomVolumeRefresh
  • [lxd-import] [lxd-import] lxd/storage/volumes: Validate source project in doVolumeCreateOrCopy
  • [lxd-import] lxd/db/cluster: Remove redunant parentheses.
  • [lxd-import] lxd/db/cluster: Rename "node" to "cluster member".
  • [lxd-import] lxd/migrate/storage/volumes: Use volume name from DB in migrationSourceWs.DoStorage
  • [lxd-import] lxd/network/network/utils: Fix incorrect conversion from int64 to int in inRoutingTable
  • [lxd-import] lxd/network/network/utils: Remove unnecessary call to fmt.Sprintf by passing base to ParseInt
  • [lxd-import] lxd/response: Use SmartError if SyncResponse success=false
  • [lxd-import] lxd/storage/drivers/driver/zfs/volumes: Fix error scoping in CreateVolumeFromCopy
  • [lxd-import] lxd/storage/drivers/driver/zfs/volumes: Define send/receive channels together in CreateVolumeFromCopy
  • [lxd-import] lxd/storage/drivers/driver/zfs/volumes: Kill sender if receiver fails in CreateVolumeFromCopy
  • [lxd-import] lxd/storage/drivers/driver/zfs/volumes: Handle multi-line errors in CreateVolumeFromCopy
  • [lxd-import] lxd/storage/drivers/driver/btrfs/volumes: Fix hang when btrfs receive fails in RefreshVolume
  • [lxd-import] lxd/storage/drivers/driver/btrfs/volumes: Handle multi-line errors in RefreshVolume
  • [lxd-import] lxd/storage/drivers/driver/zfs/volumes: Aligns RefreshVolume with BTRFS driver
  • [lxd-import] lxd/response: Fallback to error response after setting headers
  • [lxd-import] lxd/storage/backend/lxd: Use volume name from VolumeDBGet in BackupCustomVolume
  • [lxd-import] lxd/storage/drivers/btrfs: Clarify fallback in case UUID discovery times out
  • incusd/db/cluster: Fix import shadowing
  • doc/rest-api: Refresh swagger YAML
  • [lxd-import] lxd/db/cluster: Update error messages in unit tests.
  • [lxd-import] lxd/storage/drivers: Add volume param to roundVolumeBlockSizeBytes
  • [lxd-import] lxd/storage/drivers/zfs: Round to zfs.blocksize or 16KiB
  • [lxd-import] test/storage: Add non-power-of-two sized storage check
  • [lxd-import] test/storage: Remove zfs rounding test
  • [lxd-import] lxd/storage/drivers: Refactor volume size rounding logic
  • [lxd-import] lxd/storage/drivers/zfs: Wait for device to appear when activating a volume
  • [lxd-import] lxd/storage/drivers/zfs: Check for non /dev/zvol/* paths
  • incusd/storage/zfs: Fix import shadowing
  • doc/storage_backup: Fix command example
  • doc/requirements: Add OCI requirements
  • doc/images: Mention OCI images
  • incusd/instance/edk2: Support OVMF filenames on arm64
  • incusd/instance/drivers/qemu: Limit CPU flag calculation to x86_64
  • incus/s3: Fix mcli minio client executable name check
  • incusd/instance/qemu: Fix architecture check being backward
  • incusd/instance: Fix initial configuration handling for OCI
  • client/oci: Add errors for missing skopeo
  • client/oci: Add errors for missing umoci
  • Change RunDir file mode to 0711
  • incusd/instance/lxc: Handle OCI containers using real PID1
  • incusd/apparmor/qemu: Relax apparmor rules a bit
  • incus-simplestreams: Handle removal of combined images
  • incusd/instance/lxc: Add basic Mounts handling for OCI
  • incusd/apparmor/qemu: Fix typo in rule
  • incusd/apparmor/dnsmasq: Relax rules a bit
  • incusd/db/node: Fix version check in GetAPI
  • incusd/db: Allow cluster startup with differing API extensions
  • incusd: Extend heartbeat data for minimum API extension count
  • incusd/storage: Use writeback mode for qemu-img convert
  • incusd/storage: Improve unpacking message
  • incusd/operations: Handle percentage only updates
  • incusd/storage: Pass through tracker to qemu-img
  • incusd/db/cluster: Update tests for relaxed API extensions checks
  • incusd/apparmor: Implement progress tracker for qemu-img
  • incusd/cgroup: Handle unknown devices in io.stat
  • incusd/instance_post: Always set the target project
  • incusd/storage/drivers: Consistently set VolumeMultiNode
  • incusd/storage/lvm: Require an exclusive lock during snapshot
  • incusd/storage/lvm: Properly handle activation during resize
  • incusd/storage: Properly detect filesystem on remote block
  • incusd/cluster: Always attempt to forward custom volume requests
  • client: Always set GetBody
  • client: Report OIDC refresh failure
  • incus/remote: Forward OIDC auth failures
  • Make run-parts running compatible with different versions
  • client: Retry when told to by the proxy
  • Use qemu-img convert output to update progress
  • incusd/cluster: Add flagFormat
  • internal/cmd: Add support for format options
  • incusd/isntance/edk2: Move seabios to /usr/share/qemu
  • incusd/isntance/edk2: Add ArchLinux x86_64 paths
  • tests: Use future values in property test
  • incusd/db/cluster: Cleanup indentation
  • incusd/db/cluster: Update schema
  • incusd/db/generate: Add exception for cluster tables
  • api: clustering_groups_config
  • incusd/db/cluster: Add cluster_groups_config
  • incusd/db/cluster: Update schema
  • shared/api: Add Config to ClusterGroupPut
  • doc/rest-api: Refresh swagger YAML
  • incusd/db/cluster: Update functions for new cluster group config table
  • incusd/cluster: Add cluster group config logic
  • doc: Update configs
  • incus/completion: Add cmpClusterGroupConfigs
  • incus/cluster/group: Add get/set/unset
  • i18n: Update translation templates
  • doc: Add cluster group config
  • tests: Add cluster group config test
  • incusd/resources: Add sortedMapKeys
  • incusd/resources/cpu: Sort sockets, cores and threads
  • incusd/auth: Fill missing local volume location
  • incusd/cluster: Correctly record volatile.cluster.group on move
  • incusd/migration: Show source errors first
  • incusd/instance/qemu: Clarify live migration error
  • incusd/cluster: Attempt to ping the server prior to healing
  • incusd/instance/qemu: Fix bad timeout errors
  • api: instances_lxcfs_per_instance
  • incusd/instance/qemu: Send two ACPI events on shutdown
  • incusd/instance: Add progress tracking to export
  • incusd/server/config: Add instances.lxcfs.per_instance
  • incusd/instance/lxc: Add support for per-instance LXCFS
  • doc: Update configs
  • client: Remove GetBody when can't seek back
  • client: Add Seek call to GetBody
  • Translated using Weblate (Chinese (Simplified))
  • incusd: Simplify image replication
  • incusd/images: Only remove from authorizer once
  • incusd/images: Correctly record new aliases
  • incusd/images: Correctly remove aliases
  • incusd/images: Set authorizer entries at the correct time
  • incusd/images: Don't alter image info on cluster copy
  • incusd/storage_pools: Set authorizer for pending pools
  • incusd/networks: Set authorizer for pending networks
  • incusd/network: Handle long interface names
  • incus/cluster: Cleanup certificate update message
  • i18n: Update translation templates
  • incus/top: Properly filter disk usage
  • incusd/state: Add Cluster gateway
  • incusd/daemon: Sort state fields
  • incusd/daemon: Add cluster gateway to state
  • incusd/acme: Update for state change
  • incusd/images: Update for state change
  • incusd/operations: Update for state change
  • incusd/instances: Update for state change
  • incusd/patches: Update for state change
  • incusd/cluster: Update for state change
  • incusd/instances: Use targetGroupPrefix
  • incusd/cluster: Split files
  • incusd/cluster: Fix import shadowing
  • incusd/cluster: Re-factor healing logic
  • incusd/cluster: Improve logging
  • incusd/cluster: Reduce dqlite logging
  • incusd/cluster: Extend HasConnectivity to do API checking
  • incusd/cluster: Add HasConnectivity check to event handling
  • incusd: Don't block on startup for cluster queries
  • incusd/cluster: Rework offline server detection
  • incusd/cluster/evacuation: Add separate healing mode
  • incusd/cluster/healing: Try up to 5 times
  • incusd/cluster/healing: Add logging
  • incusd/device/disk: Fix bad CanMigrate logic
  • incusd/storage/lvm: Properly activate/deactivate ISOs
  • incusd/images: Expose image type as header
  • client: Skip image hash if OCI
  • incusd/instance/lxc: Add /init to OCI init paths
  • incusd/device: Don't fail on LXCFS device entry
  • client: Report source errors too on copy
  • incusd/storage: Have roundVolumeBlockSizeBytes return an error
  • incusd/instance_post: Fix cross-server live-migration
  • incus/image: Correct image copy logic
  • incusd/storage/lvm: Hardden common functions
  • incusd/api: Don't panic on missing config
  • incusd/storage: Add Deactivate flag
  • incusd/storage/lvm: Add deactivation step for clusters
  • incusd/cluster: Return clear status for servers currently starting up
  • incusd/instance/lxc: Reduce logging level
  • incusd/instance/qemu: Reduce logging level
  • incusd/migrate: Reduce logging
  • incusd/storage: Reduce logging
  • incusd/instance/qemu: Remove double lifecycle event
  • tests/clustering: Use correct target project argument
  • incusd/isntance/edk2: Fix CSM handling
  • incusd/storage/zfs: Always call tryGetVolumeDiskPathFromDataset
  • incusd/network/ovn: Require functional uplink
  • doc: add colima instructions
  • doc: Update incus_alias.md
  • incus/network/load_balancer: Fix example
  • i18n: Update translation templates
  • incusd/network/ovn: Fix crash on uninitialized external IDs
  • doc/instances: Add VM agent install instructions
  • shared/api: Add Config to ServerUntrusted
  • doc/rest-api: Refresh swagger YAML
  • incusd/api_1.0: Expose user.ui config keys to all clients
  • doc/server: Mention user.ui config keys
  • incusd/auth/oidc: Better handle logout
  • incusd/networks: Emit lifecycle event and authz entries for OVN networks
  • incusd/network/ovn: Fix indent
  • doc/storage_volume: Fix snapshot command
  • shared/api: Add EventLifecycleInstanceMigrated
  • shared/api: Sort lifecycle events
  • incusd/lifecycle: Add InstanceMigrated
  • incusd/lifecycle: Sort lifecycle events
  • incusd/isntance/operationlock: Add ActionMigrate
  • incusd/instance/common: Add support for migration operation
  • incusd/instance/qemu: Add support for migration operation and lifecycle
  • incusd/instance/lxc: Add support for migration operation and lifecycle
  • shared/api: Add lifecycle events for cluster evacuation and healing
  • incusd/lifecycle: Add lifecycle events for cluster evacuation and healing
  • incusd/cluster: Add lifecycle events for evacuation
  • incusd/request: Strip port from event address
  • incusd/instance: Properly link instance and operation
  • incusd/operations: Add CopyRequestor for nested operations
  • incusd/instance: Track operation during exec/console
  • doc/clustering: Better document healing
  • incusd/instance: Track operation during creation
  • incusd/instance: Track operation during deletion
  • incusd/instance: Keep track of API operations
  • incusd/instance: Set operations on snapshot
  • incus-migrate: Properly handle projects
  • incusd/apparmor: Allow mounting zfs when delegation is supported
  • doc/clustering: Add howto on cluster access
  • cmd/incusd: Set keep-alive timeout
  • incusd/auth/oidc: Handle cases where we can't set cookies
  • incusd/instance/qemu: Deref ceph config path
  • incusd/apparmor/qemu: Guess ceph config paths
  • Translated using Weblate (Chinese (Simplified))
  • Translated using Weblate (Chinese (Simplified))
  • Translated using Weblate (Chinese (Simplified))
  • incusd/instance/lxc: Respect LXCFS_OPTS
  • api: clustering_groups_vm_cpu_definition
  • incusd/cluster_groups: Add CPU definition keys
  • doc: Update configs
  • incusd/instance/qemu: Use cluster group config
  • incusd/instance/drivers: Extract GetClusterCPUFlags
  • incusd/cluster/group: Support for auto CPU flags
  • incusd: Switch OVN to a getter function
  • incusd/network: Port to new OVN state function
  • incus: Add support for environment file (.env)
  • i18n: Update translation templates
  • incusd/storage/lvm: Re-try activation/deactivation
  • incusd/storage/lvm: Don't activate volumes during cold migration
  • Translated using Weblate (Chinese (Simplified))
  • Translated using Weblate (Chinese (Simplified))
  • shared/cliconfig: Add CacheDir
  • incus: Configure a cache directory
  • api: disk_volume_subpath
  • incusd/device/disk: Allow relative paths within custom volumes
  • doc/devices_disk: Mention sub-paths
  • tests: Test volume subpaths
  • api: projects_limits_disk_pool
  • incusd/projects: Add new limits.disk.pool config key
  • doc: Update configs
  • incus/project: Handle pool disk limits
  • incusd/project: Add per-pool disk limits
  • incusd/project: Add HiddenStoragePools
  • incusd/storage: Hide pools with a zero limit
  • tests: Add test for per pool limits
  • incus/image/alias: Add support for column selection
  • i18n: Update translation templates
  • incusd/main_forknet: Make it so our DHCP client never fails
  • Translated using Weblate (Chinese (Simplified))
  • api: network_ovn_isolated
  • incusd/network/ovn: Harden deletion logic
  • doc/network/ovn: Cover isolated networks
  • incusd/networks: Reserve "none" for uplinks
  • incusd/network/ovn: Allow creating isolated OVN networks (no uplink)
  • incusd/device/nic_ovn: Handle networks without uplinks
  • gomod: Update dependencies
  • Release Incus 6.4

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/6.4.0

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