Correct, it should have been:
profiles:
- default
I have a similar issue:
$ lxc launch images:ubuntu/focal test
Creating test
Error: Failed instance creation: Failed creating instance record: Failed initialising instance: Invalid devices: Failed detecting root disk device: No root device could be found
I have these images installed:
lxc image list
+-------+--------------+--------+---------------------------------------------+--------------+-----------+----------+-----------------------------+
| ALIAS | FINGERPRINT | PUBLIC | DESCRIPTION | ARCHITECTURE | TYPE | SIZE | UPLOAD DATE |
+-------+--------------+--------+---------------------------------------------+--------------+-----------+----------+-----------------------------+
| | 15c63eb43b36 | no | Ubuntu focal amd64 (20210204_07:42) | x86_64 | CONTAINER | 100.44MB | Feb 4, 2021 at 3:23pm (UTC) |
+-------+--------------+--------+---------------------------------------------+--------------+-----------+----------+-----------------------------+
| | d1df9c150a9f | no | ubuntu 20.04 LTS amd64 (release) (20210201) | x86_64 | CONTAINER | 358.84MB | Feb 4, 2021 at 3:19pm (UTC) |
+-------+--------------+--------+---------------------------------------------+--------------+-----------+----------+-----------------------------+
In my case the image has the default profile:
$ lxc image show 15c63eb43b36
auto_update: true
properties:
architecture: amd64
description: Ubuntu focal amd64 (20210204_07:42)
os: Ubuntu
release: focal
serial: "20210204_07:42"
type: squashfs
variant: default
public: false
expires_at: 1970-01-01T01:00:00+01:00
profiles:
- default
What could be happening?
That suggests your default profile doesn’t have a root device.
Thanks! Because it was a new installation, I solved the issue reinstaling lxd and running lxd init again.
I installed lxd in another machine and I get the same error again. In this case, I have no root device in the default profile. So, how can I create a root device?
xc profile show default
config: {}
description: Default LXD profile
devices:
eth0:
name: eth0
nictype: bridged
parent: lxdbr0
type: nic
name: default
used_by: []
I also have a default pool storage:
lxc storage show default
config:
source: /data/pool-storages/default
volatile.initial_source: /data/pool-storages/default
description: ""
name: default
driver: btrfs
used_by: []
status: Created
locations:
- none
What is the command to create the root device in the default profile? I was trying to do this, but I got an error not found. I tried with the command below, but I guess I am doing something wrong.
lxc config device add default root disk path=/ pool=default
I am not sure if I am understanding what this command does. I guess that it creates a new device named root, in the profile default, and of type disk. The device will be mapped to the folder / in new containers, and a source for this device in each new container will be created in the default pool accordingly.
lxc profile device add default root disk path=/ pool=default
2 Likes
I think that the documentation is a bit confusing:
lxc profile device add --help
Description:
Add devices to containers or profiles
Usage:
lxc profile device add [<remote>:]<container|profile> <device> <type> [key=value...] [flags]
Examples:
lxc config device add [<remote>:]container1 <device-name> disk source=/share/c1 path=opt
Will mount the host's /share/c1 onto /opt in the container.
Global Flags:
--debug Show all debug messages
--force-local Force using the local unix socket
-h, --help Print help
-v, --verbose Show all information messages
--version Print version number
In the Usage section appears the keyword profile, but not in the Examples section.
You’re using an old LXD version:
stgraber@castiana:~/data/code/lxc/lxd (lxc/master)$ lxc profile device add --help
Description:
Add instance devices
Usage:
lxc profile device add [<remote>:]<profile> <device> <type> [key=value...] [flags]
Examples:
lxc profile device add [<remote>:]profile1 <device-name> disk source=/share/c1 path=opt
Will mount the host's /share/c1 onto /opt in the instance.
Global Flags:
--debug Show all debug messages
--force-local Force using the local unix socket
-h, --help Print help
--project string Override the source project
-q, --quiet Don't show progress information
-v, --verbose Show all information messages
--version Print version number
lxc image edit 34aefd9c3268
and setting
profiles:
- default
worked for me.
hi ,
i had the same problem and i delete the profile and copy default profile with new name and re launch the image and its work for me.
lxc profile copy default profilename
I have the same problem which appeared to occur after an lxc init failed due to the attached storage already having been created by zpool - and therefore unmanaged by lxc.
I deleted the zpool volume and recreated it using lxc init.
storage output:
root@code:/# lxc storage list
+------------+-------------+--------+------------+---------+
| NAME | DESCRIPTION | DRIVER | SOURCE | USED BY |
+------------+-------------+--------+------------+---------+
| containers | | zfs | containers | 0 |
+------------+-------------+--------+------------+---------+
profile output:
root@code:/# lxc profile show default
config: {}
description: Default LXD profile
devices:
eth0:
name: eth0
network: lxdbr0
type: nic
name: default
used_by: []
lxc storage show containers output:
root@code:/# lxc storage show containers
config:
source: containers
volatile.initial_source: /dev/disk/by-id/scsi-0DO_Volume_containers
zfs.pool_name: containers
description: ""
name: containers
driver: zfs
used_by: []
status: Created
locations:
- none
Added containers to the default profile:
root@code:/# lxc profile device add default root disk source=/mnt/containers
new profile output:
root@code:/# lxc profile show default
config: {}
description: Default LXD profile
devices:
eth0:
name: eth0
network: lxdbr0
type: nic
root:
source: /mnt/containers
type: disk
name: default
used_by: []
Attempted to create a new container:
root@code:/# lxc launch ubuntu:20.04 vscode
Creating vscode
Error: Failed instance creation: Failed creating instance record: Failed initialising instance: Invalid devices: Failed detecting root disk device: No root device could be found
I can’t express how frustrating it is to try and fix what should be dead simple, as in something as basic as:
lxc profile add default storage containers
LXD already knows about the storage and everything there is configured, including the device type, device location and mount point.
Why are the errors vague and without presenting any kind of overall understanding (of the system’s self knowledge of dependencies) of the configuration? Also, it let me add a device to the profile without first validating it.
The other problem is that this scenario isn’t covered directly anywhere and the documentation snippets found everywhere are for different versions of LXD. It is just a nightmare.
Also, none of the documentation fragments present how these things go together and their dependencies.
LXD is awesome but, man, are there some severe holes in the documentation. Most documentation failures occur because of the absence of relationships between technical details. Good documentation presents the system well as a sum of connected parts.
I removed root from the profile and successfully re-added it using:
root@code:/# lxc profile device add default root disk path=/ pool=containers
Device root added to default
profile output:
root@code:/# lxc profile show default
config: {}
description: Default LXD profile
devices:
eth0:
name: eth0
network: lxdbr0
type: nic
root:
path: /
pool: containers
type: disk
name: default
used_by: []
A new container instance was successfully created.
So now I go looking for confirmation that it was created into the containers ZFS pool/volume. This is located at /mnt/containers on the host server.
Nothing is showing inside:
root@code:/mnt/containers# ll
total 8
drwxr-xr-x 2 root root 4096 Sep 23 02:09 ./
drwxr-xr-x 4 root root 4096 Sep 23 02:09 ../
vscode container output:
root@code:/# lxc info vscode
Name: vscode
Location: none
Remote: unix://
Architecture: x86_64
Created: 2021/09/23 05:09 UTC
Status: Running
Type: container
Profiles: default
Pid: 31875
Ips:
eth0: inet 10.77.109.216 vethddb3c657
eth0: inet6 fd42:185c:bcfb:88c0:216:3eff:fea3:9e1f vethddb3c657
eth0: inet6 fe80::216:3eff:fea3:9e1f vethddb3c657
lo: inet 127.0.0.1
lo: inet6 ::1
Resources:
Processes: 42
Disk usage:
root: 8.80MB
CPU usage:
CPU usage (in seconds): 15
Memory usage:
Memory (current): 166.04MB
Memory (peak): 215.13MB
Network usage:
eth0:
Bytes received: 27.29kB
Bytes sent: 12.12kB
Packets received: 63
Packets sent: 81
lo:
Bytes received: 1.48kB
Bytes sent: 1.48kB
Packets received: 16
Packets sent: 16
root@code:/# lxc storage show containers
config:
source: containers
volatile.initial_source: /dev/disk/by-id/scsi-0DO_Volume_containers
zfs.pool_name: containers
description: ""
name: containers
driver: zfs
used_by:
- /1.0/images/a068e8daef0f88c667fd0f201ed0de1c48693ee383eeafbee6a51b79b0d29fea
- /1.0/instances/vscode
- /1.0/profiles/default
status: Created
locations:
- none
file system output at volatile.initial_source:
root@code:/dev/disk/by-id# ll
total 0
drwxr-xr-x 2 root root 100 Sep 23 02:43 ./
drwxr-xr-x 8 root root 160 Sep 23 02:43 ../
lrwxrwxrwx 1 root root 9 Sep 23 02:43 scsi-0DO_Volume_containers -> ../../sda
lrwxrwxrwx 1 root root 10 Sep 23 02:43 scsi-0DO_Volume_containers-part1 -> ../../sda1
lrwxrwxrwx 1 root root 10 Sep 23 02:43 scsi-0DO_Volume_containers-part9 -> ../../sda9
So then I check for the ZFS volume and it isn’t listed:
root@code:/mnt# df -h
Filesystem Size Used Avail Use% Mounted on
udev 474M 0 474M 0% /dev
tmpfs 99M 1004K 98M 1% /run
/dev/vda1 25G 2.4G 22G 10% /
tmpfs 491M 0 491M 0% /dev/shm
tmpfs 5.0M 0 5.0M 0% /run/lock
tmpfs 491M 0 491M 0% /sys/fs/cgroup
/dev/loop0 56M 56M 0 100% /snap/core18/2066
/dev/loop1 33M 33M 0 100% /snap/snapd/11841
/dev/loop2 68M 68M 0 100% /snap/lxd/20326
/dev/vda15 105M 5.2M 100M 5% /boot/efi
tmpfs 99M 0 99M 0% /run/user/0
tmpfs 1.0M 0 1.0M 0% /var/snap/lxd/common/ns
/dev/loop3 33M 33M 0 100% /snap/snapd/13170
/dev/loop4 56M 56M 0 100% /snap/core18/2128
/dev/loop5 62M 62M 0 100% /snap/core20/1081
/dev/loop6 68M 68M 0 100% /snap/lxd/21545
I guess LXD unmounted it, but zpool still lists it which is to be expected:
root@code:/# zpool list
NAME SIZE ALLOC FREE CKPOINT EXPANDSZ FRAG CAP DEDUP HEALTH ALTROOT
containers 19.5G 575M 18.9G - - 0% 2% 1.00x ONLINE -
Ok, so it looks like the ZFS volume was not mounted at /mnt/containers/ and was instead mounted here:
root@code:/# zpool history containers
History for 'containers':
...
2021-09-23.05:09:31 zfs create -o mountpoint=/var/snap/lxd/common/lxd/storage-pools/containers/
Directory output:
root@code:/var/snap/lxd/common/lxd/storage-pools/containers# ll
total 36
drwx--x--x 9 root root 4096 Sep 23 02:43 ./
drwx--x--x 3 root root 4096 Sep 23 02:43 ../
drwx--x--x 3 root root 4096 Sep 23 05:10 containers/
drwx--x--x 2 root root 4096 Sep 23 02:43 containers-snapshots/
drwx--x--x 2 root root 4096 Sep 23 02:43 custom/
drwx--x--x 2 root root 4096 Sep 23 02:43 custom-snapshots/
drwx--x--x 3 root root 4096 Sep 23 05:09 images/
drwx--x--x 2 root root 4096 Sep 23 02:43 virtual-machines/
drwx--x--x 2 root root 4096 Sep 23 02:43 virtual-machines-snapshots/
I had no control over where the ZFS volume was mounted, it seems. So, it was just mounted out in some random place - in the Snap directory of all places. I mean, WTF.
But, it’s still only guesswork because zpool can’t tell me the mountpoint for the containers ZFS volume (which is a 20GB Digital Ocean block volume attached to the droplet).
Yes this is the correct approach:
lxc profile device add default root disk path=/ pool=containers
You can see the ZFS volumes/datasets LXD creates by running zfs list this will show which pool they are on.
As you have correctly found the storage pool mount points are not customisable and are created inside the LXD directory (which is also inside the snap mount namespace if using the snap package).
Its not random, it is by design. The mount point is internal to LXD so it creates it inside the LXD internal directory.
We also set canmount=noauto and volmode=none on the the datasets/volumes created in the pool so that they don’t appear in the host’s filesystem or /dev directory (for block volumes) by default.
This error message appears if you try to snap install LXD and skip the lxd init (or lxd init --minimal) step before launching an image (lxd launch image:22.04 containername)
lxd init solved this problem for me