How to modify an instance

Errors or typos? Topics missing? Hard to read? Let us know or open an issue on GitHub.

See also: Instance

This document shows further ways to customize an instance outside of the launch command, via settings.

Set the CPU, RAM or disk of an instance

See also: local.<instance-name>.(cpus|disk|memory)

While instance properties can be determined at launch, some of them can be updated after the instance has been created. Specifically, an instance’s memory, disk space, and the number of its CPUs are exposed via daemon settings: local.<instance-name>.(cpus|disk|memory).

To modify one of this properties, first stop the instance and then issue the set command. For example:

$ multipass stop handsome-ling
$ multipass set local.handsome-ling.cpus=4
$ multipass set local.handsome-ling.disk=60G
$ multipass set local.handsome-ling.memory=7G

When increasing the disk size of an instance, the partition may not expand automatically to use the new available space. This usually happens if the partition was already full when trying to increase the disk size. In such cases, the partition must be expanded manually, as follows:

$ multipass shell handsome-ling
ubuntu@handsome-ling$ <b>sudo parted /dev/sda resizepart 1 100%</b>
Warning: Not all of the space available to /dev/sda appears to be used, you can fix the GPT to use all of the space (an extra 4194304 blocks) or continue with the current setting?
Fix/Ignore? <b>fix</b>
Partition number? <b>1</b>
Warning: Partition /dev/sda1 is being used. Are you sure you want to continue?
Yes/No? <b>yes</b>
ubuntu@handsome-ling$ <b>sudo resize2fs /dev/sda1</b>

These properties can be consulted with the get command. Instances do not have to be stopped for that. For example:

$ multipass get local.handsome-ling.cpus
$ multipass get local.handsome-ling.disk
$ multipass get local.handsome-ling.memory

Only properties of stopped, non-deleted instances can be updated, but all instances can have their properties fetched. Their keys can be obtained with multipass get --keys. Yet, trying to update an instance that is running, suspended, or deleted, results in an error.

Modifying instance settings is not supported when using the hyperkit driver, which has been deprecated in favor of the qemu. The qemu and virtualbox drivers on Intel macOS do support instance modification.

Set the status of an instance to primary

See also: client.primary-name

This section demonstrates how to set the status of an instance to primary. This is convenient because it makes this instance the default argument for several commands, such as shell , start , stop , restart , and suspend and also automatically mounts our $HOME directory in the instance.

To grant a regular instance the primary status, assign its name to the client.primary-name:

$ multipass set client.primary-name=<instance name>

This setting allows transferring primary status among instances. The primary name can be configured independently of whether instances with the old and new names exist. If they do, they lose and gain primary status accordingly.

This provides a means of (de)selecting an existing instance as primary.


Assign the primary status to an instance called ‘first’:

$ multipass set client.primary-name=first

Now this instance is picked up automatically by ‘multipass start’. The primary instance also automatically mounts the user’s home directory into a directory called ‘Home’:

$ multipass start
Launched: first
Mounted '/home/ubuntu' into 'first:Home'

Stop the primary instance:

$ multipass stop
$ multipass launch --name second
Launched: second

Change the primary instance to an existing instance:

$ multipass set client.primary-name=second

Now this instance is used by default by the command:

$ multipass suspend

When listing all instances, the primary one is displayed first:

$ multipass list
Name                    State             IPv4             Image
second                  Suspended         --               Ubuntu 18.04 LTS
first                   Stopped           --               Ubuntu 18.04 LTS

Let us know how this worked for you and what you’d like to see next!


I am running multipass 1.9.0 on Mac book Pro with Apple silicone.

Trying to update disk size and getting this error :

multipass set local.superior-chihuahua.disk=20G
set failed: Unrecognized settings key: ‘local.superior-chihuahua.disk’

Hi @garichd!

Sorry for the confusion. This document did not specify the Multipass version this will be in. It will be in the upcoming version 1.10 and is available only as a preview on Linux.

I’ve updated the documented to reflect this fact.

1 Like

Thanks for the reply @townsend
is there anyway I can use this feature earlier with specific patch , I noticed that this was released back in 2021 under build #2353, However I am not sure the way to get that build installed.!

My VM was installed with default settings and I am now running out of space, Its configured heavily and to redo the work will set me back weeks…! :frowning:

Hi @garichd,

It was actually merged into the main branch on April 22.

If you are brave adventurous enough, you can get an unsigned-but-built-by-ci package from that has this feature. I don’t anticipate you having any major regressions with this development package, but of course it can happen.


Hi there… quick heads up… the link
See also: local.<instance-name>.(cpus|disk|memory)
is broken.


I ran out of space so I resized with

multipass set local.charm-dev-2cpu-7g.disk=70G

That seemed to work:

$ sudo parted /dev/sda print free
Disk /dev/sda: 75.2GB
Sector size (logical/physical): 512B/512B
Partition Table: gpt
Disk Flags: 

Number  Start   End     Size    File system  Name  Flags
        17.4kB  1049kB  1031kB  Free Space
14      1049kB  5243kB  4194kB                     bios_grub
15      5243kB  116MB   111MB   fat32              boot, esp
 1      116MB   53.7GB  53.6GB  ext4
        53.7GB  75.2GB  21.5GB  Free Space

At this point I wasn’t sure how to go about resizing the partition: sudo systemctl isolate hangs. So I went on and resized it while it was mounted:

(parted) resizepart 1 75.2GB
Warning: Partition /dev/sda1 is being used. Are you sure you want to continue?
Yes/No? Yes

However, this change isn’t reflected in df -h, even after a reboot.
What am I missing here?

After I manually freed up some space and rebooted, df reflected the change. Interesting.

Hey @leon-mintz!

Glad you got it working! cloud-init will grow the partition to match the disk size automatically during boot. I suspect that something with cloud-init failed due to the lack of disk space.

That makes sense, thanks @townsend.

  • Do you mean there was no need to use parted, only (free some space and) reboot?
  • If using parted is still needed, shouldn’t we unmount first? How?

Hey @leon-mintz!

Do you mean there was no need to use parted, only (free some space and) reboot?

Yes, the growpart directive for cloud-init tells cloud-init to automatically grow the partition on boot to match the disk space. It basically does the parted for you.

1 Like

The current document is incomplete with regards to resizing the disk size of a VM. AFTER the resizing, the partitions need to be updated from within the virtual machine itself, it seems. Here is a very nicely written guide which worked for me:

Hey, @vidimuswolf! Indeed, I can reproduce this and it is a problem. We’ll investigate the issue and see if we can come up with a fix. Thanks for reporting this!