Command 'juju deploy'

The information in this doc is based on Juju version 3.5.5, and may not accurately reflect other versions of Juju.

See also: integrate, add-unit, config, expose, constraints, refresh, set-constraints, spaces, charm-resources

Summary

Deploys a new application or bundle.

Usage

juju deploy [options] <charm or bundle> [<application name>]

Options

Flag Default Usage
-B, --no-browser-login false Do not use web browser for authentication
--attach-storage Existing storage to attach to the deployed unit (not available on k8s models)
--base The base on which to deploy
--bind Configure application endpoint bindings to spaces
--channel Channel to use when deploying a charm or bundle from Charmhub
--config Either a path to yaml-formatted application config file or a key=value pair
--constraints Set application constraints
--device Charm device constraints
--dry-run false Just show what the deploy would do
--force false Allow a charm/bundle to be deployed which bypasses checks such as supported base or LXD profile allow list
-m, --model Model to operate in. Accepts [<controller name>:]<model name>|<model UUID>
--map-machines Specify the existing machines to use for bundle deployments
-n, --num-units 1 Number of application units to deploy for principal charms
--overlay Bundles to overlay on the primary bundle, applied in order
--resource Resource to be uploaded to the controller
--revision -1 The revision to deploy
--series The series on which to deploy. DEPRECATED: use --base
--storage Charm storage constraints
--to The machine and/or container to deploy the unit in (bypasses constraints)
--trust false Allows charm to run hooks that require access credentials

Examples

Deploy to a new machine:

juju deploy apache2

Deploy to machine 23:

juju deploy mysql --to 23

Deploy to a new LXD container on a new machine:

juju deploy mysql --to lxd

Deploy to a new LXD container on machine 25:

juju deploy mysql --to lxd:25

Deploy to LXD container 3 on machine 24:

juju deploy mysql --to 24/lxd/3

Deploy 2 units, one on machine 3 and one to a new LXD container on machine 5:

juju deploy mysql -n 2 --to 3,lxd:5

Deploy 3 units, one on machine 3 and the remaining two on new machines:

juju deploy mysql -n 3 --to 3

Deploy to a machine with at least 8 GiB of memory:

juju deploy postgresql --constraints mem=8G

Deploy to a specific availability zone (provider-dependent):

juju deploy mysql --to zone=us-east-1a

Deploy to a specific MAAS node:

juju deploy mysql --to host.maas

Deploy to a machine that is in the ‘dmz’ network space but not in either the ‘cms’ nor the ‘database’ spaces:

juju deploy haproxy -n 2 --constraints spaces=dmz,^cms,^database

Deploy a k8s charm that requires a single Nvidia GPU:

juju deploy mycharm --device miner=1,nvidia.com/gpu

Deploy a k8s charm that requires two Nvidia GPUs that have an attribute of ‘gpu=nvidia-tesla-p100’:

juju deploy mycharm --device \
   twingpu=2,nvidia.com/gpu,gpu=nvidia-tesla-p100

Deploy with specific resources:

juju deploy foo --resource bar=/some/file.tgz --resource baz=./docs/cfg.xml

Details

A charm or bundle can be referred to by its simple name and a base, revision, or channel can optionally be specified:

juju deploy postgresql juju deploy ch:postgresql --base ubuntu@22.04 juju deploy ch:postgresql --channel edge juju deploy ch:ubuntu --revision 17 --channel edge

All the above deployments use remote charms found in Charmhub, denoted by the ‘ch:’ prefix. Remote charms with no prefix will be deployed from Charmhub.

If a channel is specified, it will be used as the source for looking up the charm or bundle from Charmhub. When used in a bundle deployment context, the specified channel is only used for retrieving the bundle and is ignored when looking up the charms referenced by the bundle. However, each charm within a bundle is allowed to explicitly specify the channel used to look it up.

If a revision is specified, a channel must also be specified for Charmhub charms and bundles. The charm will be deployed with revision. The channel will be used when refreshing the application in the future.

A local charm may be deployed by giving the path to its directory:

juju deploy /path/to/charm juju deploy /path/to/charm --base ubuntu@22.04

You will need to be explicit if there is an ambiguity between a local and a remote charm:

juju deploy ./pig juju deploy ch:pig

A bundle can be expressed similarly to a charm:

juju deploy mediawiki-single juju deploy mediawiki-single --base ubuntu@22.04 juju deploy ch:mediawiki-single

A local bundle may be deployed by specifying the path to its YAML file:

juju deploy /path/to/bundle.yaml

The final charm/machine base is determined using an order of precedence (most preferred to least):

  • the ‘–base’ command option
  • for a bundle, the series stated in each charm URL (in the bundle file)
  • for a bundle, the series given at the top level (in the bundle file)
  • the ‘default-base’ model key
  • the first base specified in the charm’s manifest file

An ‘application name’ provides an alternate name for the application. It works only for charms; it is silently ignored for bundles (although the same can be done at the bundle file level). Such a name must consist only of lower-case letters (a-z), numbers (0-9), and single hyphens (-). The name must begin with a letter and not have a group of all numbers follow a hyphen:

Valid: myappname, custom-app, app2-scat-23skidoo Invalid: myAppName, custom–app, app2-scat-23, areacode-555-info

Use the ‘–constraints’ option to specify hardware requirements for new machines. These become the application’s default constraints (i.e. they are used if the application is later scaled out with the add-unit command). To overcome this behaviour use the set-constraints command to change the application’s default constraints or add a machine (add-machine) with a certain constraint and then target that machine with add-unit by using the ‘–to’ option.

Use the ‘–device’ option to specify GPU device requirements (with Kubernetes). The below format is used for this option’s value, where the ‘label’ is named in the charm metadata file:

<label>=[<count>,]<device-class>|<vendor/type>[,<attributes>]

Use the ‘–config’ option to specify application configuration values. This option accepts either a path to a YAML-formatted file or a key=value pair. A file should be of this format:

<charm name>: <option name>: <option value> …

For example, to deploy ‘mediawiki’ with file ‘mycfg.yaml’ that contains:

mediawiki: name: my media wiki admins: me:pwdOne debug: true

use

juju deploy mediawiki --config mycfg.yaml

Key=value pairs can also be passed directly in the command. For example, to declare the ‘name’ key:

juju deploy mediawiki --config name=‘my media wiki’

To define multiple keys:

juju deploy mediawiki --config name=‘my media wiki’ --config debug=true

If a key gets defined multiple times the last value will override any earlier values. For example,

juju deploy mediawiki --config name=‘my media wiki’ --config mycfg.yaml

Similar to the ‘juju config’ command, if the value begins with an ‘@’ character, it will be treated as a path to a config file and its contents will be assigned to the specified key. For example,

juju deploy mediawiki --config name='@wiki-name.txt"

will set the ‘name’ key to the contents of file ‘wiki-name.txt’.

If mycfg.yaml contains a value for ‘name’, it will override the earlier ‘my media wiki’ value. The same applies to single value options. For example,

juju deploy mediawiki --config name=‘a media wiki’ --config name=‘my wiki’

the value of ‘my wiki’ will be used.

Use the ‘–resource’ option to specify the resources you want to use for your charm. The format is

--resource <resource name>=<resource>

where the resource name is the name from the metadata.yaml file of the charm and where, depending on the type of the resource, the resource can be specified as follows:

(1) If the resource is type ‘file’, you can specify it by providing (a) the resource revision number or (b) a path to a local file.

(2) If the resource is type ‘oci-image’, you can specify it by providing (a) the resource revision number, (b) a path to a local file = private OCI image, (c) a link to a public OCI image.

Note: If you choose (1b) or (2b-c), i.e., a resource that is not from Charmhub: You will not be able to go back to using a resource from Charmhub.

Note: If you choose (1b) or (2b): This uploads a file from your loal disk to the juju controller to be streamed to the charm when “resource-get” is called by a hook.

Note: If you choose (2b): You will need to specify: (i) the local path to the private OCI image as well as (ii) the username/password required to access the private OCI image.

Note: If multiple resources are needed, repeat the option.

Use the ‘–to’ option to deploy to an existing machine or container by specifying a “placement directive”. The status command should be used for guidance on how to refer to machines. A few placement directives are provider-dependent (e.g.: ‘zone’).

In more complex scenarios, “network spaces” are used to partition the cloud networking layer into sets of subnets. Instances hosting units inside the same space can communicate with each other without any firewalls. Traffic crossing space boundaries could be subject to firewall and access restrictions. Using spaces as deployment targets, rather than their individual subnets, allows Juju to perform automatic distribution of units across availability zones to support high availability for applications. Spaces help isolate applications and their units, both for security purposes and to manage both traffic segregation and congestion.

When deploying an application or adding machines, the ‘spaces’ constraint can be used to define a comma-delimited list of required and forbidden spaces (the latter prefixed with ‘^’, similar to the ‘tags’ constraint).

When deploying bundles, machines specified in the bundle are added to the model as new machines. Use the ‘–map-machines=existing’ option to make use of any existing machines. To map particular existing machines to machines defined in the bundle, multiple comma separated values of the form ‘bundle-id=existing-id’ can be passed. For example, for a bundle that specifies machines 1, 2, and 3; and a model that has existing machines 1, 2, 3, and 4, the below deployment would have existing machines 1 and 2 assigned to machines 1 and 2 defined in the bundle and have existing machine 4 assigned to machine 3 defined in the bundle.

juju deploy mybundle --map-machines=existing,3=4

Only top level machines can be mapped in this way, just as only top level machines can be defined in the machines section of the bundle.

When charms that include LXD profiles are deployed the profiles are validated for security purposes by allowing only certain configurations and devices. Use the ‘–force’ option to bypass this check. Doing so is not recommended as it can lead to unexpected behaviour.

Further reading: https://juju.is/docs/olm/manage-applications


Small readability issue: Most commands in the “Details” section are in plain text instead of a code block.