The information in this doc is based on Juju version 3.5.6, and may not accurately reflect other versions of Juju.
See also: add-credential, autoload-credentials, add-model, controller-config, model-config, set-constraints, show-cloud
Summary
Initializes a cloud environment.
Usage
juju bootstrap [options] [<cloud name>[/region] [<controller name>]]
Options
| Flag | Default | Usage |
|---|---|---|
-B, --no-browser-login |
false | Do not use web browser for authentication |
--add-model |
Name of an initial model to create on the new controller | |
--agent-version |
Version of agent binaries to use for Juju agents | |
--auto-upgrade |
false | After bootstrap, upgrade to the latest patch release |
--bootstrap-base |
Specify the base of the bootstrap machine | |
--bootstrap-constraints |
Specify bootstrap machine constraints | |
--bootstrap-image |
Specify the image of the bootstrap machine (requires --bootstrap-constraints specifying architecture) | |
--bootstrap-series |
Specify the series of the bootstrap machine (deprecated use bootstrap-base) | |
--build-agent |
false | Build local version of agent binary before bootstrapping |
--clouds |
false | Print the available clouds which can be used to bootstrap a Juju environment |
--config |
Specify a controller configuration file, or one or more configuration options. Model config keys only affect the controller model. (–config config.yaml [–config key=value …]) | |
--constraints |
Set model constraints | |
--controller-charm-channel |
3.5/stable | The Charmhub channel to download the controller charm from (if not using a local charm) |
--controller-charm-path |
Path to a locally built controller charm | |
--credential |
Credentials to use when bootstrapping | |
--db-snap |
Path to a locally built .snap to use as the internal juju-db service. | |
--db-snap-asserts |
Path to a local .assert file. Requires --db-snap | |
--force |
false | Allow the bypassing of checks such as supported series |
--keep-broken |
false | Do not destroy the provisioned controller instance if bootstrap fails |
--metadata-source |
Local path to use as agent and/or image metadata source | |
--model-default |
Specify a configuration file, or one or more configuration options to be set for all models, unless otherwise specified (–model-default config.yaml [–model-default key=value …]) | |
--no-switch |
false | Do not switch to the newly created controller |
--regions |
Print the available regions for the specified cloud | |
--storage-pool |
Specify options for an initial storage pool ‘name’ and ‘type’ are required, plus any additional attributes (–storage-pool pool-config.yaml [–storage-pool key=value …]) | |
--to |
Placement directive indicating an instance to bootstrap |
Examples
juju bootstrap
juju bootstrap --clouds
juju bootstrap --regions aws
juju bootstrap aws
juju bootstrap aws/us-east-1
juju bootstrap google joe-us-east1
juju bootstrap --config=~/config-rs.yaml google joe-syd
juju bootstrap --agent-version=2.2.4 aws joe-us-east-1
juju bootstrap --config bootstrap-timeout=1200 azure joe-eastus
juju bootstrap aws --storage-pool name=secret --storage-pool type=ebs --storage-pool encrypted=true
juju bootstrap lxd --bootstrap-base=ubuntu@22.04
# For a bootstrap on k8s, setting the service type of the Juju controller service to LoadBalancer
juju bootstrap --config controller-service-type=loadbalancer
# For a bootstrap on k8s, setting the service type of the Juju controller service to External
juju bootstrap --config controller-service-type=external --config controller-external-name=controller.juju.is
Details
Used without arguments, bootstrap will step you through the process of initializing a Juju cloud environment. Initialization consists of creating a ‘controller’ model and provisioning a machine to act as controller.
Controller names may only contain lowercase letters, digits and hyphens, and may not start with a hyphen. We recommend you call your controller ‘username-region’ e.g. ‘fred-us-east-1’. See --clouds for a list of clouds and credentials. See --regions <cloud> for a list of available regions for a given cloud.
Credentials are set beforehand and are distinct from any other
configuration (see juju add-credential).
The ‘controller’ model typically does not run workloads. It should remain
pristine to run and manage Juju’s own infrastructure for the corresponding
cloud. Additional models should be created with juju add-model for workload purposes.
Note that a ‘default’ model is also created and becomes the current model
of the environment once the command completes. It can be discarded if
other models are created.
If ‘–bootstrap-constraints’ is used, its values will also apply to any future controllers provisioned for high availability (HA).
If ‘–constraints’ is used, its values will be set as the default
constraints for all future workload machines in the model, exactly as if
the constraints were set with juju set-model-constraints.
It is possible to override constraints and the automatic machine selection algorithm by assigning a “placement directive” via the ‘–to’ option. This dictates what machine to use for the controller. This would typically be used with the MAAS provider (‘–to <host>.maas’).
You can change the default timeout and retry delays used during the bootstrap by changing the following settings in your configuration (all values represent number of seconds):
# How long to wait for a connection to the controller
bootstrap-timeout: 1200 # default: 20 minutes
# How long to wait between connection attempts to a controller address.
bootstrap-retry-delay: 5 # default: 5 seconds
# How often to refresh controller addresses from the API server.
bootstrap-addresses-delay: 10 # default: 10 seconds
It is possible to override the base e.g. ubuntu@22.04, Juju attempts to bootstrap on to, by supplying a base argument to ‘–bootstrap-base’.
An error is emitted if the determined base is not supported. Using the ‘–force’ option to override this check:
juju bootstrap --bootstrap-base=ubuntu@22.04 --force
The ‘–bootstrap-series’ flag can be still used, but is deprecated in favour of ‘–bootstrap-base’.
Private clouds may need to specify their own custom image metadata and tools/agent. Use ‘–metadata-source’ whose value is a local directory.
By default, the Juju version of the agent binary that is downloaded and installed on all models for the new controller will be the same as that of the Juju client used to perform the bootstrap. However, a user can specify a different agent version via ‘–agent-version’ option to bootstrap command. Juju will use this version for models’ agents as long as the client’s version is from the same Juju release base. In other words, a 2.2.1 client can bootstrap any 2.2.x agents but cannot bootstrap any 2.0.x or 2.1.x agents. The agent version can be specified a simple numeric version, e.g. 2.2.4.
For example, at the time when 2.3.0, 2.3.1 and 2.3.2 are released and your agent stream is ‘released’ (default), then a 2.3.1 client can bootstrap:
- 2.3.0 controller by running ‘… bootstrap --agent-version=2.3.0 …’;
- 2.3.1 controller by running ‘… bootstrap …’;
- 2.3.2 controller by running ‘bootstrap --auto-upgrade’. However, if this client has a copy of codebase, then a local copy of Juju will be built and bootstrapped - 2.3.1.1.
Bootstrapping to a k8s cluster requires that the service set up to handle requests to the controller be accessible outside the cluster. Typically this means a service type of LoadBalancer is needed, and Juju does create such a service if it knows it is supported by the cluster. This is performed by interrogating the cluster for a well known managed deployment such as microk8s, GKE or EKS.
When bootstrapping to a k8s cluster Juju does not recognise, there’s no guarantee a load balancer is available, so Juju defaults to a controller service type of ClusterIP. This may not be suitable, so there are three bootstrap options available to tell Juju how to set up the controller service. Part of the solution may require a load balancer for the cluster to be set up manually first, or perhaps an external k8s service via a FQDN will be used (this is a cluster specific implementation decision which Juju needs to be informed about so it can set things up correctly). The three relevant bootstrap options are (see list of bootstrap config items below for a full explanation):
- controller-service-type
- controller-external-name
- controller-external-ips
Juju advertises those addresses to other controllers, so they must be resolveable from other controllers for cross-model (cross-controller, actually) relations to work.
If a storage pool is specified using --storage-pool, this will be created in the controller model.
Available keys for use with --config are:
Bootstrap configuration keys:
admin-secret:
type: string
description: Sets the Juju administrator password
bootstrap-addresses-delay:
type: int
description: Controls the amount of time in seconds in between refreshing the bootstrap
machine addresses
bootstrap-retry-delay:
type: int
description: Controls the amount of time in seconds between attempts to connect
to a bootstrap machine address
bootstrap-timeout:
type: int
description: Controls how long Juju will wait for a bootstrap to complete before
considering it failed in seconds
ca-cert:
type: string
description: Sets the bootstrapped controllers CA cert to use and issue certificates
from, used in conjunction with ca-private-key
ca-private-key:
type: string
description: Sets the bootstrapped controllers CA cert private key to sign certificates
with, used in conjunction with ca-cert
controller-external-ips:
type: list
description: Specifies a comma separated list of external IPs for a k8s controller
of type external
controller-external-name:
type: string
description: Sets the external name for a k8s controller of type external
controller-service-type:
type: string
description: |-
Controls the kubernetes service type for Juju controllers, see
https://kubernetes.io/docs/reference/kubernetes-api/service-resources/service-v1/#ServiceSpec
valid values are one of cluster, loadbalancer, external
Controller configuration keys:
agent-logfile-max-backups:
type: int
description: The number of old agent log files to keep (compressed)
agent-logfile-max-size:
type: string
description: The maximum size of the agent log file
agent-ratelimit-max:
type: int
description: The maximum size of the token bucket used to ratelimit agent connections
agent-ratelimit-rate:
type: string
description: The time taken to add a new token to the ratelimit bucket
allow-model-access:
type: bool
description: "Determines if the controller allows users to \nconnect to models they
have been authorized for even when \nthey don't have any access rights to the
controller itself"
api-port:
type: int
description: The port used for api connections
api-port-open-delay:
type: string
description: "The duration that the controller will wait \nbetween when the controller
has been deemed to be ready to open \nthe api-port and when the api-port is actually
opened \n(only used when a controller-api-port value is set)."
application-resource-download-limit:
type: int
description: The maximum number of concurrent resources downloads per application
audit-log-capture-args:
type: bool
description: Determines if the audit log contains the arguments passed to API methods
audit-log-exclude-methods:
type: list
description: The list of Facade.Method names that aren't interesting for audit logging
purposes.
audit-log-max-backups:
type: int
description: The number of old audit log files to keep (compressed)
audit-log-max-size:
type: string
description: The maximum size for the current controller audit log file
auditing-enabled:
type: bool
description: Determines if the controller records auditing information
autocert-dns-name:
type: string
description: The DNS name of the controller
autocert-url:
type: string
description: The URL used to obtain official TLS certificates when a client connects
to the API
caas-image-repo:
type: string
description: The docker repo to use for the jujud operator and mongo images
caas-operator-image-path:
type: string
description: |-
(deprecated) The url of the docker image used for the application operator.
Use "caas-image-repo" instead.
controller-api-port:
type: int
description: |-
An optional port that may be set for controllers
that have a very heavy load. If this port is set, this port is used by
the controllers to talk to each other - used for the local API connection
as well as the pubsub forwarders, and the raft workers. If this value is
set, the api-port isn't opened until the controllers have started properly.
controller-name:
type: string
description: The canonical name of the controller
controller-resource-download-limit:
type: int
description: The maximum number of concurrent resources downloads across all the
applications on the controller
features:
type: list
description: A list of runtime changeable features to be updated
identity-public-key:
type: string
description: The public key of the identity manager
identity-url:
type: string
description: The url of the identity manager
juju-db-snap-channel:
type: string
description: Sets channel for installing mongo snaps when bootstrapping on focal
or later
juju-ha-space:
type: string
description: The network space within which the MongoDB replica-set should communicate
juju-mgmt-space:
type: string
description: The network space that agents should use to communicate with controllers
jujud-controller-snap-source:
type: string
description: The source for the jujud-controller snap.
login-token-refresh-url:
type: string
description: The url of the jwt well known endpoint
max-agent-state-size:
type: int
description: The maximum size (in bytes) of internal state data that agents can
store to the controller
max-charm-state-size:
type: int
description: The maximum size (in bytes) of charm-specific state that units can
store to the controller
max-debug-log-duration:
type: string
description: The maximum duration that a debug-log session is allowed to run
max-prune-txn-batch-size:
type: int
description: (deprecated) The maximum number of transactions evaluated in one go
when pruning
max-prune-txn-passes:
type: int
description: (deprecated) The maximum number of batches processed when pruning
max-txn-log-size:
type: string
description: The maximum size the of capped txn log collection
metering-url:
type: string
description: The url for metrics
migration-agent-wait-time:
type: string
description: The maximum during model migrations that the migration worker will
wait for agents to report on phases of the migration
model-logfile-max-backups:
type: int
description: The number of old model log files to keep (compressed)
model-logfile-max-size:
type: string
description: The maximum size of the log file written out by the controller on behalf
of workers running for a model
model-logs-size:
type: string
description: The size of the capped collections used to hold the logs for the models
mongo-memory-profile:
type: string
description: Sets mongo memory profile
prune-txn-query-count:
type: int
description: The number of transactions to read in a single query
prune-txn-sleep-time:
type: string
description: The amount of time to sleep between processing each batch query
public-dns-address:
type: string
description: Public DNS address (with port) of the controller.
query-tracing-enabled:
type: bool
description: Enable query tracing for the dqlite driver
query-tracing-threshold:
type: string
description: "The minimum duration of a query for it to be traced. The lower the
\nthreshold, the more queries will be output. A value of 0 means all queries \nwill
be output if tracing is enabled."
set-numa-control-policy:
type: bool
description: Determines if the NUMA control policy is set
state-port:
type: int
description: The port used for mongo connections
Model configuration keys (affecting the controller model):
agent-metadata-url:
type: string
description: URL of private stream
agent-stream:
type: string
description: Version of Juju to use for deploy/upgrades.
apt-ftp-proxy:
type: string
description: The APT FTP proxy for the model
apt-http-proxy:
type: string
description: The APT HTTP proxy for the model
apt-https-proxy:
type: string
description: The APT HTTPS proxy for the model
apt-mirror:
type: string
description: The APT mirror for the model
apt-no-proxy:
type: string
description: List of domain addresses not to be proxied for APT (comma-separated)
authorized-keys:
type: string
description: Any authorized SSH public keys for the model, as found in a ~/.ssh/authorized_keys
file
automatically-retry-hooks:
type: bool
description: Determines whether the uniter should automatically retry failed hooks
backup-dir:
type: string
description: Directory used to store the backup working directory
charmhub-url:
type: string
description: The url for CharmHub API calls
cloudinit-userdata:
type: string
description: Cloud-init user-data (in yaml format) to be added to userdata for new
machines created in this model
container-image-metadata-defaults-disabled:
type: bool
description: Whether default simplestreams sources are used for image metadata with
containers.
container-image-metadata-url:
type: string
description: The URL at which the metadata used to locate container OS image ids
is located
container-image-stream:
type: string
description: The simplestreams stream used to identify which image ids to search
when starting a container.
container-inherit-properties:
type: string
description: List of properties to be copied from the host machine to new containers
created in this model (comma-separated)
container-networking-method:
type: string
description: Method of container networking setup - one of fan, provider, local
default-base:
type: string
description: The default base image to use for deploying charms, will act like --base
when deploying charms
default-space:
type: string
description: The default network space used for application endpoints in this model
development:
type: bool
description: Whether the model is in development mode
disable-network-management:
type: bool
description: Whether the provider should control networks (on MAAS models, set to
true for MAAS to control networks
disable-telemetry:
type: bool
description: Disable telemetry reporting of model information
egress-subnets:
type: string
description: Source address(es) for traffic originating from this model
enable-os-refresh-update:
type: bool
description: Whether newly provisioned instances should run their respective OS's
update capability.
enable-os-upgrade:
type: bool
description: Whether newly provisioned instances should run their respective OS's
upgrade capability.
extra-info:
type: string
description: Arbitrary user specified string data that is stored against the model.
fan-config:
type: string
description: Configuration for fan networking for this model
firewall-mode:
type: string
description: |-
The mode to use for network firewalling.
'instance' requests the use of an individual firewall per instance.
'global' uses a single firewall for all instances (access
for a network port is enabled to one instance if any instance requires
that port).
'none' requests that no firewalling should be performed
inside the model. It's useful for clouds without support for either
global or per instance security groups.
ftp-proxy:
type: string
description: The FTP proxy value to configure on instances, in the FTP_PROXY environment
variable
http-proxy:
type: string
description: The HTTP proxy value to configure on instances, in the HTTP_PROXY environment
variable
https-proxy:
type: string
description: The HTTPS proxy value to configure on instances, in the HTTPS_PROXY
environment variable
ignore-machine-addresses:
type: bool
description: Whether the machine worker should discover machine addresses on startup
image-metadata-defaults-disabled:
type: bool
description: Whether default simplestreams sources are used for image metadata.
image-metadata-url:
type: string
description: The URL at which the metadata used to locate OS image ids is located
image-stream:
type: string
description: The simplestreams stream used to identify which image ids to search
when starting an instance.
juju-ftp-proxy:
type: string
description: The FTP proxy value to pass to charms in the JUJU_CHARM_FTP_PROXY environment
variable
juju-http-proxy:
type: string
description: The HTTP proxy value to pass to charms in the JUJU_CHARM_HTTP_PROXY
environment variable
juju-https-proxy:
type: string
description: The HTTPS proxy value to pass to charms in the JUJU_CHARM_HTTPS_PROXY
environment variable
juju-no-proxy:
type: string
description: List of domain addresses not to be proxied (comma-separated), may contain
CIDRs. Passed to charms in the JUJU_CHARM_NO_PROXY environment variable
logforward-enabled:
type: bool
description: Whether syslog forwarding is enabled.
logging-config:
type: string
description: The configuration string to use when configuring Juju agent logging
(see http://godoc.org/github.com/juju/loggo#ParseConfigurationString for details)
logging-output:
type: string
description: 'The logging output destination: database and/or syslog. (default "")'
lxd-snap-channel:
type: string
description: The channel to use when installing LXD from a snap (cosmic and later)
max-action-results-age:
type: string
description: The maximum age for action entries before they are pruned, in human-readable
time format
max-action-results-size:
type: string
description: The maximum size for the action collection, in human-readable memory
format
max-status-history-age:
type: string
description: The maximum age for status history entries before they are pruned,
in human-readable time format
max-status-history-size:
type: string
description: The maximum size for the status history collection, in human-readable
memory format
mode:
type: string
description: |-
Mode is a comma-separated list which sets the
mode the model should run in. So far only one is implemented
- If 'requires-prompts' is present, clients will ask for confirmation before removing
potentially valuable resources.
(default "")
net-bond-reconfigure-delay:
type: int
description: The amount of time in seconds to sleep between ifdown and ifup when
bridging
no-proxy:
type: string
description: List of domain addresses not to be proxied (comma-separated)
num-container-provision-workers:
type: int
description: The number of container provisioning workers to use per machine
num-provision-workers:
type: int
description: The number of provisioning workers to use per model
provisioner-harvest-mode:
type: string
description: What to do with unknown machines (default destroyed)
proxy-ssh:
type: bool
description: Whether SSH commands should be proxied through the API server
resource-tags:
type: attrs
description: resource tags
saas-ingress-allow:
type: string
description: |-
Application-offer ingress allowlist is a comma-separated list of
CIDRs specifying what ingress can be applied to offers in this model.
secret-backend:
type: string
description: The name of the secret store backend. (default "auto")
snap-http-proxy:
type: string
description: The HTTP proxy value for installing snaps
snap-https-proxy:
type: string
description: The HTTPS proxy value for installing snaps
snap-store-assertions:
type: string
description: The assertions for the defined snap store proxy
snap-store-proxy:
type: string
description: The snap store proxy for installing snaps
snap-store-proxy-url:
type: string
description: The URL for the defined snap store proxy
ssh-allow:
type: string
description: |-
SSH allowlist is a comma-separated list of CIDRs from
which machines in this model will accept connections to the SSH service.
Currently only the aws & openstack providers support ssh-allow
ssl-hostname-verification:
type: bool
description: Whether SSL hostname verification is enabled (default true)
storage-default-block-source:
type: string
description: The default block storage source for the model
storage-default-filesystem-source:
type: string
description: The default filesystem storage source for the model
syslog-ca-cert:
type: string
description: The certificate of the CA that signed the syslog server certificate,
in PEM format.
syslog-client-cert:
type: string
description: The syslog client certificate in PEM format.
syslog-client-key:
type: string
description: The syslog client key in PEM format.
syslog-host:
type: string
description: The hostname:port of the syslog server.
test-mode:
type: bool
description: |-
Whether the model is intended for testing.
If true, accessing the charm store does not affect statistical
data of the store. (default false)
transmit-vendor-metrics:
type: bool
description: Determines whether metrics declared by charms deployed into this model
are sent for anonymized aggregate analytics
update-status-hook-interval:
type: string
description: How often to run the charm update-status hook, in human-readable time
format (default 5m, range 1-60m)