env_problem
def env_problem(kv: String) -> StringWhy kv can't be one of the control instance's settings (--env KEY=VALUE, the stack's ControlEnv), or empty if it can.
orca.aws
orca aws: a project's orca on AWS, with the aws and docker commands its user already has (aws login, or AWS_PROFILE for a named profile).
orca aws up --email you@example.com [--stack orca] [--hostname H] [--console-hostname H] [--budget 100] [--zone us-east-2a] [--public-data true|false] [--env KEY=VALUE ...] orca aws push [--tag T] # the project's image, to ECR orca aws register [--tag T] [--config F|--no-config] [register flags] <pipeline> [...] orca aws app <name> [--host H] [--port 8080] [--tag T] orca aws status orca aws replace-control [--plan] [--wait] [--keep-old 24h] [--smoke P=KEY] [--env KEY=VALUE ...] | --resume | --rollback orca aws down --yes
up creates or updates the stack from the template orca carries
(deploy/aws/orca.yaml, in orca.aws_assets); the first time it needs
--email, for the Budgets alert, and makes the database's password. push
builds the project's image from a staged copy of pixi.toml, pixi.lock,
bin/, pipelines/ and src/, with the project's Dockerfile.orca or
else the template (deploy/project.Dockerfile), for arm64, tagged with the
commit (or --tag) and latest, and pushes it to the stack's repository.
register and app run on the control instance through SSM: register
moves the scheduler and the UI to the image (orca-refresh) and registers
each pipeline, /opt/orca/pipelines/<pipeline>, with the flags given, as
orca register takes them, and the project's orca.toml (or --config's
file; orca.config), which it checks and sends whole with its git commit; app runs the image's /opt/orca/bin/<name>
beside the UI, at https://<host> (for console, the stack's
ConsoleHostname). --env KEY=VALUE (repeated) sets the control
instance's settings beyond the template's (ControlEnv, all of them at
once; never a secret). An update changes no running control instance: its
boot script is the stack's ControlBoot, which a new one runs, and each
slot's AMI is pinned to the one its instance has. replace-control
(orca.aws_replace) boots a new control instance from the template, with
the old one's database, and moves the Elastic IP to it. down deletes the
stack, which fails while its buckets hold objects.
Files go to build/orca-aws/. --stack (default orca) and --region
(default AWS_REGION, else the aws command's own) go anywhere.
Functions
control_env
def control_env(o: _Opts) -> StringThe ControlEnv parameter --env gives.
deploy
def deploy(o: _Opts, params: List[String], show: Bool)cloudformation deploy of the template orca carries with params (the others keep their values), printing them if show.
image_of
def image_of(o: _Opts, instance: String) -> StringThe AMI instance was started from.
slot_pins
def slot_pins(o: _Opts, outs: Dict[String, String]) -> List[String]BlueAmi= and GreenAmi= the images the slots' instances have, so that an update, whose Ami resolves again, replaces neither.
ssm_run
def ssm_run(o: _Opts, instance: String, commands: List[String]) -> Tuple[Int, String]Runs commands on instance through SSM and prints what they wrote; 0 and their output if they succeeded, else 3 and it.
aws_command
def aws_command(args: List[String]) -> Intorca aws <cmd> ...: args is the whole command line.
Aliases
comptime AWS_USAGE = "usage:\n orca aws up --email ADDR [--hostname H] [--console-hostname H] [--budget N]\n [--zone AZ] [--public-data true|false] [--env KEY=VALUE ...]\n orca aws push [--tag T]\n orca aws register [--tag T] [--config F|--no-config] [register flags]\n <pipeline> [...]\n orca aws app <name> [--host H] [--port 8080] [--tag T]\n orca aws status\n orca aws replace-control [--plan] [--wait] [--keep-old 24h] [--smoke P=KEY]\n [--env KEY=VALUE ...] | --resume | --rollback\n orca aws down --yes\n (each takes --stack NAME, default orca, and --region R)"