NAME¶

podman-auto-update - Auto update containers according to their auto-update policy

SYNOPSIS¶

podman auto-update [options]

DESCRIPTION¶

podman auto-update pulls down new container images and restarts containers configured for auto updates. To make use of auto updates, the container or Kubernetes workloads must run inside a systemd unit. After a successful update of an image, the containers using the image get updated by restarting the systemd units they run in. Please refer to podman-systemd.unit(5) on how to run Podman under systemd.

To configure a container for auto updates, it must be created with the io.containers.autoupdate label or the AutoUpdate field in podman-systemd.unit(5) with one of the following two values:

registry: If the label is present and set to registry, Podman reaches out to the corresponding registry to check if the image has been updated. The label image is an alternative to registry maintained for backwards compatibility. An image is considered updated if the digest in the local storage is different than the one of the remote image. If an image must be updated, Podman pulls it down and restarts the systemd unit executing the container. The registry policy requires a fully-qualified image reference (e.g., quay.io/podman/stable:latest) to be used to create the container. This enforcement is necessary to know which image to actually check and pull. If an image ID was used, Podman would not know which image to check/pull anymore.
local: If the autoupdate label is set to local, Podman compares the image digest of the container to the one in the local container storage. If they differ, the local image is considered to be newer and the systemd unit gets restarted.

Auto Updates and Kubernetes YAML¶

Podman supports auto updates for Kubernetes workloads. The auto-update policy can be configured directly via podman-systemd.unit(5) or inside the Kubernetes YAML with the Podman-specific annotations mentioned below:

io.containers.autoupdate: “registry|local” to apply the auto-update policy to all containers
io.containers.autoupdate/$container: “registry|local” to apply the auto-update policy to $container only
io.containers.sdnotify: “conmon|container” to apply the sdnotify policy to all containers
io.containers.sdnotify/$container: “conmon|container” to apply the sdnotify policy to $container only

By default, the autoupdate policy is set to “disabled”, the sdnotify policy is set to “conmon”.

Systemd Unit and Timer¶

Podman ships with a podman-auto-update.service systemd unit. This unit is triggered daily at midnight by the podman-auto-update.timer systemd timer. The timer can be altered for custom time-based updates if desired. The unit can further be invoked by other systemd units (e.g., via the dependency tree) or manually via systemctl start podman-auto-update.service.

OPTIONS¶

--authfile=path¶

Path of the authentication file. Default is ${XDG_RUNTIME_DIR}/containers/auth.json on Linux, and $HOME/.config/containers/auth.json on Windows/macOS. The file is created by podman login. If the authorization state is not found there, $HOME/.docker/config.json is checked, which is set using docker login.

Note: There is also the option to override the default path of the authentication file by setting the REGISTRY_AUTH_FILE environment variable. This can be done with export REGISTRY_AUTH_FILE=path.

Alternatively, the io.containers.autoupdate.authfile container label can be configured. In that case, Podman will use the specified label’s value instead.

--dry-run¶

Check for the availability of new images but do not perform any pull operation or restart any service or container. The UPDATED field indicates the availability of a new image with “pending”.

--filter, -f=filter¶

Filter what containers auto update. Multiple filters can be given with multiple uses of the --filter flag. Filters with the same key work inclusive with the only exception being label which is exclusive. Filters with different keys always work exclusive.

Valid filters are listed below:

Filter	Description
id	[ID] Container’s ID (CID prefix match by default; accepts regex)
name	[Name] Container’s name (accepts regex)
annotation	[Key] or [Key=Value] Annotation assigned to a container
annotation!	[Key] or [Key=Value] Annotation NOT assigned to a container
label	[Key] or [Key=Value] Label assigned to a container
label!	[Key] or [Key=Value] Label NOT assigned to a container
exited	[Int] Container’s exit code
status	[Status] Container’s status: ‘created’, ‘initialized’, ‘running’, ‘stopped’, ‘paused’, ‘exited’, ‘removing’, ‘stopping’, ‘unknown’
ancestor	[ImageName] Image or descendant used to create container (accepts regex)
before	[ID] or [Name] Containers created before this container
since	[ID] or [Name] Containers created since this container
volume	[VolumeName] or [MountpointDestination] Volume mounted in container
health	[Status] healthy or unhealthy
pod	[Pod] name or full or partial ID of pod
network	[Network] name or full ID of network
restart-policy	[Policy] Container’s restart policy (e.g., ‘no’, ‘on-failure’, ‘always’, ‘unless-stopped’)
until	[DateTime] Containers created before the given duration or time.
command	[Command] the command the container is executing, only argv[0] is taken
should-start-on-boot	[Bool] Containers that need to be restarted after system reboot. True for containers with restart policy ‘always’, or ‘unless-stopped’ that were not explicitly stopped by the user

Only matching running containers with an enabled auto-update policy are considered for updates. The label filter matches labels assigned to containers. For example, --filter label=app=web selects containers labeled app=web. Use --filter ancestor=IMAGE to select containers by image. Without filters, all eligible containers are considered. A filter that matches no eligible containers succeeds without updating anything.

Updates restart the systemd unit associated with a selected container. For containers in a pod, this may be the pod’s unit. Other containers in the same unit or affected by systemd dependencies may also restart, even if they do not match the filters. Those containers may use updated images when recreated, including images shared with selected containers.

--format=format¶

Change the default output format. This can be of a supported type like ‘json’ or a Go template. Valid placeholders for the Go template are listed below:

Placeholder	Description
.Container	ID and name of the container
.ContainerID	ID of the container
.ContainerName	Name of the container
.Image	Name of the image
.Policy	Auto-update policy of the container
.Unit	Name of the systemd unit
.Updated	Update status: true,false,failed,pending

--rollback¶

If restarting a systemd unit after updating the image has failed, rollback to using the previous image and restart the unit another time. Default is true.

Note that detecting if a systemd unit has failed is best done by the container sending the READY message via SDNOTIFY. This way, restarting the unit waits until having received the message or a timeout kicked in. Without that, restarting the systemd unit may succeed even if the container has failed shortly after.

For a container to send the READY message via SDNOTIFY it must be created with the --sdnotify=container option (see podman-run(1)). The application running inside the container can then execute systemd-notify --ready when ready or use the sdnotify bindings of the specific programming language (e.g., sd_notify(3)).

--tls-verify¶

Require HTTPS and verify certificates when contacting registries (default: true). If explicitly set to true, TLS verification is used. If set to false, TLS verification is not used. If not specified, TLS verification is used unless the target registry is listed as an insecure registry in containers-registries.conf(5)

EXAMPLES¶

Create a Quadlet file configured for auto updates:

$ cat ~/.config/containers/systemd/sleep.container
[Container]
Image=registry.fedoraproject.org/fedora:latest
Exec=sleep infinity
AutoUpdate=registry

Generate a systemd service from the Quadlet file by reloading the systemd user daemon:

$ systemctl --user daemon-reload

Start the systemd service and make sure the container is running

$ systemctl --user start sleep.service
$ podman ps
CONTAINER ID  IMAGE                                     COMMAND         CREATED        STATUS        PORTS       NAMES
f8e4759798d4  registry.fedoraproject.org/fedora:latest  sleep infinity  2 seconds ago  Up 2 seconds              systemd-sleep

Check if a new image is available via --dry-run:

$ podman auto-update --dry-run --format "{{.Image}} {{.Updated}}"
registry.fedoraproject.org/fedora:latest   pending

Update the service:

$ podman auto-update
UNIT           CONTAINER                     IMAGE                                     POLICY      UPDATED
sleep.service  f8e4759798d4 (systemd-sleep)  registry.fedoraproject.org/fedora:latest  registry    true

Update containers belonging to a deployment, selected by a container label:

$ podman auto-update --filter label=com.example.deploy.repository=example/web

Check for updates to containers using a particular image:

$ podman auto-update --dry-run --filter ancestor=quay.io/example/web:latest