podman-run(1)

📦
version 6.1.4 · 8 versions
, in arch rolling, nixpkgs unstable

🫶 podman-run(1) 🫶

General Commands Manual

🌸 NAME 🌸

podman-run — run a command in a new container 💖

✨ SYNOPSIS ✨

podman run [options] image [command [arg ...]]

podman container run [options] image [command [arg ...]]

🦋 DESCRIPTION 🦋

Okay bestie, so podman run runs a process in a brand-new container. It starts that process with its own file system, its own networking, and its own isolated process tree — like its own tiny universe, no cap.

The image that starts the process may define defaults for the process that will be run in the container, the networking to expose, and more — but podman run gives the final say to the operator or administrator who starts the container from the image. That’s literally why podman run has more options than any other Podman command. It’s the main-character command 👑.

If the image is not already loaded, then podman run will pull the image, and all image dependencies, from the repository in the same way as running podman pull image, before it starts the container from that image.

Several files get automatically created inside the container. These include /etc/hosts, /etc/hostname, and /etc/resolv.conf to manage networking. They’re based on the host’s version of the files, though you can customize them with options — for example, --dns may override the host’s DNS servers in the created resolv.conf.

Additionally, a container environment file is created in each container to tell programs they’re running in a container. This file is located at /run/.containerenv (or /var/run/.containerenv for FreeBSD containers). When you use the --privileged flag, the .containerenv contains name/value pairs indicating the container engine version, whether the engine is running in rootless mode, the container name and ID, and the image name and ID that the container is based on.

⚠️ Note: /run/.containerenv will not be created when a volume is mounted on /run.

When running from a user defined network namespace, the /etc/netns/NSNAME/resolv.conf will be used if it exists; otherwise /etc/resolv.conf will be used.

Default settings are defined in containers.conf. Most settings for remote connections use the server’s containers.conf, except when documented in man pages.

🎀 IMAGE 🎀

The image is specified using transport:path format. If no transport is specified, the docker (container registry) transport is used by default.

💡 For remote Podman, including Mac and Windows (excluding WSL2) machines, docker is the only allowed transport.
  • dir:path — an existing local directory path storing the manifest, layer tarballs, and signatures as individual files. This is a non-standardized format, primarily useful for debugging or noninvasive container inspection.
$ podman save --format docker-dir fedora -o /tmp/fedora
$ podman run dir:/tmp/fedora echo hello
  • docker://docker-reference (Default) — an image reference stored in a remote container image registry. Example: "quay.io/podman/stable:latest". The reference can include a path to a specific registry; if it does not, the registries listed in registries.conf are queried to find a matching image. By default, credentials from podman login (stored at $XDG_RUNTIME_DIR/containers/auth.json by default) are used to authenticate; otherwise it falls back to using credentials in $HOME/.docker/config.json.
$ podman run registry.fedoraproject.org/fedora:latest echo hello
  • docker-archive:path[:docker-reference] — an image stored in the docker save formatted file. docker-reference is only used when creating such a file, and it must not contain a digest.
$ podman save --format docker-archive fedora -o /tmp/fedora
$ podman run docker-archive:/tmp/fedora echo hello
⚠️ Note: On remote clients, including Mac and Windows (excluding WSL2) machines, this transport is not supported.
  • docker-daemon:docker-reference — an image in docker-reference format stored in the docker daemon internal storage. The docker-reference can also be an image ID (docker-daemon:algo:digest).
$ sudo docker pull fedora
$ sudo podman run docker-daemon:docker.io/library/fedora echo hello
⚠️ Note: On remote clients, including Mac and Windows (excluding WSL2) machines, this transport is not supported.
  • oci-archive:path:tag — an image in a directory compliant with the “Open Container Image Layout Specification” at the specified path and specified with a tag.
$ podman save --format oci-archive fedora -o /tmp/fedora
$ podman run oci-archive:/tmp/fedora echo hello
⚠️ Note: On remote clients, including Mac and Windows (excluding WSL2) machines, this transport is not supported.

💅 OPTIONS 💅

🌈 --add-host=hostname[;hostname[;...]]:ip

Add a custom host-to-IP mapping to the container’s /etc/hosts file.

The option takes one or multiple semicolon-separated hostnames to be mapped to a single IPv4 or IPv6 address, separated by a colon. It can also be used to overwrite the IP addresses of hostnames Podman adds to /etc/hosts by default — also see the --name and --hostname options. This option can be specified multiple times to add additional mappings to /etc/hosts. It conflicts with the --no-hosts option and conflicts with no_hosts=true in containers.conf.

Instead of an IP address, the special flag host-gateway can be given. This resolves to an IP address the container can use to connect to the host. The IP address chosen depends on your network setup, so there’s no guarantee that Podman can determine the host-gateway address automatically — which will then cause Podman to fail with an error message. You can overwrite this IP address using the host_containers_internal_ip option in containers.conf.

The host-gateway address is also used by Podman to automatically add the host.containers.internal and host.docker.internal hostnames to /etc/hosts. You can prevent that by either giving the --no-hosts option, or by setting host_containers_internal_ip="none" in containers.conf. If no host-gateway address was configured manually and Podman fails to determine the IP address automatically, Podman will silently skip adding these internal hostnames to /etc/hosts. If Podman is running in a virtual machine using podman machine (this includes Mac and Windows hosts), Podman will silently skip adding the internal hostnames to /etc/hosts, unless an IP address was configured manually; the internal hostnames are resolved by the gvproxy DNS resolver instead.

Podman will use the /etc/hosts file of the host as a basis by default — i.e. any hostname present in this file will also be present in the /etc/hosts file of the container. A different base file can be configured using the base_hosts_file config in containers.conf.

🌟 --annotation=key=value

Add an annotation to the container. This option can be set multiple times.

🦋 --arch=ARCH

Override the architecture, defaults to the host, of the image to be pulled. For example, arm. Unless overridden, subsequent lookups of the same image in the local storage matches this architecture, regardless of the host.

💖 --attach, -a=stdin | stdout | stderr

Attach to STDIN, STDOUT, or STDERR.

In foreground mode (the default when -d is not specified), podman run can start the process in the container and attach the console to the process’s standard input, output, and error. It can even pretend to be a TTY — this is what most command-line executables expect — and pass along signals. The -a option can be set for each of stdin, stdout, and stderr.

🎀 --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.

🍓 --blkio-weight=weight

Block IO relative weight. The weight is a value between 10 and 1000.

This option is not supported on cgroups V1 rootless systems.

🧚 --blkio-weight-device=device:weight

Block IO relative device weight.

🔥 --cap-add=capability

Add Linux capabilities.

Granting additional capabilities increases the privileges of the processes running inside the container and potentially allows it to break out of confinement. Capabilities like CAP_SYS_ADMIN, CAP_SYS_PTRACE, CAP_MKNOD, and CAP_SYS_MODULE are particularly dangerous when they are not used within a user namespace. Please refer to user_namespaces(7) for a more detailed explanation of the interaction between user namespaces and capabilities.

Before adding any capability, review its security implications and ensure it is really necessary for the container’s functionality. See capabilities(7) for more information.

💀 --cap-drop=capability

Drop these capabilities from the default podman capability set, or all to drop all capabilities.

This is a space separated list of capabilities.

💜 --cert-dir=path

Use certificates at path (*.crt, *.cert, *.key) to connect to the registry. Default: /etc/containers/certs.d. For details, see containers-certs.d(5).

⚠️ This option is not available with the remote Podman client, including Mac and Windows (excluding WSL2) machines.

🌈 --cgroup-conf=KEY=VALUE

When running on cgroup v2, specify the cgroup file to write to and its value. For example, --cgroup-conf=memory.high=1073741824 sets the memory.high limit to 1GB.

🌸 --cgroup-parent=path

Path to cgroups under which the cgroup for the container is created. If the path is not absolute, the path is considered to be relative to the cgroups path of the init process. Cgroups are created if they do not already exist.

✨ --cgroupns=mode

Set the cgroup namespace mode for the container.

  • host: use the host’s cgroup namespace inside the container.
  • container:id: join the namespace of the specified container.
  • private: create a new cgroup namespace.
  • ns:path: join the namespace at the specified path.

If the host uses cgroups v1, the default is set to host. On cgroups v2, the default is private.

🦋 --cgroups=how

Determines whether the container creates cgroups.

Default is enabled.

The enabled option creates a new cgroup under the cgroup-parent. The disabled option forces the container to not create cgroups, and thus conflicts with cgroup options (--cgroupns and --cgroup-parent). The no-conmon option disables a new cgroup only for the conmon process. The split option splits the current cgroup in two sub-cgroups: one for conmon and one for the container payload. It is not possible to set --cgroup-parent with split.

  • --chrootdirs=path — bestie, this one's giving "extra home lol". It's a path inside the container that gets treated like a chroot dir: literally any Podman-managed file (think /etc/resolv.conf, /etc/hosts, /etc/hostname) that gets mounted into the root directory also gets mounted in there too. Need more than one? Separate them with a comma. Slay.
  • --cidfile=file — writes the container ID to file. That file peaces out along with the container… with one lil exception: it sticks around when you're using podman --remote run on detached containers. Noted, no cap.
  • --conmon-pidfile=file — drops the pid of the conmon process into a file. Why do we care? Because conmon runs in a whole separate process from Podman, so if you want systemd to restart Podman containers you need this. 💅
Heads up: this one is not available with the remote Podman client — that includes Mac and Windows (excluding WSL2) machines. Rip.
This option is not supported on cgroups V1 rootless systems. 😭
Not supported on cgroups V1 rootless systems.
  • --cpu-rt-period=microseconds — limits the CPU real-time period, in microseconds. This tells the kernel to restrict the container's Real Time CPU usage to whatever period you specified. Iconic.
This option is only supported on cgroups V1 rootful systems. Like, only that. Don't come at me.
  • --cpu-rt-runtime=microseconds — limits the CPU real-time runtime, in microseconds. This tells the kernel to limit how much time Real Time tasks can consume within a given CPU period. Example straight from the source: a period of 1,000,000us and a runtime of 950,000us means the container can eat 95% of available CPU and leaves the remaining 5% for normal priority tasks. Leave some for the girlies!! 🫶 Also: the sum of all runtimes across containers cannot exceed the amount allotted to the parent cgroup.
Again: only supported on cgroups V1 rootful systems.
  • --cpu-shares, -c=shares — CPU shares, aka relative weight. By default every container gets the same proportion of CPU cycles, but you can modify that proportion by changing a container's CPU share weighting relative to the combined weight of all the running containers. Default weight is 1024. 🎀 Plot twist: the proportion only applies when CPU-intensive processes are running. When tasks in one container are idle, the other containers get to use the leftover CPU time. So the actual amount of CPU time varies depending on how many containers are running on the system — it's giving ✨circumstantial✨.

Example time, because this is the part everyone asks about: say you've got three containers — one with a cpu-share of 1024 and two others with a cpu-share of 512. When processes in all three containers try to use 100% of CPU, the first container receives 50% of the total CPU time. Now add a fourth container with a cpu-share of 1024, and the first container only gets 33% of the CPU. The remaining containers get 16.5%, 16.5% and 33%. Math ate, honestly.

On a multi-core system, the shares of CPU time get distributed over all CPU cores. So even if a container is limited to less than 100% of CPU time, it can still use 100% of each individual CPU core. Main character behavior tbh.

For example, consider a system with more than three cores. If container C0 is started with --cpu-shares=512 running one process, and another container C1 with --cpu-shares=1024 running two processes, this can result in the following division of CPU shares:

┌─────┬───────────┬─────┬──────────────┐
│ PID │ container │ CPU │ CPU share    │
├─────┼───────────┼─────┼──────────────┤
│ 100 │ C0        │ 0   │ 100% of CPU0 │
├─────┼───────────┼─────┼──────────────┤
│ 101 │ C1        │ 1   │ 100% of CPU1 │
├─────┼───────────┼─────┼──────────────┤
│ 102 │ C1        │ 2   │ 100% of CPU2 │
└─────┴───────────┴─────┴──────────────┘

You already know the drill: on some systems changing resource limits isn't allowed for non-root users, more details here https://github.com/containers/podman/blob/main/troubleshooting.md#26-running-containers-with-resource-limits-fails-with-a-permissions-error

Not supported on cgroups V1 rootless systems. 🦋
Not supported on cgroups V1 rootless systems.
Not supported on cgroups V1 rootless systems.
Not supported on cgroups V1 rootless systems.
  • --creds=[username[:password]] — the [username[:password]] to use when authenticating with the registry, if required. If one or both values aren't supplied, a command line prompt pops up and you can type the value in — and the password gets entered without echo, so nobody's shoulder-surfing your secrets. 🙈
Real talk though: the credentials you specify here are only used to authenticate against target registries. They are not used for mirrors or when the registry gets rewritten (see containers-registries.conf(5)). For those, consider using a containers-auth.json(5) file instead. Iykyk.
  • --decryption-key=key[:passphrase] — the [key[:passphrase]] used to decrypt images. The key can point to keys and/or certificates. Decryption is tried with all keys. And if the key is protected by a passphrase, you're required to pass it in the argument — omit it otherwise. 🔥
  • --detach, -d — detached mode, babes: run the container in the background and print the new container ID. Default is false. At any time, run podman ps in the other shell to view a list of the running containers. Want back in? Reattach to a detached container with the podman attach command. 🫶 And when you're attached via tty mode, you detach from the container (and leave it running!) using a configurable key sequence. Default sequence is ctrl-p,ctrl-q. Specify your own with the --detach-keys option, or configure it in the containers.conf file — see containers.conf(5) for more info.
  • --detach-keys=sequence — specifies the key sequence for detaching a container. Format is either a single character [a-Z] or one or more ctrl-<value> characters where <value> is one of: a-z, @, ^, [, , or _. Specifying "" disables this feature entirely. Default is ctrl-p,ctrl-q. 🎀 This option can also be set in the containers.conf(5) file.
  • --device=host-device[:container-device][:permissions] — adds a host device to the container. The format of this is HOST-DEVICE[:CONTAINER-DEVICE][:PERMISSIONS], where HOST-DEVICE is the path of the device node on the host, CONTAINER-DEVICE is the path of the device node in the container, and PERMISSIONS is a list of permissions combining 'r' for read, 'w' for write, and 'm' for mknod(2). Example: --device=/dev/sdc:/dev/xvdc:rwm. Note: if host-device is a symbolic link, it gets resolved first — the container only stores the major and minor numbers of the host device. Sneaky but correct. Podman may load kernel modules required for using the specified device. The devices Podman loads modules for when necessary are: /dev/fuse. 🦋 In rootless mode, the new device gets bind mounted in the container from the host rather than Podman creating it within the container space. And because the bind mount retains its SELinux label on SELinux systems, the container can get permission denied when accessing the mounted device 😭. To fix that, modify SELinux settings to allow containers to use all device labels with this:
$ sudo setsebool -P  container_use_devices=true
Also note: if the user only has access rights via a group, accessing the device from inside a rootless container fails. Use the --group-add keep-groups flag to pass the user's supplementary group access into the container. Save this one, it's a lifesaver fr.
  • --device-cgroup-rule="type major:minor mode" — adds a rule to the cgroup allowed devices list. The rule is expected to be in the format specified in the Linux kernel documentation admin-guide/cgroup-v1/devices:
    • type: a (all), c (char), or b (block)
    • major and minor: either a number, or * for all
    • mode: a composition of r (read), w (write), and m (mknod(2))

Very "build your own permission slip" of it.

That last little note belongs to the option right above it — <pink>still true tho</pink>:

This option is not supported on cgroups V1 rootless systems.

  • --device-read-iops=path:rate — caps the read rate in IO operations per second coming off a device, like --device-read-iops=/dev/sda:1000. 🍓

On some systems, changing the resource limits may not be allowed for non-root users. For more details, see https://github.com/containers/podman/blob/main/troubleshooting.md#26-running-containers-with-resource-limits-fails-with-a-permissions-error

This option is not supported on cgroups V1 rootless systems.

  • --device-write-bps=path:rate — caps the write rate in bytes per second going to a device — for example --device-write-bps=/dev/sda:1mb. ✨

On some systems, changing the resource limits may not be allowed for non-root users. For more details, see https://github.com/containers/podman/blob/main/troubleshooting.md#26-running-containers-with-resource-limits-fails-with-a-permissions-error

This option is not supported on cgroups V1 rootless systems.

  • --device-write-iops=path:rate — caps the write rate in IO operations per second going to a device, e.g. --device-write-iops=/dev/sda:1000. 🦋

On some systems, changing the resource limits may not be allowed for non-root users. For more details, see https://github.com/containers/podman/blob/main/troubleshooting.md#26-running-containers-with-resource-limits-fails-with-a-permissions-error

This option is not supported on cgroups V1 rootless systems.

  • --disable-content-trust — ok so this one is a whole nothingburger 💀 it's a Docker-specific option to disable image verification to a container registry, and Podman does not support it. It's a NOOP (literally does nothing) and exists solely for scripting compatibility. So if a script passes it, fine, nothing happens. That's the whole thing.
  • --dns=ipaddr — set custom DNS servers. 🧚

You'd use this to override the DNS configuration passed to the container. Typically this is necessary when the host DNS configuration is invalid for the container (like, say, 127.0.0.1 — which means nothing from inside the container, lowkey useless). When that's the case, the --dns flag is necessary for every run. No skipping, bestie.

The special value none can be specified to disable creation of /etc/resolv.conf in the container by Podman. Then the /etc/resolv.conf file in the image gets used without changes. 🫶

Note that ipaddr may be added directly to the container's /etc/resolv.conf. This is not guaranteed though. For example, passing a custom network whose dns_enabled is set to true to --network will result in /etc/resolv.conf only referring to the aardvark-dns server. aardvark-dns then forwards to the supplied ipaddr for all non-container name queries. iykyk 🌈

This option cannot be combined with --network that is set to none or container:id.
  • --dns-option=option — set custom DNS options. 💜

Invalid if using --dns-option with --network that is set to none or container:id.

  • --dns-search=domain — set custom DNS search domains. 🌸

Invalid if using --dns-search with --network that is set to none or container:id.

Use --dns-search=. to remove the search domain. (Period. Literally just a dot. iconic)

  • --entrypoint="command" | '["command", arg1 , ...]' — override the default ENTRYPOINT from the image. 👑

So the ENTRYPOINT of an image is similar to a COMMAND, because it specifies what executable to run when the container starts — but it's (purposely) more difficult to override. The ENTRYPOINT gives a container its default nature or behavior. When the ENTRYPOINT is set, the container runs as if it were that binary, complete with default options. More options can be passed in via the COMMAND. But, if a user wants to run something else inside the container, the --entrypoint= option allows a new ENTRYPOINT to be specified. main character energy, truly.

Specify multi option commands in the form of a JSON string.

  • --env, -e=env — set environment variables. 💅

This option allows arbitrary environment variables that are available for the process to be launched inside of the container. Here's the cute part: if an environment variable is specified without a value, Podman checks the host environment for a value and sets the variable only if it is set on the host. And as a special case 💖 — if an environment variable ending in * is specified without a value, Podman searches the host environment for variables starting with the prefix and adds those variables to the container. Chef's kiss.

See ⟨#environment⟩ note below for precedence and examples.

  • --env-file=file — read in a line-delimited file of environment variables. 📄

See ⟨#environment⟩ note below for precedence and examples.

  • --env-host — use host environment inside of the container. 🏠

See Environment note below for precedence.

(This option is not available with the remote Podman client, including Mac and Windows (excluding WSL2) machines)
  • --env-merge=env — preprocess default environment variables for the containers. 🎀

For example, if the image contains the environment variable hello=world, the user can preprocess it using --env-merge hello=${hello}-some so the new value is hello=world-some.

Please note 👀 that if the environment variable hello is not present in the image, then it'll be replaced by an empty string — and so using --env-merge hello=${hello}-some would result in the new value of hello=-some. Notice the leading - delimiter. Yeah. Watch for that.

  • --expose=port[/protocol] — expose a port or a range of ports (like --expose=3300-3310). 🔥

The protocol can be tcp, udp or sctp and if not given tcp is assumed. This option matches the EXPOSE instruction for image builds, and it has no effect on the actual networking rules unless -P/--publish-all is used to forward to all exposed ports from random host ports. To forward specific ports from the host into the container, use the -p/--publish option instead. tysm for coming to my TED talk 🫶

  • --gidmap=[flags]container_gid:from_gid[:amount] — run the container in a new user namespace using the supplied GID mapping. 🌟

This option conflicts with the --userns and --subgidname options. It gives you a way to map host GIDs to container GIDs in the same way as --uidmap maps host UIDs to container UIDs. For details see --uidmap.

Note: the --gidmap option cannot be called in conjunction with the --pod option, as a gidmap cannot be set on the container level when in a pod.
  • --gpus=ENTRY — start the container with GPU support. 🚀

Where ENTRY can be all to request all GPUs, or a vendor-specific identifier. Currently, NVIDIA and AMD devices are supported. If both NVIDIA and AMD devices are present, the NVIDIA devices will be preferred, and a CDI device name must be specified using the --device flag to request a set of GPUs from a specific vendor. slay 💅

  • --group-add=group | keep-groups — assign additional groups to the primary user running within the container process. 🌈
    • keep-groups is a special flag that tells Podman to keep the supplementary group access.

Allows the container to use the user's supplementary group access. If file systems or devices are only accessible by the rootless user's group, this flag tells the OCI runtime to pass the group access into the container. Currently only available with the crun OCI runtime. Note: keep-groups is exclusive — other groups cannot be specified with this flag. (Not available for remote commands, including Mac and Windows (excluding WSL2) machines)

  • --group-entry=ENTRY — customize the entry that is written to the /etc/group file within the container when --user is used. 📝

The variables $GROUPNAME, $GID, and $USERLIST are automatically replaced with their value at runtime if present. cute magic ✨

  • --health-cmd="command" | '["command", arg1 , ...]' — set or alter a healthcheck command for a container. 💖

The command is a command to be executed inside the container that determines the container health. The command is required for other healthcheck options to be applied. A value of none disables existing healthchecks.

Multiple options can be passed in the form of a JSON array; otherwise, the command is interpreted as an argument to /bin/sh -c.

Note: The default values are used even if healthcheck is configured in the image.
  • --health-interval=interval — set an interval for the healthchecks. ⏰

An interval of disable results in no automatic timer setup. The default is 30s.

Note: This parameter will overwrite related healthcheck configuration from the image.
  • --health-log-destination=directory_path — set the destination of the HealthCheck log. 🌸

Directory path, local or events_logger (local uses container state file) (Default: local)

  • local: (default) HealthCheck logs are stored in overlay containers. (For example: $runroot/healthcheck.log)
  • directory: creates a log file named <container-ID>-healthcheck.log with HealthCheck logs in the specified directory.
  • events_logger: The log will be written with logging mechanism set by events_logger. It also saves the log to a default directory, for performance on a system with a large number of logs.
  • --health-max-log-count=number of stored logs — set maximum number of attempts in the HealthCheck log file. 🗂️

('0' value means an infinite number of attempts in the log file) (Default: 5 attempts)

  • --health-max-log-size=size of stored logs — set maximum length in characters of stored HealthCheck log. 📏

("0" value means an infinite log length) (Default: 500 characters)

  • --health-on-failure=action — action to take once the container transitions to an unhealthy state. The default is none. 🩺
    • none: Take no action.
    • kill: Kill the container.
    • restart: Restart the container. Do not combine the restart action with the --restart flag. When running inside of a systemd unit, consider using the kill or stop action instead to make use of systemd's restart policy.
    • stop: Stop the container.
  • --health-retries=retries — the number of retries allowed before a healthcheck is considered to be unhealthy. 🔁

The default value is 3.

Note: This parameter can overwrite the healthcheck configuration from the image.
  • --health-start-period=period — the initialization time needed for a container to bootstrap. 🌱

The value can be expressed in time format like 2m3s. The default value is 0s.

Note: The health check command is executed as soon as a container is started — if the health check is successful the container's health state will be updated to healthy. However, if the health check fails, the health state will stay as starting until either the health check is successful or until the --health-start-period time is over. If the health check command fails after the --health-start-period time is over, the health state will be updated to unhealthy. The health check command is executed periodically based on the value of --health-interval.
Note: This parameter will overwrite related healthcheck configuration from the image.

ok that's the health stuff done — and yes there are more letters after h in the alphabet, we're not done yet 💀

  • --health-startup-cmd="command" / '["command", arg1 , ...]' — okay so this sets a startup healthcheck for your container bestie 💖. It runs inside the container and its whole job is to gate the regular healthcheck — like a bouncer at the door of the healthcheck club 🚪✨. When the startup command succeeds, the regular healthcheck kicks in and the startup one ceases (iconic behavior, honestly). Optionally, if the command fails a set number of times, the container gets restarted. Why would you want this? Because containers with a long startup period shouldn't get marked unhealthy before they're actually fully started — that's just rude 😭. Also super important: startup healthchecks can only be used when a regular healthcheck (from the container's image or the --health-cmd option) is also set. No regular healthcheck, no startup healthcheck. Periodt.
  • --health-startup-interval=interval — sets the interval for the startup healthcheck ⏰. If you set it to disable, no automatic timer setup happens. Default? 30s.
  • --health-startup-retries=retries — how many attempts are allowed before the startup healthcheck restarts the container 💀. If you set it to 0, the container never gets restarted. Default is 0.
  • --health-startup-success=retries — the number of successful runs required before the startup healthcheck succeeds and the regular healthcheck begins 🎀. A value of 0 means any single success starts the regular healthcheck. Default is 0.
  • --health-startup-timeout=timeout — the max time a startup healthcheck command has to complete before it's marked as failed ⏳. You can express the value in a time format like 2m3s. Default value is 30s.
  • --health-timeout=timeout — the maximum time allowed to complete the healthcheck before an interval is considered failed 😬. Just like start-period, the value can be expressed in a time format such as 1m22s. The default value is 30s.
⚠️ Note: a timeout marks the healthcheck as failed. And if the healthcheck command itself runs longer than the specified timeout, it gets sent a SIGKILL signal. Brutal but fair 💅.
⚠️ Note: this parameter will overwrite related healthcheck configuration from the image. So yeah, your image's settings? Gone. Overwritten. That's the vibe.
  • --help — print the usage statement 🫶. The classic "what do I even do here" button.
  • --hostname, -h=name — sets the container's hostname inside the container ✨.

This option can only be used with a private UTS namespace — --uts=private (which is the default). If --pod is given and the pod shares the same UTS namespace (default), the pod's hostname gets used instead. The hostname you give is also added to the /etc/hosts file using the container's primary IP address (also peep the --add-host option) 🎀.

When --hostname is not used and the container uses a private UTS namespace (default), Podman sets the hostname to the first 12 characters of the container ID 🌸. The container name assigned with --name is not used unless container_name_as_hostname=true is set in containers.conf. Iykyk.

Podman network DNS registers the container name, the short container ID (first 12 characters), and any explicitly set --hostname as DNS names 🔥. The default hostname matches the short ID alias. See podman-network(1).

  • --hosts-file=path | none | image — the base file used to create the /etc/hosts file inside the container 📄. This must either be an absolute path to a file on the host system, or one of these special flags:
    • "" — follow the base_hosts_file configuration in containers.conf (this is the default)
    • none — don't use a base file (i.e. start with an empty file)
    • image — use the container image's /etc/hosts file as the base file
  • --hostuser=name — adds a user account to /etc/passwd from the host into the container 🧑‍🤝‍🧑. The username or UID must exist on the host system. No imaginary friends allowed 🚫.
  • --http-proxy — so by default, proxy environment variables get passed into the container if they're set for the Podman process ✨. You can disable that by setting the value to false. The env vars passed in include http_proxy, https_proxy, ftp_proxy, no_proxy, plus the uppercase versions of those. This option is only needed when the host system must use a proxy but the container doesn't use any proxy 🤷‍♀️. Proxy environment variables specified for the container in any other way override the values passed through from the host. (Other ways to specify the proxy for the container include passing the values with the --env flag, or hard coding the proxy environment at container build time.) When used with the remote client, it uses the proxy environment variables set on the server process 🖥️.

Defaults to true. Slay.

  • --image-volume=bind | tmpfs | ignore — tells Podman how to handle the builtin image volumes 🎀. Default is bind.
    • bind: an anonymous named volume is created and mounted into the container.
    • tmpfs: the volume is mounted onto the container as a tmpfs, which lets users create content that disappears when the container is stopped 🫧 (poof, gone).
    • ignore: all volumes are just ignored and no action is taken. Nothing happens. Zero. Nada.
  • --init — run an init inside the container that forwards signals and reaps processes 👑. The container-init binary is mounted at /run/podman-init.
🚨 Mounting over /run breaks container execution. Like, do not do it. Consider yourself warned 💀.
  • --init-path=path — path to the container-init binary 📍. In case the default one isn't where you want it.
  • --interactive, -i — when set to true, make stdin available to the contained process 💬. If false, the stdin of the contained process is empty and immediately closed. Interesting.

If attached, stdin is piped to the contained process. If detached, reading stdin will block until later attached 😴.

⚠️ Caveat: Podman will consume input from stdin as soon as it becomes available, even if the contained process doesn't request it. It's thirsty like that 🙈.
  • --ip=IPv4 — specify a static IPv4 address for the container, for example 10.88.64.128 📍. This option can only be used if the container is joined to only a single network — i.e. --network=network-name is used at most once — and if the container isn't joining another container's network namespace via --network=container:id. The address must be within the network's IP address pool (default 10.88.0.0/16). No cap.

To specify multiple static IP addresses per container, use the --network option with multiple comma-separated ip values:

--network mynet:ip=10.88.0.10,ip=10.88.0.11,ip=10.88.0.12

This assigns multiple static IPv4 addresses (10.88.0.10, 10.88.0.11, 10.88.0.12) to the same network interface 🔥.

Multi-Subnet Networks: when a network has multiple subnets, you can assign IPs from different subnets to the same container 🌈. The IPs get applied to a single network interface, with the first IP as primary and additional IPs as secondary addresses.

IP Assignment Order: for multi-subnet networks, IPs are grouped and ordered by their corresponding subnet, following the order in which subnets were defined during network creation (via --subnet flags) ✨. The order you specify IPs in the command does not affect the final assignment order. Example:

podman network create --subnet 10.89.0.0/24 --subnet 10.90.0.0/24 mynet
podman run --network mynet:ip=10.90.0.20,ip=10.89.0.10,ip=10.89.0.11 alpine

Results in IPs ordered by subnet: 10.89.0.10 (primary), 10.89.0.11 (secondary), 10.90.0.20 (secondary), since 10.89.0.0/24 was defined first 👑

Dynamic Allocation: if fewer IPs are specified than available subnets, the remaining subnets receive dynamically allocated IPs 🦋. Dynamic IPs are assigned in subnet order after all static IPs are applied.

Example with a multi-subnet network:

podman network create --subnet 10.89.0.0/24 --subnet 10.90.0.0/24 mynet
podman run --network mynet:ip=10.89.0.10,ip=10.90.0.20 alpine

This configures eth0 with 10.89.0.10 (primary) and 10.90.0.20 (secondary). Chef's kiss 💅

  • --ip6=IPv6 — specify a static IPv6 address for the container, for example fd46:db93:aa76:ac37::10 🌸. This option can only be used if the container is joined to only a single network — i.e. --network=network-name is used at most once — and if the container isn't joining another container's network namespace via --network=container:id. The address must be within the network's IPv6 address pool.

To specify multiple static IPv6 addresses per container, set multiple networks using the --network option with a static IPv6 address specified for each one using the ip6 mode for that option 🎀.

  • --ipc=ipc — set the IPC namespace mode for a container 🗣️. The default is to create a private IPC namespace.
    • "": use Podman's default, defined in containers.conf.
    • container:id: reuses another container's shared memory, semaphores, and message queues.
    • host: use the host's shared memory, semaphores, and message queues inside the container.
⚠️ Note: the host mode gives the container full access to local shared memory and is therefore considered insecure. Just so we're clear 💀.
  • none: private IPC namespace, with /dev/shm not mounted.
  • ns:path: path to an IPC namespace to join.
  • private: private IPC namespace.
  • shareable: private IPC namespace with a possibility to share it with other containers.
  • --label, -l=key=value — add metadata to a container 🏷️. Little name tags but for vibes.
  • --label-file=file — read in a line-delimited file of labels 📋. Bulk label energy.
  • --link-local-ip=ip — not implemented. Yeah. Nothing here. Moving on 🫠.
  • --log-driver=driver — the logging driver for the container 📝. Currently available options are k8s-file, journald, none, passthrough and passthrough-tty, with json-file aliased to k8s-file for scripting compatibility. Default is journald.

The podman info command below displays the default log-driver for the system:

$ podman info --format '{{ .Host.LogDriver }}'
journald

The passthrough driver passes down the standard streams (stdin, stdout, stderr) to the container 🌊. It is not allowed with the remote Podman client — including Mac and Windows (excluding WSL2) machines — and it's not allowed on a tty, since it's vulnerable to attacks via TIOCSTI 🚨. Dangerous behavior, be careful.

The passthrough-tty driver is the same as passthrough except that it also allows it to be used on a TTY if the user really wants it 🙈.

  • --log-opt=name=value — logging driver specific options 🎛️.

Set custom logging configuration. These *name*s are supported:

  • path: specify a path to the log file (e.g. --log-opt path=/var/log/container/mycontainer.json);
  • max-size: specify a max size of the log file (e.g. --log-opt max-size=10mb);
  • tag: specify a custom log tag for the container (e.g. --log-opt tag="{{.ImageName}}". It supports the same keys as podman inspect --format. This option is currently supported only by the journald log driver.)
  • label: specify a custom log label for the container (e.g. --log-opt label="CONTAINER_IMAGE={{.ImageName}}". It supports the same keys as podman inspect --format. This option can be repeated multiple times. This option is currently supported only by the journald log driver.)
  • --mac-address=address — container network interface MAC address (e.g. 92:d0:c6:0a:29:33) 🦋. This option can only be used if the container is joined to only a single network — i.e. --network=network-name is used at most once — and if the container isn't joining another container's network namespace via --network=container:id.

Remember bestie: the MAC address on an Ethernet network has to be unique, no cap. And the IPv6 link-local address? It's built from the device's MAC address according to RFC4862. ✨

Wanna give a container multiple static MAC addresses? Set up multiple networks with the --network option, and specify a static MAC for each one using the mac mode for that option. That's it, that's the trick. 🎀

  • *--memory, -m=number[unit]* ** — the memory limit, i.e. the hard ceiling. The unit can be b (bytes), k (kibibytes), m (mebibytes), or g (gibibytes).

This is your girl for constraining how much memory a container gets. If the host supports swap memory, then the --memory setting can actually be larger than physical RAM — yes really. If a limit of 0 is specified (i.e. you're not using --memory), then the container's memory is not limited at all. Heads up though: the actual limit may get rounded up to a multiple of the operating system's page size (and that value is very large — we're talking millions of trillions 😭).

This option is not supported on cgroups V1 rootless systems. 💀
  • --memory-reservation=number[unit] ** — the memory *soft limit. *Unit can be b (bytes), k (kibibytes), m (mebibytes), or g (gibibytes).

So after you set a memory reservation, when the system detects memory contention or low memory, containers get forced to restrict their consumption down to their reservation. Which means: always set this value below --memory, otherwise the hard limit takes precedence and your reservation is basically doing nothing. By default, memory reservation is the same as the memory limit.

This option is not supported on cgroups V1 rootless systems. 💀
  • --memory-swap=number[unit] ** — a limit value equal to memory *plus swap. *Unit can be b (bytes), k (kibibytes), m (mebibytes), or g (gibibytes).

Must be used with the -m (--memory) flag, okay? And the argument value must be larger than that of -m (--memory). By default, it's set to double the value of --memory.

Set number to -1 to enable unlimited swap. 🌀

This option is not supported on cgroups V1 rootless systems. 💀
  • *--memory-swappiness=number* ** — tunes a container's memory swappiness behavior. Accepts an integer between 0 and 100.
This flag is only supported on cgroups V1 rootful systems. So if you're on anything else, it's giving "not for you" 🙈

🦋 --mount=type=TYPE,TYPE-SPECIFIC-OPTION[,...] 🦋

Attach a filesystem mount to the container. This is the fancy, explicit one — it's the main character of mount options. 👑

The currently supported mount TYPEs are: artifact, bind, devpts, glob, image, ramfs, tmpfs and volume.

Options common to all mount types 🌈

  • src, source — the mount source spec for bind, glob, and volume. Mandatory for artifact, bind, glob, and image. Optional for volume — if you omit it, an anonymous volume gets created.
  • dst, dest, destination, target — the mount destination spec.

Okay so, when source globs are specified without the destination directory, the files and directories are mounted with their complete path within the container. When the destination is specified, the files and directories matching the glob on the base file name on the destination directory are mounted. So the option type=glob,src=/foo*,destination=/tmp/bar tells container engines to mount host files matching /foo* to the /tmp/bar/ directory in the container. Slay. ✨

Options specific to type=artifact 🎀

  • digest — if the artifact source contains multiple blobs, you can specify a digest to only mount that one specific blob with the digest.
  • title — if the artifact source contains multiple blobs, a title can be set, which gets compared against the org.opencontainers.image.title annotation.
  • name — this one lets you overwrite the filename used inside the container for mounting. On a single blob artifact, the name is used as-is if dst is a directory, and otherwise ignored. With a multi blob artifact, the name will be used with an index suffix <name>-x where x is the layer index in the artifact, starting with 0.

The src argument contains the name of the artifact, which must already exist locally. The dst argument contains the target path — if the path in the container is a directory, the blob title (org.opencontainers.image.title annotation) will be used as the filename and joined to the path. If the annotation doesn't exist, the digest gets used as the filename instead. This results in all blobs of the artifact being mounted into the container at the given path. 💅

However! If the dst path is an existing file in the container, then the blob will be mounted directly on it. This only works when the artifact contains a single blob, or when either digest or title are specified.

And if the dst path does not already exist in the container, then: if the artifact contains a single blob, it behaves like the existing-file case and mounts directly to that path. If the artifact has more than one blob, it works like the existing-directory case and mounts each blob as a file within the dst path.

Options specific to type=volume 🍓

  • ro, readonly — true or false (default if unspecified: false).
  • U, chown — true or false (default if unspecified: false). Recursively changes the owner and group of the source volume based on the UID and GID of the container.
  • subpath — mount only a specific subpath within the volume, instead of the whole entire volume.
  • relabel — shared, private. Recursively walk the mount and set the security system (e.g. SELinux) label on each file, to grant access permissions from the container context. shared applies a label that grants permissions to all containers, while private applies a label that grants permissions to this specific container.
  • z, Z — shorthand for relabel=shared and relabel=private, respectively.
  • idmap — if specified, create an idmapped mount to the target user namespace in the container. The idmap option is only supported by Podman in rootful mode — the Linux kernel does not allow the use of idmapped file systems for unprivileged users. The idmap option supports a custom mapping that can be different from the user namespace used by the container. The mapping can be specified after the idmap option like this: idmap=uids=0-1-10#10-11-10;gids=0-100-10. For each triplet, the first value is the start of the backing file system IDs that are mapped to the second value on the host. The length of this mapping is given in the third value. Multiple ranges are separated with #. If the specified mapping is prepended with a @, then the mapping is considered relative to the container user namespace — the host ID for the mapping is changed to account for the relative position of the container user in the container user namespace.

Options specific to type=image 🦋

  • rw, readwrite — true or false (default if unspecified: false).
  • subpath — mount only a specific path within the image, instead of the whole image.

Options specific to bind and glob 💜

  • ro, readonly — true or false (default if unspecified: false).
  • bind-propagation — shared, slave, private, unbindable, rshared, rslave, runbindable, or rprivate (default). That's footnote [1] — and yeah, go see also mount(2) for the deep lore.
  • bind-nonrecursive — do not set up a recursive bind mount. By default it is recursive.
  • relabel — shared, private. Recursively walk the mount and set the security system (e.g. SELinux) label on each file, to grant access permissions from the container context. shared applies a label that grants permissions to all containers, while private applies a label that grants permissions to this specific container.
  • z, Z — shorthand for relabel=shared and relabel=private, respectively.
  • idmap — if specified, create an idmapped mount to the target user namespace in the container. The idmap option is only supported by Podman in rootful mode. The Linux kernel does not allow the use of idmapped file systems for unprivileged users. The idmap option supports a custom mapping that can be different from the user namespace used by the container. The mapping can be specified after the idmap option like: idmap=uids=0-1-10#10-11-10;gids=0-100-10. For each triplet, the first value is the start of the backing file system IDs that are mapped to the second value on the host. The length of this mapping is given in the third value. Multiple ranges are separated with #. If the specified mapping is prepended with a @, then the mapping is considered relative to the container user namespace. The host ID for the mapping is changed to account for the relative position of the container user in the container user namespace.
  • U, chown — true or false (default if unspecified: false). Recursively change the owner and group of the source volume based on the UID and GID of the container.
  • no-dereference — do not dereference symlinks, but copy the link source into the mount destination.

Options specific to type=tmpfs and ramfs 🌸

  • ro, readonly — true or false (default if unspecified: false).
  • tmpfs-size — size of the tmpfs/ramfs mount, in bytes. Unlimited by default in Linux.
  • tmpfs-mode — octal file mode of the tmpfs/ramfs (e.g. 700 or 0700.).
  • tmpcopyup — enable copyup from the image directory at the same location to the tmpfs/ramfs. Used by default.

When the tmpfs destination is inside a volume or bind mount, files from the parent mount are also copied into the tmpfs, so the parent content remains visible. Use notmpcopyup to mount an empty tmpfs that shadows the parent mount's subtree.

  • noatime — disable updating file access times when the file is read.
  • notmpcopyup — disable copying files from the image to the tmpfs/ramfs.

Use this option when mounting a tmpfs inside a volume or bind mount, to ensure the tmpfs properly shadows the parent mount's subtree.

  • U, chown — true or false (default if unspecified: false). Sets the uid and gid options for the tmpfs filesystem based on the UID and GID of the container. Heads up: this is not recursive.

Options specific to type=devpts 🧚

  • uid — numeric UID of the file owner (default: 0).
  • gid — numeric GID of the file owner (default: 0).
  • mode — octal permission mask for the file (default: 600).
  • max — maximum number of PTYs (default: 1048576).

✨ Examples, because you know you want them ✨

  • type=bind,source=/path/on/host,destination=/path/in/container
  • type=bind,src=/path/on/host,dst=/path/in/container,relabel=shared
  • type=bind,src=/path/on/host,dst=/path/in/container,relabel=shared,U=true
  • type=bind,src=/path/on/host,dst=/path/in/container,idmap
  • type=bind,src=/path/on/host,dst=/path/in/container,idmap=uids=0-1-10;gids=0-100-10
  • type=devpts,destination=/dev/pts
  • type=glob,src=/usr/lib/libfoo*,destination=/usr/lib,ro=true
  • type=image,source=fedora,destination=/fedora-image,rw=true
  • type=ramfs,tmpfs-size=512M,destination=/path/in/container
  • type=tmpfs,tmpfs-size=512M,destination=/path/in/container
  • type=tmpfs,destination=/path/in/container,noswap
  • type=artifact,src=quay.io/libpod/testartifact:20250206-single,dst=/data
  • type=artifact,src=quay.io/libpod/testartifact:20250206-multi,dst=/data,title=test1
  • type=volume,destination=/path/in/container
  • type=volume,src=test_vol,dst=/data,subpath=/code/docs

🎀 --name=name

Give your container a name, bestie. Like, an actual human one 💖

Fun fact: an operator can clock a container in three ways:

  • UUID long identifier (f78375b1c487e03c9438c729345e54db9d20cfa2ac1fc3494b6eb60872e74778) — the whole entire saga
  • UUID short identifier (f78375b1c487) — the lil teaser version
  • Name (jonah) — the vibe

Podman generates a UUID for each container, and if you don't hand it a name with --name, Podman makes up a random string name like exciting_chebyshev — that's adjective_noun, and yes it's compatible with Docker 🫶

Container names are not required to be valid hostnames or domain names. Underscores and other characters allowed by naming rules are permitted, so go off. On Podman networks with DNS enabled, container-to-container name resolution still uses the name as given — for example exciting_chebyshev. The name is lowkey useful as a more human-friendly way to identify containers, and it works for both background and foreground containers. Obsessed.

The container's name is also added to the /etc/hosts file using the container's primary IP address (also see the --add-host option).

Ngl, the name is not the hostname inside the container — that's a whole different thing, see --hostname. And see podman-network(1) for more on network DNS ✨

🌐 --network=mode, --net

Set the network mode for the container. Main character energy for your networking 🌈

Valid mode values are:

  • bridge[:OPTIONS,...]: Create a network stack on the default bridge. This is the default for rootful containers. It is possible to specify these additional options:
    • alias=name: Add network-scoped alias for the container.
    • ip=IPv4: Specify a static IPv4 address for this container.
    • ip6=IPv6: Specify a static IPv6 address for this container.
    • mac=MAC: Specify a static MAC address for this container.
    • interface_name=name: Specify a name for the created network interface inside the container.
    • host_interface_name=name: Specify a name for the created network interface outside the container.

Any other options will be passed through to netavark without validation. This can be useful to pass arguments to netavark plugins.

For rootless bridge networks, port forwarding uses rootlessport by default. Setting rootless_port_forwarder="pasta" in the [network] section of containers.conf(5) switches to pasta's kernel-level forwarding (via pesto), which preserves the original client source IP address inside the container. This option is experimental and its behavior is subject to change 🫣

For example, to set a static IPv4 address and a static mac address, use --network bridge:ip=10.88.0.10,mac=44:33:22:11:00:99.

  • <network name or ID>[:OPTIONS,...]: Connect to a user-defined network — this is the network name or ID from a network created by podman network create. It is possible to specify the same options described under the bridge mode above. Use the --network option multiple times to specify additional networks.

For backwards compatibility it is also possible to specify comma-separated networks on the first --network argument, however this prevents you from using the options described under the bridge section above. (iykyk, legacy stuff 💀)

  • none: Create a network namespace for the container but do not configure network interfaces for it, thus the container has no network connectivity.
  • container:id: Reuse another container's network stack. Sharing is caring 🫶
  • host: Use the host's network namespace for the container instead of creating an isolated namespace.
⚠️ Warning: This gives the container full access to abstract Unix domain sockets and to TCP/UDP sockets bound to localhost. Since these mechanisms are often used to prevent access to sensitive system services, isolating them from access by external entities, use of this option may be considered a security vulnerability. Like, fr fr be careful.
  • ns:path: Path to a network namespace to join.
  • private: Create a new namespace for the container. This uses the bridge mode for rootful containers and pasta for rootless ones.
  • pasta[:OPTIONS,...]: use pasta(1) to create a user-mode networking stack.

This is the default for rootless containers and only supported in rootless mode. By default, IPv4 and IPv6 addresses and routes, as well as the pod interface name, are copied from the host. Port forwarding preserves the original source IP address. Options described in pasta(1) can be specified as comma-separated arguments.

In terms of pasta(1) options, --config-net is given by default, in order to configure networking when the container is started, and --no-map-gw is also assumed by default, to avoid direct access from container to host using the gateway address. The latter can be overridden by passing --map-gw in the pasta-specific options (despite not being an actual pasta(1) option).

For better integration with DNS handling, --dns-forward 169.254.1.1 is passed by default, and this address is added to resolv.conf(5) as first resolver. It is possible to pass --dns-forward explicitly in case a different IP address should be used.

To make the host.containers.internal /etc/hosts entry work and allow connections to the host, --map-guest-addr 169.254.1.2 is passed by default. Again, it can be set explicitly to choose a different IP address.

Also, -t none and -u none are passed if, respectively, no TCP or UDP port forwarding from host to container is configured (via Podman's --publish or by passing the pasta -t/-u options directly), to disable automatic port forwarding based on bound ports. Similarly, -T none and -U none are given to disable the same functionality from container to host.

All options can also be set in containers.conf(5); see the pasta_options key under the network section in that file.

Some examples:

  • pasta:--map-gw: Allow the container to directly reach the host using the gateway address.
  • pasta:--mtu,1500: Specify a 1500 bytes MTU for the tap interface in the container.
  • pasta:--ipv4-only,-a,10.0.2.0,-n,24,-g,10.0.2.2,--dns-forward,10.0.2.3,-m,1500,--no-ndp,--no-dhcpv6,--no-dhcp: disable IPv6, assign 10.0.2.0/24 to the tap0 interface in the container, with gateway 10.0.2.3, enable DNS forwarder reachable at 10.0.2.3, set MTU to 1500 bytes, disable NDP, DHCPv6 and DHCP support.
  • pasta:-I,tap0,--ipv4-only,-a,10.0.2.0,-n,24,-g,10.0.2.2,--dns-forward,10.0.2.3,--no-ndp,--no-dhcpv6,--no-dhcp: same as above, but leave the MTU to 65520 bytes
  • pasta:-t,auto,-u,auto,-T,auto,-U,auto: enable automatic port forwarding based on observed bound ports from both host and container sides
  • pasta:-T,5201: enable forwarding of TCP port 5201 from container to host, using the loopback interface instead of the tap interface for improved performance
🚫 Invalid if using --dns, --dns-option, or --dns-search with --network set to none or container:id. Don't do it, bestie.

If used together with --pod, the container joins the pod's network namespace.

🏷️ --network-alias=alias

Add a network-scoped alias for the container, setting the alias for all networks that the container joins. To set a name only for a specific network, use the alias option as described under the --network option. If the network has DNS enabled (podman network inspect -f {{.DNSEnabled}} <name>), these aliases can be used for name resolution on the given network. This option can be specified multiple times — stack 'em up 💅

🩺 --no-healthcheck

Disable any defined healthchecks for container. No wellness checks today, the container is thriving (allegedly) 🙈

📛 --no-hostname

Do not create the /etc/hostname file in the containers.

By default, Podman manages the /etc/hostname file, adding the container's own hostname. When the --no-hostname option is set, the image's /etc/hostname will be preserved unmodified if it exists. Iconic, leave it be ✨

🗂️ --no-hosts

Do not modify the /etc/hosts file in the container.

Podman assumes control over the container's /etc/hosts file by default and adds entries for the container's name (see --name option) and hostname (see --hostname option), the internal host.containers.internal and host.docker.internal hosts, as well as any hostname added using the --add-host option. Refer to the --add-host option for details. Passing --no-hosts disables this, so that the image's /etc/hosts file is kept unmodified. The same can be achieved globally by setting no_hosts=true in containers.conf.

💀 This option conflicts with --add-host. Pick a lane.

☠️ --oom-kill-disable

Whether to disable OOM Killer for the container or not.

🫣 This flag is not supported on cgroups V2 systems. Just so you know.

📊 --oom-score-adj=num

Tune the host's OOM preferences for containers (accepts values from -1000 to 1000).

When running in rootless mode, the specified value can't be lower than the oom_score_adj for the current process. In this case, the oom-score-adj is clamped to the current process value. The system said no, respectfully 🌸

🖥️ --os=OS

Override the OS, defaults to the host, of the image to be pulled. For example, windows. Unless overridden, subsequent lookups of the same image in the local storage matches this OS, regardless of the host. It's giving commitment 💍

🔐 --passwd

Allow Podman to add entries to /etc/passwd and /etc/group when used in conjunction with the --user option. This is used to override the Podman provided user setup in favor of entrypoint configurations such as libnss-extrausers. Very specific, very slay.

✍️ --passwd-entry=ENTRY

Customize the entry that is written to the /etc/passwd file within the container when --passwd is used.

The variables $USERNAME, $UID, $GID, $NAME, $HOME are automatically replaced with their value at runtime. Mad libs but make it sysadmin 🎀

🎭 --personality=persona

Personality sets the execution domain via Linux personality(2). Yes it's literally called personality, no I'm not joking 👑

🪪 --pid=mode

Set the PID namespace mode for the container. The default is to create a private PID namespace for the container.

  • container:id: join another container's PID namespace;
  • host: use the host's PID namespace for the container. Note the host mode gives the container full access to local PID and is therefore considered insecure;
  • ns:path: join the specified PID namespace;
  • private: create a new namespace for the container (default).

📌 --pidfile=path

When the pidfile location is specified, the container process' PID is written to the pidfile. (This option is not available with the remote Podman client, including Mac and Windows (excluding WSL2) machines) If the pidfile option is not specified, the container process' PID is written to /run/containers/storage/${storage-driver}-containers/$CID/userdata/pidfile.

After the container is started, the location for the pidfile can be discovered with the following podman inspect command:

$ podman inspect --format '{{ .PidFile }}' $CID
/run/containers/storage/${storage-driver}-containers/$CID/userdata/pidfile

Like a little find-my-friend tracker but for PIDs 🦋

  • --pids-limit=limit — tunes the container's pids limit, bestie. Set it to -1 and the container gets unlimited pids (go off). Default is 2048 on systems that support the "pids" cgroup controller. 💅
  • --platform=OS/ARCH — tells Podman which platform to pick the image for. Heads up: it conflicts with --arch and --os. This one can override the current architecture and operating system. And unless you override it, subsequent lookups of that same image in local storage match this platform — regardless of the host. ✨
  • --pod=name — runs the container in an existing pod. Cute trick: Podman makes the pod automatically if the pod name is prefixed with new:. Want more granular options? Use the podman pod create command first, before creating a container. And when a container runs with a pod that has an infra-container, the infra-container starts first. 🫶
  • --pod-id-file=file — runs the container in an existing pod, but reads that pod's ID from the file you specify. Same deal: if the pod has an infra-container, the infra-container starts first. 💖
  • --preserve-fd=FD1[,FD2,...] — passes down to the process the extra file descriptors you list, comma separated. Can be specified multiple times. Only supported with the crun OCI runtime. Ngl, it might be a security risk to use this option with other OCI runtimes. 😭
Not available with the remote Podman client — including Mac and Windows (excluding WSL2) machines.
  • --preserve-fds=N — passes down to the process N additional file descriptors (on top of 0, 1, 2). Total FDs are 3+N. Math slay. 🧮
Also not available with the remote Podman client, including Mac and Windows (excluding WSL2) machines.
  • --privileged — gives extended privileges to this container. Default is false.

So by default, Podman containers are unprivileged (=false) and cannot, for example, modify parts of the operating system — because by default a container only gets limited access to devices. A "privileged" container gets the same access to devices as the user launching it, except virtual consoles (/dev/tty\d+) when running in systemd mode (--systemd=always).

A privileged container turns off the security features that isolate the container from the host. Dropped capabilities, limited devices, read-only mount points, Apparmor/SELinux separation, and Seccomp filters are all disabled. And because of all that, the privileged field should almost never be set — containers can easily break out of confinement. 🚨

Containers running in a user namespace (e.g., rootless containers) cannot have more privileges than the user that launched them.

  • --publish, -p=[[hostIP:][hostPort]:]containerPort[/protocol] — publishes a container's port, or a range of ports, to the host. 💅

Both hostPort and containerPort can be a range of ports. When you give ranges for both, the number of container ports in the range must match the number of host ports in the range.

If hostIP is 0.0.0.0, only IPv4 addresses get bound. If it's [::], only IPv6 addresses are bound. In all other cases the exact IP address as given is bound. If hostIP isn't set at all, the port is bound on all IP addresses on the host — both v4 and v6 (if possible).

The default hostIP can be configured in containers.conf(5) using the default_host_ips option under the [network] section. That option accepts an array so you can set multiple bind IP addresses. For example, setting default_host_ips = ["127.0.0.1", "::1"] makes the port publish on the ipv4 and ipv6 localhost address respectively. An explicit hostIP set on the command always overrides the config option. 👑

By default, Podman publishes TCP ports. To publish a UDP port instead, give udp as the protocol. To publish both TCP and UDP, set --publish twice, with tcp and udp as protocols respectively. Rootful containers can also publish ports using the sctp protocol.

Host port does not have to be specified (e.g. podman run -p 127.0.0.1::80). If it isn't, the container port is randomly assigned a port on the host.

Use podman port to see the actual mapping: podman port $CONTAINER $CONTAINERPORT.

Port publishing is only supported for containers using their own network namespace through bridge networks, or the pasta network mode.

For rootless bridge networks, port forwarding uses rootlessport by default — that's a userspace proxy that does not preserve client source IPs. Setting rootless_port_forwarder="pasta" in the [network] section of containers.conf(5) switches to pasta's kernel-level forwarding via pesto, which preserves the original client IP address inside the container. Heads up: this option is experimental and its behavior is subject to change. 🔥

Note: if a container runs within a pod, you don't need to publish the port for the containers in the pod. The port must only be published by the pod itself. Pod network stacks act like the network stack on the host — meaning a variety of containers in the pod and programs in the container all share a single interface, IP address, and associated ports. If one container binds to a port, no other container can use that port within the pod while it's in use. Containers in the pod can also communicate over localhost by having one container bind to localhost in the pod, and another connect to that port.
  • --publish-all, -P — publishes all exposed ports to random ports on the host interfaces. Default is false.

When set to true, all exposed ports get published to the host interfaces. If the operator uses -P (or -p) then Podman makes the exposed port accessible on the host, and the ports are available to any client that can reach the host. 🌈

When using this option, Podman binds any exposed port to a random port on the host within an ephemeral port range defined by /proc/sys/net/ipv4/ip_local_port_range. To find the mapping between the host ports and the exposed ports, use podman port.

  • --pull=policy — the pull image policy. Default is missing. The options, bestie:
    • always: always pull the image and throw an error if the pull fails.
    • missing: pull the image only when the image is not in the local containers storage. Throw an error if no image is found and the pull fails.
    • never: never pull the image, just use the one from the local containers storage. Throw an error if no image is found.
    • newer: pull if the image on the registry is newer than the one in the local containers storage. An image is considered newer when the digests are different — comparing the time stamps is prone to errors. Pull errors are suppressed if a local image was found. 🦋
  • --quiet, -q — suppresses output information when pulling images. 🤫
  • --rdt-class=intel-rdt-class-of-service — sets the class of service (CLOS or COS) for the container to run in. Based on the Cache Allocation Technology (CAT) feature that's part of Intel's Resource Director Technology (RDT) feature set, all container processes will run within the pre-configured COS, representing a part of the cache. The COS has to be created and configured using a pseudo file system (usually mounted at /sys/fs/resctrl) that the resctrl kernel driver provides. Assigning the container to a COS requires root privileges, so it does not work in a rootless environment. Currently the feature is only supported using runc as a runtime. See https://docs.kernel.org/arch/x86/resctrl.html for more details on creating a COS before a container can be assigned to it. 🌸
  • --read-only — mounts the container's root filesystem as read-only.

By default, container root filesystems are writable, letting processes write files anywhere. By specifying the --read-only flag, the container's root filesystem is mounted read-only, prohibiting any writes. 🎀

  • --read-only-tmpfs — when running --read-only containers, mounts a read-write tmpfs on /dev, /dev/shm, /run, /tmp, and /var/tmp. Default is true.

Here's the vibe check on how these two options play together:

  • --read-only=true + --read-only-tmpfs=true → / is r/o, and /run, /tmp, /var/tmp are r/w.
  • --read-only=true + --read-only-tmpfs=false → / is r/o, and /run, /tmp, /var/tmp are r/o.
  • --read-only=false + --read-only-tmpfs=false → / is r/w, and /run, /tmp, /var/tmp are r/w.
  • --read-only=false + --read-only-tmpfs=true → / is r/w, and /run, /tmp, /var/tmp are r/w.

When --read-only=true and --read-only-tmpfs=true, additional tmpfs are mounted on the /tmp, /run, and /var/tmp directories.

When --read-only=true and --read-only-tmpfs=false, /dev and /dev/shm are marked Read/Only and no tmpfs are mounted on /tmp, /run and /var/tmp. The directories are exposed from the underlying image, meaning they're read-only by default. This makes the container totally read-only — no writable directories exist within the container. In this mode, writable directories need to be added via external volumes or mounts. 😭

By default, when --read-only=false, /dev and /dev/shm are read/write, and /tmp, /run, and /var/tmp are read/write directories from the container image.

  • --replace — if another container with the same name already exists, replace and remove it. Default is false. ✨
  • --requires=container — specify one or more requirements. A requirement is a dependency container that's started before this one. Containers can be specified by name or ID, with multiple containers being separated by commas. 🧚
  • --restart=policy — the restart policy to follow when containers exit. Important: restart policy does not take effect if a container is stopped via the podman kill or podman stop commands. 💀

Valid policy values are:

  • no: do not restart containers on exit.
  • never: synonym for no; do not restart containers on exit.
  • on-failure[:max_retries] — restarts the container when it exits with a non-zero exit code, retrying indefinitely or until it hits that optional max_retries count 🌀
  • always — restarts the container when it exits, no matter what the status was, retrying indefinitely. forever. no thoughts, just restart 🔁
  • unless-stopped — restarts the container when it exits, unless the container was explicitly stopped by the user. After a system reboot, containers with this policy get restarted by podman-restart.service only if they were not explicitly stopped by the user before the reboot. This differs from always, which restarts containers after a system reboot regardless of whether they were user-stopped. iykyk 💅

Podman provides a systemd unit file, podman-restart.service, which restarts containers after a system reboot.

🚨 bestie listen — when you're running containers in systemd services, use the restart functionality provided by systemd. In other words, do not use this option in a container unit; instead set the Restart= systemd directive in the [Service] section. See podman-systemd.unit(5) and systemd.service(5). that's the whole assignment, periodt ✨
  • --retry=attempts — how many times to retry pulling or pushing images between the registry and local storage if it fails. Default is 3. three strikes but make it container 🎀
  • --retry-delay=duration — how long to wait between those retry attempts when pulling or pushing images between the registry and local storage fails. The default is to start at two seconds and then exponentially back off. But heads up: the delay is used when this value is set, and no exponential back off occurs. so like, you override it, you own it 🙈
  • --rm — automatically remove the container and any anonymous unnamed volume associated with the container when it exits. The default is false. cleanup queen behavior 🧹✨
  • --rmi — after the container exits, remove the image too, unless another container is using it. Implies --rm on the new container. The default is false. she takes everything with her 💀
  • --rootfs — if specified, the first argument refers to an exploded container on the file system.

This is useful to run a container without requiring any image management — the rootfs of the container is assumed to be managed externally. very "i brought my own files" energy 🦋

Overlay Rootfs Mounts

The :O flag tells Podman to mount the directory from the rootfs path as storage using the overlay file system. Container processes can modify content within the mount point, and that content is stored in the container storage in a separate directory. In overlay terms, the source directory is the lower, and the container storage directory is the upper. Modifications to the mount point are destroyed when the container finishes executing — similar to a tmpfs mount point being unmounted. poof, gone, that's the vibe 😭

⚠️ Note: on SELinux systems, the rootfs needs the correct label, which is by default unconfined_u:object_r:container_file_t:s0. don't skip this bestie, SELinux will absolutely clock you 🔥

idmap

If idmap is specified, create an idmapped mount to the target user namespace in the container. The idmap option supports a custom mapping that can be different than the user namespace used by the container. The mapping can be specified after the idmap option like: idmap=uids=0-1-10#10-11-10;gids=0-100-10. For each triplet, the first value is the start of the backing file system IDs that are mapped to the second value on the host. The length of this mapping is given in the third value. Multiple ranges are separated with #. math but make it filesystem 🧮💜

  • --sdnotify=container | conmon | healthy | ignore — determines how to use the NOTIFY_SOCKET, as passed with systemd and Type=notify.

Default is container, which means allow the OCI runtime to proxy the socket into the container to receive ready notification. Podman sets the MAINPID to conmon's pid. The conmon option sets MAINPID to conmon's pid, and sends READY when the container has started — the socket is never passed to the runtime or the container. The healthy option sets MAINPID to conmon's pid, and sends READY when the container has turned healthy; requires a healthcheck to be set. The socket is never passed to the runtime or the container. The ignore option removes NOTIFY_SOCKET from the environment for itself and child processes, for the case where some other process above Podman uses NOTIFY_SOCKET and Podman does not use it. four whole personalities, pick one 👑

  • --seccomp-policy=policy — specify the policy to select the seccomp profile. If set to image, Podman looks for a "io.containers.seccomp.profile" label in the container-image config and uses its value as a seccomp profile. Otherwise, Podman follows the default policy by applying the default profile unless specified otherwise via --security-opt seccomp as described below.
🌸 Note that this feature is experimental and may change in the future. she's still in beta, be nice to her
  • --secret=secret[,opt=opt ...] — give the container access to a secret. Can be specified multiple times. 🤫✨

A secret is a blob of sensitive data which a container needs at runtime but is not stored in the image or in source control — think usernames and passwords, TLS certificates and keys, SSH keys or other important generic strings or binary content (up to 512 kB in size).

When secrets are specified as type mount, the secrets are copied and mounted into the container when a container is created. When secrets are specified as type env, the secret is set as an environment variable within the container. Secrets are written in the container at the time of container creation, and modifying the secret using podman secret commands after the container is created affects the secret inside the container. so yes, it's live, it updates, iconic 💅

Secrets and its storage are managed using the podman secret command.

Secret Options

  • type=mount|env — how the secret is exposed to the container. mount mounts the secret into the container as a file. env exposes the secret as an environment variable. Defaults to mount.
  • target=target — target of secret. For mounted secrets, this is the path to the secret inside the container. If a fully qualified path is provided, the secret is mounted at that location. Otherwise, the secret is mounted to /run/secrets/target for Linux containers or /var/run/secrets/target for FreeBSD containers. If the target is not set, the secret is mounted to /run/secrets/secretname by default. For env secrets, this is the environment variable key. Defaults to secretname.
  • uid=0 — UID of secret. Defaults to 0. Mount secret type only.
  • gid=0 — GID of secret. Defaults to 0. Mount secret type only.
  • mode=0 — mode of secret. Defaults to 0444. Mount secret type only.

Examples

Mount at /my/location/mysecret with UID 1:

--secret mysecret,target=/my/location/mysecret,uid=1

Mount at /run/secrets/customtarget with mode 0777:

--secret mysecret,target=customtarget,mode=0777

Create a secret environment variable called ENVSEC:

--secret mysecret,type=env,target=ENVSEC
  • --security-opt=option — security options. ok this one's a whole mood board, buckle up 🎀

Security Options

  • apparmor=unconfined — turn off apparmor confinement for the container
  • apparmor=alternate-profile — set the apparmor confinement profile for the container
  • label=user:USER — set the label user for the container processes
  • label=role:ROLE — set the label role for the container processes
  • label=type:TYPE — set the label process type for the container processes
  • label=level:LEVEL — set the label level for the container processes
  • label=filetype:TYPE — set the label file type for the container files
  • label=disable — turn off label separation for the container
💜 Note: labeling can be disabled for all containers by setting label=false in the containers.conf (/etc/containers/containers.conf or $HOME/.config/containers/containers.conf) file.
  • label=nested — allows SELinux modifications within the container. Containers are allowed to modify SELinux labels on files and processes, as long as SELinux policy allows. Without nested, containers view SELinux as disabled, even when it is enabled on the host. Containers are prevented from setting any labels.
  • mask=/path/1:/path/2 — the paths to mask, separated by a colon. A masked path cannot be accessed inside the container. 🙈
  • no-new-privileges — disable container processes from gaining additional privileges through the execve(2) system call (e.g. via setuid or setgid bits, or via file capabilities). Programs that rely on setuid/setgid bits set on their executable to change user id or group id are no longer able to do so, and any file capabilities added to the executable (e.g. via setcap) are not added to the permitted capability set. For more details, see: https://docs.kernel.org/userspace-api/no_new_privs.html. no glow-ups allowed 🔥
  • seccomp=unconfined — turn off seccomp confinement for the container.
  • seccomp=profile.json — JSON file to be used as a seccomp filter. Note that the io.podman.annotations.seccomp annotation is set with the specified value as shown in podman inspect.
  • proc-opts=OPTIONS — comma-separated list of options to use for the /proc mount. More details for the possible mount options are specified in the proc(5) man page.
  • unmask=ALL or /path/1:/path/2, or shell expanded paths (/proc/*) — paths to unmask, separated by a colon. If set to ALL, it unmasks all the paths that are masked or made read-only by default. The default masked paths are /proc/acpi, /proc/kcore, /proc/keys, /proc/latency_stats, /proc/sched_debug, /proc/scsi, /proc/timer_list, /proc/timer_stats, /sys/firmware, and /sys/fs/selinux, /sys/devices/virtual/powercap. The default paths that are read-only are /proc/asound, /proc/bus, /proc/fs, /proc/irq, /proc/sys, /proc/sysrq-trigger, /sys/fs/cgroup. that's a lot of paths, she did the reading so you don't have to 📚✨
💜 Note: labeling can be disabled for all containers by setting label=false in the containers.conf(5) file.
  • --shm-size=number[unit] — size of /dev/shm. A unit can be b (bytes), k (kibibytes), m (mebibytes), or g (gibibytes). If the unit is omitted, the system uses bytes. If the size is omitted, the default is 64m. When size is 0, there is no limit on the amount of memory used for IPC by the container. This option conflicts with --ipc=host. ⚠️ pick a lane bestie
  • --shm-size-systemd=number[unit] — size of systemd-specific tmpfs mounts such as /run, /run/lock, /var/log/journal and /tmp. A unit can be b (bytes), k (kibibytes), m (mebibytes), or g (gibibytes). If the unit is omitted, the system uses bytes. If the size is omitted, the default is 64m. When size is 0, the usage is limited to 50% of the host's available memory. halfsies, that's the rule 🍓
  • --sig-proxy — proxy received signals to the container process. SIGCHLD, SIGURG, SIGSTOP, and SIGKILL are not proxied. 🚫

The default is true. she's a messenger and she's not touching those four, no cap 💅

  • --stop-signal=signal — signal to stop a container. Default is SIGTERM. ✨
  • --stop-timeout=seconds — Timeout to stop a container, bestie. Default is 10. Remote connections use local containers.conf for defaults.
  • --subgidname=name — Run the container in a new user namespace using the map with name in the /etc/subgid file. When running rootless, the user needs the necessary permissions to use the mapping. See subgid(5). This flag conflicts with --userns and --gidmap. Together with --gidmap it acts as if --gidmap was passed with all mappings of the name group in /etc/subgid.
  • --subuidname=name — Run the container in a new user namespace using the map with name in the /etc/subuid file. When running rootless, the user needs the necessary permissions to use the mapping. See subuid(5). This flag conflicts with --userns and --uidmap. Together with --uidmap it acts as if --uidmap was passed with all mappings of the name user in /etc/subuid.
  • --sysctl=name=value — Configure namespaced kernel parameters at runtime. Kernel vibes, but make it container-scoped. ✨

For the IPC namespace, the following sysctls are allowed:

  • kernel.msgmax
  • kernel.msgmnb
  • kernel.msgmni
  • kernel.sem
  • kernel.shmall
  • kernel.shmmax
  • kernel.shmnni
  • kernel.shm_rmid_forced
  • Sysctls beginning with fs.mqueue.*
Note: if using the --ipc=host option, the above sysctls are not allowed.

For the network namespace, only sysctls beginning with net.* are allowed.

Note: if using the --network=host option, the above sysctls are not allowed.
  • --systemd=true | false | always — Run container in systemd mode. The default is true. Systemd girlies, this one's for you. 🫶
    • true enables systemd mode only when the command executed inside the container is systemd, /usr/sbin/init, /sbin/init or /usr/local/sbin/init.
    • false disables systemd mode.
    • always enforces the systemd mode to be enabled.

Running the container in systemd mode causes the following changes:

  • Podman mounts tmpfs file systems on the following directories:
    • /run
    • /run/lock
    • /tmp
    • /sys/fs/cgroup/systemd (on a cgroup v1 system)
    • /var/lib/journal
  • Podman sets the default stop signal to SIGRTMIN+3.
  • Podman sets container_uuid environment variable in the container to the first 32 characters of the container ID.
  • Podman does not mount virtual consoles (/dev/tty\d+) when running with --privileged.
  • On cgroup v2, /sys/fs/cgroup is mounted writable.

This allows systemd to run in a confined container without any modifications.

Note: on SELinux systems, systemd attempts to write to the cgroup file system. Containers writing to the cgroup file system are denied by default. The container_manage_cgroup boolean must be enabled for this to be allowed on an SELinux separated system.
setsebool -P container_manage_cgroup true
  • --timeout=seconds — Maximum time a container is allowed to run before conmon sends it the kill signal. By default containers run until they exit or are stopped by podman stop. So like, unless you say otherwise, it's gonna live its best life until it's done. ⏳
  • --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). Certificate check era 💅
  • --tmpfs=fs — Create a tmpfs mount. Temporary filesystem, permanent slay. 🦋

Mount a temporary filesystem (tmpfs) mount into a container, for example:

$ podman run -d --tmpfs /tmp:rw,size=787448k,mode=1777 my_image

This command mounts a tmpfs at /tmp within the container. The supported mount options are the same as the Linux default mount flags. If no options are specified, the system uses the following options: rw,noexec,nosuid,nodev.

By default, Podman enables tmpcopyup on tmpfs mounts, which copies the contents of the underlying image directory into the tmpfs before mounting it. This also applies when the tmpfs destination is inside a volume or bind mount: files from the parent mount are copied into the tmpfs, so the parent content remains visible. To mount an empty tmpfs that shadows a parent mount's subtree, use the notmpcopyup option:

$ podman run --volume myvolume:/data --tmpfs /data/sub:notmpcopyup my_image
  • --tty, -t — Allocate a pseudo-TTY. The default is false. Interactive terminal moment ✨

When set to true, Podman allocates a pseudo-tty and attach to the standard input of the container. This can be used, for example, to run a throwaway interactive shell.

NOTE: The --tty flag prevents redirection of standard output. It combines STDOUT and STDERR, it can insert control characters, and it can hang pipes. This option is only used when run interactively in a terminal. When feeding input to Podman, use -i only, not -it. Do not come for me with -it when you're piping, bestie, it's giving chaos. 💀
echo "asdf" | podman run --rm -i someimage /bin/cat
  • --tz=timezone — Set timezone in container. This flag takes area-based timezones, GMT time, as well as local, which sets the timezone in the container to match the host machine. See /usr/share/zoneinfo/ for valid timezones. Remote connections use local containers.conf for defaults. Time zone check: passed 🕰️
  • --uidmap=[flags]container_uid:from_uid[:amount] — Run the container in a new user namespace using the supplied UID mapping. This option conflicts with the --userns and --subuidname options. This option provides a way to map host UIDs to container UIDs. It can be passed several times to map different ranges. UID mapping lore drop incoming 👑

The possible values of the optional flags are discussed further down on this page. The amount value is optional and assumed to be 1 if not given.

The from_uid value is based upon the user running the command, either rootful or rootless users.

  • rootful user: [flags]container_uid:host_uid[:amount]
  • rootless user: [flags]container_uid:intermediate_uid[:amount]

Rootful mappings

When podman run is called by a privileged user, the option --uidmap works as a direct mapping between host UIDs and container UIDs.

host UID -> container UID

The amount specifies the number of consecutive UIDs that is mapped. If for example amount is 4 the mapping looks like:

  • from_uid → container_uid
  • from_uid + 1 → container_uid + 1
  • from_uid + 2 → container_uid + 2
  • from_uid + 3 → container_uid + 3

Rootless mappings

When podman run is called by an unprivileged user (i.e. running rootless), the value from_uid is interpreted as an "intermediate UID". In the rootless case, host UIDs are not mapped directly to container UIDs. Instead the mapping happens over two mapping steps:

host UID -> intermediate UID -> container UID

The --uidmap option only influences the second mapping step.

The first mapping step is derived by Podman from the contents of the file /etc/subuid and the UID of the user calling Podman.

First mapping step: host UID -> intermediate UID

  • UID for Podman user → 0
  • 1st subordinate UID → 1
  • 2nd subordinate UID → 2
  • 3rd subordinate UID → 3
  • nth subordinate UID → n

To be able to use intermediate UIDs greater than zero, the user needs to have subordinate UIDs configured in /etc/subuid. See subuid(5).

The second mapping step is configured with --uidmap.

If for example amount is 5 the second mapping step looks like:

  • from_uid → container_uid
  • from_uid + 1 → container_uid + 1
  • from_uid + 2 → container_uid + 2
  • from_uid + 3 → container_uid + 3
  • from_uid + 4 → container_uid + 4

When running as rootless, Podman uses all the ranges configured in the /etc/subuid file.

The current user ID is mapped to UID=0 in the rootless user namespace.

Every additional range is added sequentially afterward:

okay bestie, here’s the mapping table for that, and yes it’s giving spreadsheet but make it cute: 💖

  • host $UID → rootless user namespace 0, length 1
  • host 1 → rootless user namespace $FIRST_RANGE_ID, length $FIRST_RANGE_LENGTH
  • host 1+$FIRST_RANGE_LENGTH → rootless user namespace $SECOND_RANGE_ID, length $SECOND_RANGE_LENGTH

🦋 Referencing a host ID from the parent namespace 🦋

As a rootless user, the given host ID in --uidmap or --gidmap is mapped from the intermediate namespace generated by Podman. Sometimes it is desirable to refer directly to the host namespace. It is possible to manually do so, by running podman unshare cat /proc/self/gid_map, finding the desired host id at the second column of the output, and getting the corresponding intermediate id from the first column. iykyk 👀

Podman can perform all that by preceding the host id in the mapping with the @ symbol. For instance, by specifying --gidmap 100000:@2000:1, podman will look up the intermediate id corresponding to host id 2000 and it will map the found intermediate id to the container id 100000. The given host id must have been subordinated (otherwise it would not be mapped into the intermediate space in the first place). periodt. 💅

If the length is greater than one, for instance with --gidmap 100000:@2000:2, Podman will map host ids 2000 and 2001 to 100000 and 100001, respectively, regardless of how the intermediate mapping is defined. slay ✨

🌸 Extending previous mappings 🌸

Some mapping modifications may be cumbersome. For instance, a user starts with a mapping such as --gidmap="0:0:65000", that needs to be changed such as the parent id 1 is mapped to container id 100000 instead, leaving container id 1 unassigned. The corresponding --gidmap becomes --gidmap="0:0:1" --gidmap="2:2:65534" --gidmap="100000:1:1". lowkey a lot, ngl 😭

This notation can be simplified using the + flag, that takes care of breaking previous mappings removing any conflicting assignment with the given mapping. The flag is given before the container id as follows: --gidmap="0:0:65000" --gidmap="+100000:1:1". iconic behavior 🎀

  • flag + — example +100000:1:1 — description: Extend the previous mapping

This notation leads to gaps in the assignment, so it may be convenient to fill those gaps afterwards: --gidmap="0:0:65000" --gidmap="+100000:1:1" --gidmap="1:65001:1". no cap, that’s the move 🫶

One specific use case for this flag is in the context of rootless users. A rootless user may specify mappings with the + flag as in --gidmap="+100000:1:1". Podman will then “fill the gaps” starting from zero with all the remaining intermediate ids. This is convenient when a user wants to map a specific intermediate id to a container id, leaving the rest of subordinate ids to be mapped by Podman at will. main character mapping energy fr fr 👑

💜 Passing only one of --uidmap or --gidmap 💜

Usually, subordinated user and group ids are assigned simultaneously, and for any user the subordinated user ids match the subordinated group ids. For convenience, if only one of --uidmap or --gidmap is given, podman assumes the mapping refers to both UIDs and GIDs and applies the given mapping to both. If only one value of the two needs to be changed, the mappings should include the u or the g flags to specify that they only apply to UIDs or GIDs and should not be copied over. the way that saves you from typing everything twice? obsessed 💖

  • flag u — example u20000:2000:1 — description: The mapping only applies to UIDs
  • flag g — example g10000:1000:1 — description: The mapping only applies to GIDs

For instance given the command

podman run --gidmap "0:0:1000" --gidmap "g2000:2000:1"

Since no --uidmap is given, the --gidmap is copied to --uidmap, giving a command equivalent to

podman run --gidmap "0:0:1000" --gidmap "2000:2000:1" --uidmap "0:0:1000"

The --gidmap "g2000:2000:1" used the g flag and therefore it was not copied to --uidmap. see? little flag, big impact ✨

🧚 Rootless mapping of additional host GIDs 🧚

A rootless user may desire to map a specific host group that has already been subordinated within /etc/subgid without specifying the rest of the mapping.

This can be done with --gidmap "+gcontainer_gid:@host_gid"

Where:

  • The host GID is given through the @ symbol
  • The mapping of this GID is not copied over to --usermap thanks to the g flag.
  • The rest of the container IDs will be mapped starting from 0 to n, with all the remaining subordinated GIDs, thanks to the + flag.

For instance, if a user belongs to the group 2000 and that group is subordinated to that user (with usermod --add-subgids 2000-2000 $USER), the user can map the group into the container with: --gidmap=+g100000:@2000. 🍓

If this mapping is combined with the option, --group-add=keep-groups, the process in the container will belong to group 100000, and files belonging to group 2000 in the host will appear as being owned by group 100000 inside the container. the way this just works? iconic 🔥

podman run --group-add=keep-groups --gidmap="+g100000:@2000" ...

🎀 No subordinate UIDs 🎀

Even if a user does not have any subordinate UIDs in /etc/subuid, --uidmap can be used to map the normal UID of the user to a container UID by running podman run --uidmap $container_uid:0:1 --user $container_uid .... bestie, it’s giving “make it work with what you have” 💅

👑 Pods 👑

The --uidmap option cannot be called in conjunction with the --pod option as a uidmap cannot be set on the container level when in a pod. no cap, that’s just the rule 🫶

  • --ulimit=option — Ulimit options. Sets the ulimits values inside of the container.

--ulimit with a soft and hard limit in the format =[:]. For example:

$ podman run --ulimit nofile=1024:1024 --rm ubi9 ulimit -n 1024

Set -1 for the soft or hard limit to set the limit to the maximum limit of the current process. In rootful mode this is often unlimited. 🌈

If nofile and nproc are unset, a default value of 1048576 will be used, unless overridden in containers.conf(5). However, if the default value exceeds the hard limit for the current rootless user, the current hard limit will be applied instead. highkey important 💜

Use host to copy the current configuration from the host.

Don’t use nproc with the ulimit flag as Linux uses nproc to set the maximum number of processes available to a user, not to a container. literally not the same thing, bestie 😭

Use the --pids-limit option to modify the cgroup control to limit the number of processes within a container.

  • --umask=umask — Set the umask inside the container. Defaults to 0022. Remote connections use the local containers.conf for defaults.
  • --unsetenv=env — Unset default environment variables for the container. Default environment variables include variables provided natively by Podman, environment variables configured by the image, and environment variables from containers.conf. bye bye defaults 👋
  • --unsetenv-all — Unset all default environment variables for the container. Default environment variables include variables provided natively by Podman, environment variables configured by the image, and environment variables from containers.conf. not me clearing the whole environment 💀
  • --user, -u=user[:group] — Sets the username or UID used and, optionally, the groupname or GID for the specified command. Both user and group may be symbolic or numeric.

Without this argument, the command runs as the user specified in the container image. Unless overridden by a USER command in the Containerfile or by a value passed to this option, this user generally defaults to root. ✨

When a user namespace is not in use, the UID and GID used within the container and on the host match. When user namespaces are in use, however, the UID and GID in the container may correspond to another UID and GID on the host. In rootless containers, for example, a user namespace is always used, and root in the container by default corresponds to the UID and GID of the user invoking Podman. the way identity gets a little remix in rootless mode 🦋

  • --userns=mode — Set the user namespace mode for the container.

If --userns is not set, the default value is determined as follows.

  • If --pod is set, --userns is ignored and the user namespace of the pod is used.
  • If the environment variable PODMAN_USERNS is set its value is used.
  • If userns is specified in containers.conf this value is used.
  • Otherwise, --userns=host is assumed.

--userns="" (i.e., an empty string) is an alias for --userns=host. 🎀

This option is incompatible with --gidmap, --uidmap, --subuidname and --subgidname. no mixing, bestie 💅

Rootless user --userns=Key mappings:

  • auto — Host User is $UID, and the container user is nil (the host user's UID just isn't mapped into the container, periodt).
  • host — Host User is $UID, container user is 0 (that's the default user account, mapped to root inside the container, iykyk).
  • keep-id — Host User is $UID, container user is $UID (the user account gets mapped to the same UID inside the container 💅).
  • keep-id:uid=200,gid=210 — Host User is $UID, container user is 200:210 (maps the user account to the UID and GID values you specify inside the container ✨).
  • nomap — Host User is $UID, container user is nil (again: host user's UID is not mapped into the container).

Valid mode values are, bestie:

auto[:OPTIONS,...] — automatically creates a unique user namespace 🦋

  • rootful mode: the --userns=auto flag requires that the user name containers be specified in the /etc/subuid and /etc/subgid files, with an unused range of subordinate user IDs that Podman containers are allowed to allocate. Example: containers:2147483647:2147483648.
  • rootless mode: the users range from the /etc/subuid and /etc/subgid files gets used. Note that running a single container without --userns=auto uses the entire range of UIDs and doesn't allow further subdividing. See subuid(5).

Podman allocates unique ranges of UIDs and GIDs from the containers subordinate user IDs. The size of those ranges is based on the number of UIDs required in the image — and you can override the number of UIDs and GIDs with the size option.

The option --userns=keep-id uses all the subuids and subgids of the user. The option --userns=nomap uses all the subuids and subgids of the user except the user's own ID. And heads up: using --userns=auto when starting new containers does not work as long as any containers exist that were started with --userns=nomap or --userns=keep-id without limiting the user namespace size 😭

Valid auto options:

  • gidmapping=CONTAINER_GID:HOST_GID:SIZE — forces a GID mapping to be present in the user namespace.
  • size=SIZE — specifies an explicit size for the automatic user namespace, e.g. --userns=auto:size=8192. If size isn't specified, auto estimates a size for the user namespace.
  • uidmapping=CONTAINER_UID:HOST_UID:SIZE — forces a UID mapping to be present in the user namespace.

The host UID and GID in gidmapping and uidmapping can optionally be prefixed with the @ symbol. In that case, podman looks up the intermediate ID corresponding to the host ID and maps the found intermediate ID to the container id. For details see --uidmap.

container:id — joins the user namespace of the specified container 🫶

host or "" (empty string) — runs in the user namespace of the caller. The processes running in the container have the same privileges on the host as any other process launched by the calling user.

keep-id — creates a user namespace where the current user's UID:GID are mapped to the same values in the container. For containers created by root, the current mapping is created into a new user namespace.

In addition, the init process within the container runs under the current user's UID. This behavior overrides the image's USER instruction — unless you explicitly set --user.

Valid keep-id options:

  • uid=UID — overrides the UID inside the container that's used to map the current user to.
  • gid=GID — overrides the GID inside the container that's used to map the current user to.
  • size=SIZE — overrides the size of the configured user namespace. It's useful so you don't saturate all the available IDs. Not supported when running as root.

nomap — creates a user namespace where the current rootless user's UID:GID are not mapped into the container. This option is not allowed for containers created by the root user 🚫

ns:namespace — runs the container in the given existing user namespace.

🌸 --uts=mode 🌸

Sets the UTS namespace mode for the container. These are the supported values:

  • host — use the host's UTS namespace inside the container.
  • private — create a new namespace for the container (this is the default).
  • ns:[path] — run the container in the given existing UTS namespace.
  • container:[container] — join the UTS namespace of the specified container.

🎀 --variant=VARIANT 🎀

Uses VARIANT instead of the default architecture variant of the container image. Some images can use multiple variants of the arm architectures, like arm/v5 and arm/v7. Slay, flexible architecture ✨

💖 --volume, -v=[[SOURCE-VOLUME|HOST-DIR:]CONTAINER-DIR[:OPTIONS]] 💖

Creates a bind mount. If -v /HOST-DIR:/CONTAINER-DIR is specified, Podman bind mounts /HOST-DIR from the host into /CONTAINER-DIR in the Podman container. Similarly, -v SOURCE-VOLUME:/CONTAINER-DIR mounts the named volume from the host into the container. If no such named volume exists, Podman creates one 🪄 — but the nocreate option can be used to disable that behavior and require the volume to already exist. If no source is given, the volume is created as an anonymously named volume with a randomly generated name, and it gets removed when the container is removed via the --rm flag or the podman rm --volumes command.

(Note: when using the remote client — including Mac and Windows machines, excluding WSL2 — the volumes are mounted from the remote server, not necessarily the client machine. Important!! 🙈)

The OPTIONS is a comma-separated list and can be one or more of:

  • rw|ro
  • z|Z
  • [O]
  • [U]
  • [no]copy
  • [no]dev
  • [no]exec
  • [no]suid
  • [r]bind
  • [r]shared|[r]slave|[r]private[r]unbindable [1]
  • idmap[=options]
  • nocreate

The CONTAINER-DIR must be an absolute path such as /src/docs. The volume is mounted into the container at this directory.

If a volume source is specified, it must be a path on the host or the name of a named volume. Host paths are allowed to be absolute or relative — relative paths are resolved relative to the directory Podman is run in. If the source doesn't exist, Podman returns an error. So yeah, you gotta pre-create the source files or directories yourself 💀

Any source that doesn't begin with a . or / is treated as the name of a named volume. If a volume with that name doesn't exist, it gets created. Volumes created with names are not anonymous, and they are not removed by the --rm option or the podman rm --volumes command 👑

The nocreate option can be specified for named volumes to prevent automatic volume creation. If nocreate is set and the volume doesn't exist, Podman returns an error instead of creating the volume. This is useful when you want to make sure a volume was explicitly created before use.

$ podman run -v myvolume:/data:nocreate alpine

Specify multiple -v options to mount one or more volumes into a container.

🔒 Write Protected Volume Mounts

Add the :ro or :rw option to mount a volume in read-only or read-write mode, respectively. By default, volumes are mounted read-write. See examples.

🫶 Chowning Volume Mounts

When a named volume is first mounted to a container, Podman automatically adjusts the ownership of the volume's mount point during container initialization. This chown operation occurs under the following conditions:

  • The volume was not used yet (has NeedsChown set to true)
  • The volume is empty or has not been copied up yet
  • The volume is not managed by an external volume driver
  • The volume driver is not "image"

For volumes with idmapped mounts (using the idmap option), the ownership change takes into account the container's user namespace mappings — but the idmapped volume retains proper UID/GID mapping. For volumes without idmapping, the mount point is chowned to match the container's process user and group, mapped to the host user namespace if user namespace remapping is enabled.

If a container is created in a new user namespace, the UID and GID in the container may correspond to another UID and GID on the host.

The :U suffix tells Podman to use the correct host UID and GID based on the UID and GID within the container, to recursively change the owner and group of the source volume. Chowning walks the file system under the volume and changes the UID/GID on each file. If the volume has thousands of inodes, this process takes a long time, delaying the start of the container 😭

Warning: use with caution, since this modifies the host filesystem. Like, actually be careful bestie.

🏷️ Labeling Volume Mounts

Labeling systems like SELinux require that proper labels are placed on volume content mounted into a container. Without a label, the security system might prevent the processes running inside the container from using the content. By default, Podman does not change the labels set by the OS.

So to change a label in the container context, you just add one of two suffixes to the volume mount: :z or :Z. These suffixes are basically telling Podman "bestie, relabel the file objects on the shared volumes." 💅

Here's the tea on which one to use:

  • :z — tells Podman that two or more containers share the volume content. Podman labels the content with a shared content label. Shared volume labels allow all containers to read/write content. 🫶
  • :Z — tells Podman to label the content with a private, unshared label. Only the current container can use a private volume. 🔒
Note: all containers within a pod share the same SELinux label. That means all containers within said pod can read/write volumes shared into the container created with the :Z on any one of the containers. Relabeling walks the file system under the volume and changes the label on each file — so if the volume has thousands of inodes, this process takes a long time, delaying the start of the container. If the volume was previously relabeled with the z option, Podman is optimized to not relabel a second time. If files are moved into the volume, then the labels can be manually changed with the chcon -Rt container_file_t PATH command.
Note: do not relabel system files and directories, ok? Relabeling system content might cause other confined services on the machine to fail, and nobody wants that kind of chaos. 💀 For these types of containers we recommend disabling SELinux separation. The option --security-opt label=disable disables SELinux separation for the container. Like, if a user wanted to volume mount their entire home directory into a container, they need to disable SELinux separation. That's just how it is.
$ podman run --security-opt label=disable -v $HOME:/home/user fedora touch /home/user/file

🦋 Overlay Volume Mounts

The :O flag tells Podman to mount the directory from the host as a temporary storage using the overlay file system. The container processes can modify content within the mountpoint, and that modified content is stored in the container storage in a separate directory. In overlay terms: the source directory is the lower, and the container storage directory is the upper. Modifications to the mount point are destroyed when the container finishes executing — similar to a tmpfs mount point being unmounted. ✨

For advanced users, the overlay option also supports custom non-volatile upperdir and workdir for the overlay mount. Custom upperdir and workdir can be fully managed by the users themselves, and Podman does not remove it on lifecycle completion. Example:

:O,upperdir=/some/upper,workdir=/some/work

Subsequent executions of the container sees the original source directory content — any changes from previous container executions no longer exist. Poof, gone. 🙈

One use case of the overlay mount is sharing the package cache from the host into the container to allow speeding up builds.

Note: the O flag conflicts with other options listed above. Not a vibe to mix them.

Content mounted into the container is labeled with the private label. On SELinux systems, labels in the source directory must be readable by the container label. Usually containers can read/execute container_share_t and can read/write container_file_t. If unable to change the labels on a source volume, SELinux container separation must be disabled for the container to work.

Do not modify the source directory mounted into the container with an overlay mount — it can cause unexpected failures. Only modify the directory after the container finishes running. I'm serious, bestie. 😭

🌀 Mounts propagation

By default, bind-mounted volumes are private. That means any mounts done inside the container are not visible on the host and vice versa. One can change this behavior by specifying a volume mount propagation property. When a volume is shared, mounts done under that volume inside the container are visible on host and vice versa. Making a volume slave[1] ⟨#Footnote1⟩ enables only one-way mount propagation: mounts done on the host under that volume are visible inside the container but not the other way around.

To control mount propagation property of a volume one can use the [r]shared, [r]slave, [r]private or the [r]unbindable propagation flag. Propagation property can be specified only for bind mounted volumes and not for internal volumes or named volumes. For mount propagation to work the source mount point (the mount point where source dir is mounted on) has to have the right propagation properties. For shared volumes, the source mount point has to be shared. And for slave volumes, the source mount point has to be either shared or slave. [1] ⟨#Footnote1⟩

To recursively mount a volume and all of its submounts into a container, use the rbind option. By default the bind option is used, and submounts of the source directory is not mounted into the container.

Mounting the volume with a copy option tells podman to copy content from the underlying destination directory onto newly created internal volumes. The copy only happens on the initial creation of the volume. Content is not copied up when the volume is subsequently used on different containers. The copy option is ignored on bind mounts and has no effect.

Mounting volumes with the nosuid options means that SUID executables on the volume can not be used by applications to change their privilege. By default volumes are mounted with nosuid.

Mounting the volume with the noexec option means that no executables on the volume can be executed within the container.

Mounting the volume with the nodev option means that no devices on the volume can be used by processes within the container. By default volumes are mounted with nodev.

If the HOST-DIR is a mount point, then dev, suid, and exec options are ignored by the kernel. Just how the kernel rolls. 💀

Use df HOST-DIR to figure out the source mount, then use findmnt -o TARGET,PROPAGATION source-mount-dir to figure out propagation properties of source mount. If findmnt(1) utility is not available, then one can look at the mount entry for the source mount point in /proc/self/mountinfo. Look at the "optional fields" and see if any propagation properties are specified. In there, shared:N means the mount is shared, master:N means mount is slave, and if nothing is there, the mount is private. [1] ⟨#Footnote1⟩

To change propagation properties of a mount point, use mount(8) command. For example, if one wants to bind mount source directory /foo, one can do mount --bind /foo /foo and mount --make-private --make-shared /foo. This converts /foo into a shared mount point. Alternatively, one can directly change propagation properties of source mount. Say / is source mount for /foo, then use mount --make-shared / to convert / into a shared mount.

Note: if the user only has access rights via a group, accessing the volume from inside a rootless container fails. Yeah. Annoying but true. 🥲

🌈 Idmapped mount

If idmap is specified, create an idmapped mount to the target user namespace in the container. The idmap option supports a custom mapping that can be different than the user namespace used by the container. The mapping can be specified after the idmap option like: idmap=uids=0-1-10#10-11-10;gids=0-100-10. For each triplet, the first value is the start of the backing file system IDs that are mapped to the second value on the host. The length of this mapping is given in the third value. Multiple ranges are separated with #.

Use the --group-add keep-groups option to pass the user's supplementary group access into the container. 🫶

👑 --volumes-from=CONTAINER[:OPTIONS]

Mount volumes from the specified container(s). Used to share volumes between containers. The options is a comma-separated list with the following available elements:

  • rw|ro
  • z

Mounts already mounted volumes from a source container onto another container. CONTAINER may be a name or ID. To share a volume, use the --volumes-from option when running the target container. Volumes can be shared even if the source container is not running. Slay. ✨

By default, Podman mounts the volumes in the same mode (read-write or read-only) as it is mounted in the source container. This can be changed by adding a ro or rw option.

Labeling systems like SELinux require that proper labels are placed on volume content mounted into a container. Without a label, the security system might prevent the processes running inside the container from using the content. By default, Podman does not change the labels set by the OS.

To change a label in the container context, add z to the volume mount. This suffix tells Podman to relabel file objects on the shared volumes. The z option tells Podman that two entities share the volume content. As a result, Podman labels the content with a shared content label. Shared volume labels allow all containers to read/write content. 💖

If the location of the volume from the source container overlaps with data residing on a target container, then the volume hides that data on the target. Sneaky but that's the rules.

🎀 --workdir, -w=dir

Working directory inside the container.

The default working directory for running binaries within a container is the root directory (/). The image developer can set a different default with the WORKDIR instruction. The operator can override the working directory by using the -w option. Main character energy: you get the final say. 💅

⭐ Exit Status ⭐

The exit code from podman run gives information about why the container failed to run or why it exited. When podman run exits with a non-zero code, the exit codes follow the chroot(1) standard, see below:

125 — The error is with Podman itself

$ podman run --foo busybox; echo $?
Error: unknown flag: --foo
125

126 — The contained command cannot be invoked

$ podman run busybox /etc; echo $?
Error: container_linux.go:346: starting container process caused "exec: \"/etc\": permission denied": OCI runtime error
126

127 — The contained command cannot be found

$ podman run busybox foo; echo $?
Error: container_linux.go:346: starting container process caused "exec: \"foo\": executable file not found in $PATH": OCI runtime error
127

Exit code contained command exit code

$ podman run busybox /bin/sh -c 'exit 3'; echo $?
3

🌸 EXAMPLES 🌸

🦋 Running container in read-only mode

During container image development, containers often need to write to the image content. Installing packages into /usr, for example. In production, applications seldom need to write to the image. Container applications write to volumes if they need to write to file systems at all. Applications can be made more secure by running them in read-only mode using the --read-only switch. This protects the container's image from modification. By default read-only containers can write to temporary data. Podman mounts a tmpfs on /run and /tmp within the container. 🎀

$ podman run --read-only -i -t fedora /bin/bash

If the container does not write to any file system within the container, including tmpfs, set --read-only-tmpfs=false. Locked down and iconic. 🔥

$ podman run --read-only --read-only-tmpfs=false --tmpfs /run -i -t fedora /bin/bash

🌈 Exposing shared libraries inside of container as read-only using a glob 🌈

$ podman run --mount type=glob,src=/usr/lib64/libnvidia\*,ro=true -i -t fedora /bin/bash

🦋 Exposing log messages from the container to the host's log 🦋

Okay so this one is literally so useful — bind mount the /dev/log directory and suddenly every message logged inside the container shows up in the host's syslog/journal. Bestie. That's the whole trick.

$ podman run -v /dev/log:/dev/log -i -t fedora /bin/bash

Then from inside the container, test it by sending a message to the log:

(bash)# logger "Hello from my container"

Then exit and check the journal:

(bash)# exit
$ journalctl -b | grep Hello

And it lists the message you sent to the logger. Chef's kiss. 👑

🎀 Attaching to one or more from STDIN, STDOUT, STDERR 🎀

If you don't specify the -a option, Podman attaches everything (stdin, stdout, stderr) — all of it, no cap. You override that default by specifying -a (stdin, stdout, stderr), like this:

$ podman run -a stdin -a stdout -i -t fedora /bin/bash

💜 Sharing IPC between containers 💜

Using shm_server.c available here:

https://www.cs.cf.ac.uk/Dave/C/node27.html

Testing --ipc=host mode: the host shows a shared memory segment with 7 pids attached, happens to be from httpd:

$ sudo ipcs -m
------ Shared Memory Segments --------
key        shmid      owner      perms      bytes      nattch     status
0x01128e25 0          root       600        1000       7

Now run a regular container, and it correctly does NOT see the shared memory segment from the host:

$ podman run -it shm ipcs -m
------ Shared Memory Segments --------
key        shmid      owner      perms      bytes      nattch     status

Run a container with the new --ipc=host option, and it now sees the shared memory segment from the host httpd:

$ podman run -it --ipc=host shm ipcs -m
------ Shared Memory Segments --------
key        shmid      owner      perms      bytes      nattch     status
0x01128e25 0          root       600        1000       7

Now testing --ipc=container:id mode. Start a container with a program to create a shared memory segment:

$ podman run -it shm bash
$ sudo shm/shm_server &
$ sudo ipcs -m
------ Shared Memory Segments --------
key        shmid      owner      perms      bytes      nattch     status
0x0000162e 0          root       666        27         1

Create a 2nd container and it correctly shows no shared memory segment from the 1st container:

$ podman run shm ipcs -m
------ Shared Memory Segments --------
key        shmid      owner      perms      bytes      nattch     status

Create a 3rd container using the --ipc=container:id option, and now it shows the shared memory segment from the first:

$ podman run -it --ipc=container:ed735b2264ac shm ipcs -m
$ sudo ipcs -m
------ Shared Memory Segments --------
key        shmid      owner      perms      bytes      nattch     status
0x0000162e 0          root       666        27         1

🔥 Mapping Ports for External Usage 🔥

The exposed port of an application can be mapped to a host port using the -p flag. For example, an httpd port 80 can be mapped to the host port 8080 using the following:

$ podman run -p 8080:80 -d fedora/httpd

🌸 Mounting External Volumes 🌸

To mount a host directory as a container volume, specify the absolute path to the directory and the absolute path for the container directory, separated by a colon. If the source is a named volume maintained by Podman, it is recommended to use its name rather than the path to the volume — otherwise the volume is considered an orphan and gets wiped by the podman volume prune command. 😭 So don't do that.

$ podman run -v /var/db:/data1 -i -t fedora bash
$ podman run -v data:/data2 -i -t fedora bash
$ podman run -v /var/cache/dnf:/var/cache/dnf:O -ti fedora dnf -y update

If the container needs a writable mounted volume by a non-root user inside the container, use the U option. This option tells Podman to chown the source volume to match the default UID and GID used within the container.

$ podman run -d -e MARIADB_ROOT_PASSWORD=root --user mysql --userns=keep-id -v ~/data:/var/lib/mysql:Z,U mariadb

Alternatively, if the container needs a writable volume by a non-root user inside of the container, the --userns=keep-id option allows users to specify the UID and GID of the user executing Podman to specific UIDs and GIDs within the container. Since the processes running in the container run as the user's UID, they can read/write files owned by the user. Slay.

$ podman run -d -e MARIADB_ROOT_PASSWORD=root --user mysql --userns=keep-id:uid=999,gid=999 -v ~/data:/var/lib/mysql:Z mariadb

Using --mount flags to mount a host directory as a container folder: specify the absolute path to the directory or the volume name, and the absolute path within the container directory:

$ podman run --mount type=bind,src=/var/db,target=/data1 busybox sh
$ podman run --mount type=bind,src=volume-name,target=/data1 busybox sh

The --mount flag provides a structured, key-value style for defining mounts inside containers. It's similar in purpose to -v or --volume, but offers greater clarity for complex or multi-option configurations. The general syntax is:

--mount type=<TYPE>,src=<SOURCE>,dst=<TARGET>[,options...]

Supported mount types include bind, volume, tmpfs, artifact, devpts, image, glob, and ramfs. Each type serves a different purpose in how data is attached to the container. Periodt.

🧚 Bind mounts 🧚

Bind mounts directly link a directory or file on the host into the container. Changes made in one are immediately visible in the other. Use bind mounts when both the host and container need access to the same files, such as configuration files or source code.

Example:

podman run --mount type=bind,src=/etc/config,dst=/app/config alpine cat /app/config/file.conf

🍓 Volume mounts 🍓

Volume mounts use Podman-managed named volumes that persist independently of containers. They're ideal for persistent data such as databases or logs, and they're isolated from direct host paths.

Example:

podman volume create mydata
podman run --mount type=volume,src=mydata,dst=/var/lib/data postgres

✨ Tmpfs mounts ✨

A tmpfs mount creates an in-memory filesystem on the host that is mounted inside the container. Data stored here is temporary and removed when the container stops or the host reboots. tmpfs mounts are useful for temporary caches or sensitive data that should not persist to disk.

Example:

podman run --mount type=tmpfs,dst=/cache,tmpfs-size=64m alpine

🫶 Artifact, devpts, image, glob, and ramfs mounts 🫶

Other specialized mount types are available for advanced use cases:

  • artifact — mounts read-only content from a container image or artifact.
  • devpts — provides a pseudo-terminal device inside the container.
  • image — mounts files directly from another container image.
  • glob — mounts multiple host files matching a glob pattern.
  • ramfs — similar to tmpfs but backed directly by system RAM without size limits.

These mount types are less commonly used and often appear in internal or advanced Podman workflows.

In summary, --mount provides a single consistent interface for connecting external storage to containers. Choose the mount type that best fits your use case: bind for direct host access, volume for persistent data managed by Podman, and tmpfs for ephemeral in-memory storage. 💅

⚠️ When using SELinux, be aware that the host has no knowledge of container SELinux policy. Therefore, in the above example, if SELinux policy is enforced, the /var/db directory is not writable to the container. A "Permission Denied" message occurs, and an avc: message is added to the host's syslog.

To work around this, at time of writing this man page, the following command needs to be run in order for the proper SELinux policy type label to be attached to the host directory:

$ chcon -Rt svirt_sandbox_file_t /var/db

Now, writing to the /data1 volume in the container is allowed and the changes are reflected on the host in /var/db. Iconic.

💀 Using alternative security labeling 💀

Override the default labeling scheme for each container by specifying the --security-opt flag. For example, specify the MCS/MLS level — a requirement for MLS systems. Specifying the level in the following command allows the same content to be shared between containers.

podman run --security-opt label=level:s0:c100,c200 -i -t fedora bash

An MLS example might be:

$ podman run --security-opt label=level:TopSecret -i -t rhel7 bash

To disable the security labeling for this container versus running with the --permissive flag, use the following command:

$ podman run --security-opt label=disable -i -t fedora bash

Tighten the security policy on the processes within a container by specifying an alternate type for the container. For example, run a container that is only allowed to listen on Apache ports by executing the following command:

$ podman run --security-opt label=type:svirt_apache_t -i -t centos bash
Note that an SELinux policy defining a svirt_apache_t type must be written.

To mask additional specific paths in the container, specify the paths separated by a colon using the mask option with the --security-opt flag.

$ podman run --security-opt mask=/foo/bar:/second/path fedora bash

To unmask all the paths that are masked by default, set the unmask option to ALL. Or, to only unmask specific paths, specify the paths as shown above with the mask option.

$ podman run --security-opt unmask=ALL fedora bash

To unmask all the paths that start with /proc, set the unmask option to /proc/*.

$ podman run --security-opt unmask=/proc/* fedora bash
$ podman run --security-opt unmask=/foo/bar:/sys/firmware fedora bash

⭐ Setting device weight via --blkio-weight-device flag ⭐

$ podman run -it --blkio-weight-device "/dev/sda:200" ubuntu

🌈 Using a podman container with input from a pipe 🌈

$ echo "asdf" | podman run --rm -i --entrypoint /bin/cat someimage
asdf

👑 Setting automatic user namespace separated containers 👑

# podman run --userns=auto:size=65536 ubi8-micro cat /proc/self/uid_map
0 2147483647      65536
# podman run --userns=auto:size=65536 ubi8-micro cat /proc/self/uid_map
0 2147549183      65536

🎀 Setting Namespaced Kernel Parameters (Sysctls) 🎀

The --sysctl sets namespaced kernel parameters (sysctls) in the container. For example, to turn on IP forwarding in the container's network namespace, run this command:

$ podman run --sysctl net.ipv4.ip_forward=1 someimage
Heads up, bestie: not all sysctls are namespaced. Podman does not support changing sysctls inside of a container that also modify the host system. As the kernel evolves, we expect to see more sysctls become namespaced.

See the definition of the --sysctl option above for the current list of supported sysctls. 💖

💖 Set UID/GID mapping in a new user namespace 💖

Running a container in a new user namespace needs a mapping of the UIDs and GIDs from the host, bestie. Like, it literally cannot vibe without it.

$ podman run --uidmap 0:30000:7000 --gidmap 0:30000:7000 fedora echo hello

🎀 Configuring Storage Options from the command line 🎀

Podman lets you configure storage by changing the values in /etc/container/storage.conf or by using global options. This example shows how to use an additional image store for a one-time run of busybox using global options. So iconic of her.

podman --log-level=debug --storage-opt "additionalimagestore=/tmp/readonly-images" run busybox /bin/sh

🌸 Configure timezone in a container 🌸

$ podman run --tz=local alpine date
$ podman run --tz=Asia/Shanghai alpine date
$ podman run --tz=US/Eastern alpine date

🦋 Adding dependency containers 🦋

The first container, container1, is not started initially, but must be running before container2 starts. The podman run command starts the container automatically before starting container2. Lowkey magical.

$ podman create --name container1 -t -i fedora bash
$ podman run --name container2 --requires container1 -t -i fedora bash

Multiple containers can be required. Not just one bestie — a whole friend group.

$ podman create --name container1 -t -i fedora bash
$ podman create --name container2 -t -i fedora bash
$ podman run --name container3 --requires container1,container2 -t -i fedora bash

🧚 Configure keep supplemental groups for access to volume 🧚

$ podman run -v /var/lib/design:/var/lib/design --group-add keep-groups ubi8

⭐ Configure execution domain for containers using personality flag ⭐

$ podman run --name container1 --personality=LINUX32 fedora bash

🌈 Run a container with external rootfs mounted as an overlay 🌈

$ podman run --name container1 --rootfs /path/to/rootfs:O bash

👑 Handling Timezones in java applications in a container. 👑

In order to use a timezone other than UTC when running a Java application within a container, the TZ environment variable must be set within the container. Java applications ignores the value set with the --tz option. Rude, but true.

# Example run
podman run -ti --rm  -e TZ=EST mytzimage
lrwxrwxrwx. 1 root root 29 Nov  3 08:51 /etc/localtime -> ../usr/share/zoneinfo/Etc/UTC
Now with default timezone:
Fri Nov 19 18:10:55 EST 2021
Java default sees the following timezone:
2021-11-19T18:10:55.651130-05:00
Forcing UTC:
Fri Nov 19 23:10:55 UTC 2021

💅 Run a container connected to two networks (called net1 and net2) with a static ip 💅

$ podman run --network net1:ip=10.89.1.5 --network net2:ip=10.89.10.10 alpine ip addr

🍓 Define a healthcheck at run time and verify it 🍓

$ podman create -d --name healthcheck-example -p 8080:80
  --health-cmd '["CMD-SHELL","wget -q --spider http://127.0.0.1:80 || exit 1"]'
  --health-interval 30s --health-timeout 5s --health-retries 3
  docker.io/library/nginx:alpine

$ podman inspect healthcheck-example | jq '.[0].State.Health'

🫶 Rootless Containers 🫶

Podman runs as a non-root user on most systems. This feature requires that a new enough version of shadow-utils be installed. The shadow-utils package must include the newuidmap(1) and newgidmap(1) executables.

In order for users to run rootless, there must be an entry for their username in /etc/subuid and /etc/subgid which lists the UIDs for their user namespace.

Rootless Podman works better if the fuse-overlayfs package is installed. The fuse-overlayfs package provides a userspace overlay storage driver, otherwise users need to use the vfs storage driver, which can be disk space expensive and less performant than other drivers. So like, install the cute package if you can. 💖

To enable VPN on the container, pasta networking is used by default; otherwise, containers need to be run with the --network=host flag.

🌟 ENVIRONMENT 🌟

Environment variables within containers can be set using multiple different options, in the following order of precedence (later entries override earlier entries):

  • Container image: Any environment variables specified in the container image.
  • --http-proxy: By default, several environment variables are passed in from the host, such as http_proxy and no_proxy. See --http-proxy for details.
  • --env-host: Host environment of the process executing Podman is added.
  • --env-file: Any environment variables specified via env-files. If multiple files are specified, then they override each other in order of entry.
  • --env: Any environment variables specified overrides previous settings.

Run containers and set the environment ending with a *. The trailing * glob functionality is only active when no value is specified:

$ export ENV1=a
$ podman run --env 'ENV*' alpine env | grep ENV
ENV1=a
$ podman run --env 'ENV*=b' alpine env | grep ENV
ENV*=b

👀 CONMON 👀

When Podman starts a container it actually executes the conmon program, which then executes the OCI Runtime. Conmon is the container monitor. It is a small program whose job is to watch the primary process of the container, and if the container dies, save the exit code. It also holds open the tty of the container, so that it can be attached to later. This is what allows Podman to run in detached mode (backgrounded), so Podman can exit but conmon continues to run. Each container has their own instance of conmon. Conmon waits for the container to exit, gathers and saves the exit code, and then launches a Podman process to complete the container cleanup, by shutting down the network and storage. For more information about conmon, see the conmon(8) man page.

📁 FILES 📁

/etc/subuid

/etc/subgid

NOTE: Use the environment variable TMPDIR to change the temporary storage location of downloaded container images. Podman defaults to use /var/tmp.

💖 SEE ALSO 💖

podman(1), podman-save(1), podman-ps(1), podman-attach(1), podman-pod-create(1), podman-port(1), podman-start(1), podman-kill(1), podman-stop(1), podman-generate-systemd(1), podman-rm(1), subgid(5), subuid(5), containers.conf(5), podman-systemd.unit(5), setsebool(8), pasta(1), fuse-overlayfs(1), proc(5), conmon(8), personality(2)

🌈 Troubleshooting 🌈

See podman-troubleshooting(7) for solutions to common issues.

See podman-rootless(7) for rootless issues.

🕰️ HISTORY 🕰️

  • September 2018, updated by Kunal Kushwaha <kushwaha_kunal_v7@lab.ntt.co.jp>
  • October 2017, converted from Docker documentation to Podman by Dan Walsh for Podman <dwalsh@redhat.com>
  • November 2015, updated by Sally O'Malley <somalley@redhat.com>
  • June 2014, updated by Sven Dowideit <SvenDowideit@home.org.au>
  • April 2014, Originally compiled by William Henry <whenry@redhat.com> based on docker.com source material and internal work.

🫶 FOOTNOTES 🫶

1: The Podman project is committed to inclusivity, a core value of open source. The master and slave mount propagation terminology used here is problematic and divisive, and needs to be changed. However, these terms are currently used within the Linux kernel and must be used as-is at this time. When the kernel maintainers rectify this usage, Podman will follow suit immediately.

Footer check, bestie: that’s source/version podman-run(1), with the date stamp down there doing its official little bow — and that’s the whole page. Go run something iconic. 💅✨🫶

📦 where this page comes from