Change MTU for rootless containers

[volker-raschek]

Hi,
I am experiencing an issue where the MTU of the default podman network is 65520, which causes connection drops.

After a deeper analysis, I found that if I create a custom network with an MTU of 1500, the connection drops do not occur. The question I have is: How can I set the MTU to 1500 for both root and rootless containers?

I tried to solve the problem myself. I edited the file /etc/containers/containers.conf, defined slirp4netns as the default_rootless_network_cmd, and added --mtu=1500 to engine.network_cmd_options.

When I start a container as root, I see that the MTU is correctly set to 1500:

$ sudo podman run --rm docker.io/library/busybox:latest ip link show
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
2: eth0@if20: <BROADCAST,MULTICAST,UP,LOWER_UP,M-DOWN> mtu 1500 qdisc noqueue qlen 1000
    link/ether 06:1a:8b:1d:2c:b5 brd ff:ff:ff:ff:ff:ff

However, running the same command without sudo results in the following error:

$ podman run --rm docker.io/library/busybox:latest ip link show
Error: unknown option for slirp4netns: "--mtu=1500"

When I run slirp4netns now, I see that there is a --mtu flag, which has the default of MTU 1500, but seems to be set by podman. I don't understand why podman has a problem with this option, when I configure it via containers.conf to restore it's default behavior. Can anyone help me with this? I can't solve this problem on my own.

$ slirp4netns -h
Usage: slirp4netns [OPTION]... PID|PATH|FD [TAPNAME]
User-mode networking for unprivileged network namespaces.

-c, --configure          bring up the interface
-e, --exit-fd=FD         specify the FD for terminating slirp4netns
-r, --ready-fd=FD        specify the FD to write to when the network is configured
-m, --mtu=MTU            specify MTU (default=1500, max=65521)
-6, --enable-ipv6        enable IPv6 (experimental)
-a, --api-socket=PATH    specify API socket path
--cidr=CIDR              specify network address CIDR (default=10.0.2.0/24)
--disable-host-loopback  prohibit connecting to 127.0.0.1:* on the host namespace
--netns-type=TYPE 	 specify network namespace type ([path|pid|tapfd], default=pid)
--userns-path=PATH	 specify user namespace path
--enable-sandbox         create a new mount namespace (and drop all caps except CAP_NET_BIND_SERVICE if running as the root)
--enable-seccomp         enable seccomp to limit syscalls (experimental)
--outbound-addr=IPv4     sets outbound ipv4 address to bound to (experimental)
--outbound-addr6=IPv6    sets outbound ipv6 address to bound to (experimental)
--disable-dns            disables 10.0.2.3 (or configured internal ip) to host dns redirect (experimental)
--macaddress=MAC         specify the MAC address of the TAP (only valid with -c)
--target-type=TYPE       specify the target type ([netns|bess], default=netns)
-h, --help               show this help and exit
-v, --version            show version and exit

Here is some additional information:

host:
  arch: amd64
  buildahVersion: 1.42.2
  cgroupControllers:
  - memory
  - pids
  cgroupManager: systemd
  cgroupVersion: v2
  conmon:
    package: conmon-1:2.1.13-1
    path: /usr/bin/conmon
    version: 'conmon version 2.1.13, commit: 82de887596ed8ee6d9b2ee85e4f167f307bb569b'
  cpuUtilization:
    idlePercent: 97.79
    systemPercent: 0.44
    userPercent: 1.77
  cpus: 24
  databaseBackend: sqlite
  distribution:
    distribution: arch
    version: unknown
  eventLogger: journald
  freeLocks: 2048
  hostname: markus-pc
  idMappings:
    gidmap:
    - container_id: 0
      host_id: 1000
      size: 1
    - container_id: 1
      host_id: 200000
      size: 65536
    uidmap:
    - container_id: 0
      host_id: 1000
      size: 1
    - container_id: 1
      host_id: 200000
      size: 65536
  kernel: 6.12.63-1-lts
  linkmode: dynamic
  logDriver: journald
  memFree: 28432691200
  memTotal: 67337887744
  networkBackend: netavark
  networkBackendInfo:
    backend: netavark
    dns:
      package: aardvark-dns-1.17.0-1
      path: /usr/lib/podman/aardvark-dns
      version: aardvark-dns 1.17.0
    package: netavark-1.17.1-1
    path: /usr/lib/podman/netavark
    version: netavark 1.17.1
  ociRuntime:
    name: crun
    package: crun-1.25.1-1
    path: /usr/bin/crun
    version: |-
      crun version 1.25.1
      commit: 156ae065d4a322d149c7307034f98d9637aa92a2
      rundir: /run/user/1000/crun
      spec: 1.0.0
      +SYSTEMD +SELINUX +APPARMOR +CAP +SECCOMP +EBPF +CRIU +LIBKRUN +YAJL
  os: linux
  pasta:
    executable: /usr/bin/pasta
    package: passt-2025_12_23.2ba9fd5-1
    version: |
      pasta 2025_12_23.2ba9fd5
      Copyright Red Hat
      GNU General Public License, version 2 or later
        <https://www.gnu.org/licenses/old-licenses/gpl-2.0.html>
      This is free software: you are free to change and redistribute it.
      There is NO WARRANTY, to the extent permitted by law.
  remoteSocket:
    exists: true
    path: /run/user/1000/podman/podman.sock
  rootlessNetworkCmd: slirp4netns
  security:
    apparmorEnabled: false
    capabilities: CAP_CHOWN,CAP_DAC_OVERRIDE,CAP_FOWNER,CAP_FSETID,CAP_KILL,CAP_NET_BIND_SERVICE,CAP_SETFCAP,CAP_SETGID,CAP_SETPCAP,CAP_SETUID,CAP_SYS_CHROOT
    rootless: true
    seccompEnabled: true
    seccompProfilePath: /etc/containers/seccomp.json
    selinuxEnabled: false
  serviceIsRemote: false
  slirp4netns:
    executable: /usr/bin/slirp4netns
    package: slirp4netns-1.3.3-1
    version: |-
      slirp4netns version 1.3.3
      commit: 944fa94090e1fd1312232cbc0e6b43585553d824
      libslirp: 4.9.1
      SLIRP_CONFIG_VERSION_MAX: 6
      libseccomp: 2.5.6
  swapFree: 31689011200
  swapTotal: 31689011200
  uptime: 3h 40m 56.00s (Approximately 0.12 days)
  variant: ""
plugins:
  authorization: null
  log:
  - k8s-file
  - none
  - passthrough
  - journald
  network:
  - bridge
  - macvlan
  - ipvlan
  volume:
  - local
registries: []
store:
  configFile: /home/markus/.config/containers/storage.conf
  containerStore:
    number: 0
    paused: 0
    running: 0
    stopped: 0
  graphDriverName: btrfs
  graphOptions: {}
  graphRoot: /home/markus/.local/share/containers/storage
  graphRootAllocated: 966367641600
  graphRootUsed: 360473034752
  graphStatus:
    Build Version: Btrfs v6.17.1
    Library Version: "104"
  imageCopyTmpDir: /var/tmp
  imageStore:
    number: 32
  runRoot: /run/user/1000/containers
  transientStore: false
  volumePath: /home/markus/.local/share/containers/storage/volumes
version:
  APIVersion: 5.7.1
  Built: 1765447754
  BuiltTime: Thu Dec 11 11:09:14 2025
  GitCommit: f845d14e941889ba4c071f35233d09b29d363c75
  GoVersion: go1.25.5 X:nodwarf5
  Os: linux
  OsArch: linux/amd64
  Version: 5.7.1

$ cat /etc/containers/containers.conf
# The containers configuration file specifies all of the available configuration
# command-line options/flags for container engine tools like Podman & Buildah,
# but in a TOML format that can be easily modified and versioned.

# Please refer to containers.conf(5) for details of all configuration options.
# Not all container engines implement all of the options.
# All of the options have hard coded defaults and these options will override
# the built in defaults. Users can then override these options via the command
# line. Container engines will read containers.conf files in up to three
# locations in the following order:
#  1. /usr/share/containers/containers.conf
#  2. /etc/containers/containers.conf
#  3. $XDG_CONFIG_HOME/containers/containers.conf or
#     $HOME/.config/containers/containers.conf if $XDG_CONFIG_HOME is not set
#  Items specified in the latter containers.conf, if they exist, override the
# previous containers.conf settings, or the default settings.

[containers]

# List of annotation. Specified as
# "key = value"
# If it is empty or commented out, no annotations will be added
#
#annotations = []

# Used to change the name of the default AppArmor profile of container engine.
#
#apparmor_profile = "container-default"

# Base file 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 the following
# special flags:
# ""      Use the host's `/etc/hosts` file (the default)
# `none`  Do not use a base file (i.e. start with an empty file)
# `image` Use the container image's `/etc/hosts` file as base file
#
#base_hosts_file = ""

# List of cgroup_conf entries specifying a list of cgroup files to write to and
# their values. For example `memory.high=1073741824` sets the
# memory.high limit to 1GB.
#
# cgroup_conf = []

# Default way to to create a cgroup namespace for the container
# Options are:
# `private` Create private Cgroup Namespace for the container.
# `host`    Share host Cgroup Namespace with the container.
#
#cgroupns = "private"

# Control container cgroup configuration
# Determines  whether  the  container will create CGroups.
# Options are:
# `enabled`   Enable cgroup support within container
# `disabled`  Disable cgroup support, will inherit cgroups from parent
# `no-conmon` Do not create a cgroup dedicated to conmon.
#
#cgroups = "enabled"

# When no hostname is set for a container, use the container's name, with
# characters not valid for a hostname removed, as the hostname instead of
# the first 12 characters of the container's ID. Containers not running
# in a private UTS namespace will have their hostname set to the host's
# hostname regardless of this setting.
#
#container_name_as_hostname = false

# List of default capabilities for containers. If it is empty or commented out,
# the default capabilities defined in the container engine will be added.
#
#default_capabilities = [
#  "CHOWN",
#  "DAC_OVERRIDE",
#  "FOWNER",
#  "FSETID",
#  "KILL",
#  "NET_BIND_SERVICE",
#  "SETFCAP",
#  "SETGID",
#  "SETPCAP",
#  "SETUID",
#  "SYS_CHROOT",
#]

# A list of sysctls to be set in containers by default,
# specified as "name=value",
# for example:"net.ipv4.ping_group_range=0 0".
#
default_sysctls = [
  "net.ipv4.ping_group_range=0 0",
]

# A list of ulimits to be set in containers by default, specified as
# "<ulimit name>=<soft limit>:<hard limit>", for example:
# "nofile=1024:2048"
# See setrlimit(2) for a list of resource names.
# Any limit not specified here will be inherited from the process launching the
# container engine.
# Ulimits has limits for non privileged container engines.
#
#default_ulimits = [
#  "nofile=1280:2560",
#]

# List of devices. Specified as
# "<device-on-host>:<device-on-container>:<permissions>", for example:
# "/dev/sdc:/dev/xvdc:rwm".
# If it is empty or commented out, only the default devices will be used
#
#devices = []

# List of default DNS options to be added to /etc/resolv.conf inside of the container.
#
#dns_options = []

# List of default DNS search domains to be added to /etc/resolv.conf inside of the container.
#
#dns_searches = []

# Set default DNS servers.
# This option can be used to override the DNS configuration passed to the
# container. The special value "none" can be specified to disable creation of
# /etc/resolv.conf in the container.
# The /etc/resolv.conf file in the image will be used without changes.
#
#dns_servers = []

# Environment variable list for the conmon process; used for passing necessary
# environment variables to conmon or the runtime.
#
#env = [
#  "PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin",
#]

# Pass all host environment variables into the container.
#
#env_host = false

# Set the IP address the container should expect to connect to the host. The IP
# address is used by Podman to automatically add the `host.containers.internal`
# and `host.docker.internal` hostnames to the container's `/etc/hosts` file. It
# is also used for the *host-gateway* flag of Podman's `--add-host` CLI option.
# If no IP address is configured (the default), Podman will try to determine it
# automatically, but might fail to do so depending on the container's network
# setup. Adding these internal hostnames to `/etc/hosts` is silently skipped then.
# Set this config to `none` to never add the internal hostnames to `/etc/hosts`.
#
# Note: 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. This config
# has no effect on gvproxy. However, since `/etc/hosts` bypasses the DNS resolver,
# a manually configured IP address still takes precedence.
#
# Note: This config doesn't affect the actual network setup, it just tells Podman
# the IP address it should expect. Configuring an IP address here doesn't ensure
# that the container can actually reach the host using this IP address.
#
#host_containers_internal_ip = ""

# Default proxy environment variables passed into the container.
# The environment variables passed in include:
# http_proxy, https_proxy, ftp_proxy, no_proxy, and the upper case versions of
# these. This option is needed when host system uses a proxy but container
# should not use proxy. Proxy environment variables specified for the container
# in any other way will override the values passed from the host.
#
#http_proxy = true

# Run an init inside the container that forwards signals and reaps processes.
#
#init = false

# Container init binary, if init=true, this is the init binary to be used for containers.
# If this option is not set catatonit is searched in the directories listed under
# the helper_binaries_dir option. It is recommended to just install catatonit
# there instead of configuring this option here.
#
#init_path = "/usr/lib/podman/catatonit"

# Default way to to create an IPC namespace (POSIX SysV IPC) for the container
# Options are:
#  "host"     Share host IPC Namespace with the container.
#  "none"     Create shareable IPC Namespace for the container without a private /dev/shm.
#  "private"  Create private IPC Namespace for the container, other containers are not allowed to share it.
#  "shareable" Create shareable IPC Namespace for the container.
#
#ipcns = "shareable"

# Default way to set an interface name inside container. Defaults to legacy
# pattern of ethX, where X is a integer, when left undefined.
# Options are:
#  "device"   Uses the network_interface name from the network config as interface name.
#             Falls back to the ethX pattern if the network_interface is not set.
#interface_name = ""

# keyring tells the container engine whether to create
# a kernel keyring for use within the container.
#
#keyring = true

# label tells the container engine whether to use container separation using
# MAC(SELinux) labeling or not.
# The label flag is ignored on label disabled systems.
#
#label = true

# label_users indicates whether to enforce confined users in containers on
# SELinux systems. This option causes containers to maintain the current user
# and role field of the calling process. By default SELinux containers run with
# the user system_u, and the role system_r.
#label_users = false

# Logging driver for the container. Available options: k8s-file and journald.
#
#log_driver = "k8s-file"

# Default path for container logs to be stored in. When empty, logs will be stored
# in the container's default storage and removed when the container is removed.
# A subdirectory named with the container ID will be created under the specified
# path, and the log file will have the default name `ctr.log` within that directory.
# This option can be overridden by the `--log-opt` flag.
#
#log_path = ""

# Maximum size allowed for the container log file. Negative numbers indicate
# that no size limit is imposed. If positive, it must be >= 8192 to match or
# exceed conmon's read buffer. The file is truncated and re-opened so the
# limit is never exceeded.
#
#log_size_max = -1

# Specifies default format tag for container log messages.
# This is useful for creating a specific tag for container log messages.
# Containers logs default to truncated container ID as a tag.
#
#log_tag = ""

# List of mounts. Specified as
# "type=TYPE,source=<directory-on-host>,destination=<directory-in-container>,<options>", for example:
# "type=bind,source=/var/lib/foobar,destination=/var/lib/foobar,ro".
# If it is empty or commented out, no mounts will be added
#
#mounts = []

# Default way to create a NET namespace for the container.
# The option is mapped to the **--network** argument for the podman commands, it accepts the same values as that option.
# For example it can be set to `bridge`, `host`, `none`, `pasta` and more, see the podman-create(1)
# manual for all available options.
#
#netns = ""

# Do not modify the `/etc/hosts` file in the container. Podman assumes control
# over the container's `/etc/hosts` file by default; refer to the `--add-host`
# CLI option for details. To disable this, either set this config to `true`, or
# use the functionally identical `--no-hosts` CLI option.
#
#no_hosts = false

# Tune the host's OOM preferences for containers
# (accepts values from -1000 to 1000).
#oom_score_adj = 0

# Default way to to create a PID namespace for the container
# Options are:
# `private` Create private PID Namespace for the container.
# `host`    Share host PID Namespace with the container.
#
#pidns = "private"

# Maximum number of processes allowed in a container.
#
#pids_limit = 2048

# Copy the content from the underlying image into the newly created volume
# when the container is created instead of when it is started. If false,
# the container engine will not copy the content until the container is started.
# Setting it to true may have negative performance implications.
#
#prepare_volume_on_create = false

# Give extended privileges to all containers. 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. Due to the disabled
# security features the privileged field should almost never be set as
# containers can easily break out of confinment.
#
# Containers running in a user namespace (e.g., rootless containers) cannot
# have more privileges than the user that launched them.
#
#privileged = false

# Run all containers with root file system mounted read-only
#
# read_only = false

# Path to the seccomp.json profile which is used as the default seccomp profile
# for the runtime.
#
#seccomp_profile = "/usr/share/containers/seccomp.json"

# Size of /dev/shm. Specified as <number><unit>.
# Unit is optional, values:
# b (bytes), k (kilobytes), m (megabytes), or g (gigabytes).
# If the unit is omitted, the system uses bytes.
#
#shm_size = "65536k"

# Set timezone in container. Takes IANA timezones as well as "local",
# which sets the timezone in the container to match the host machine.
#
#tz = ""

# Set umask inside the container
#
#umask = "0022"

# Default way to create a USER namespace for the container.
# Options are:
#   `private` Create private USER Namespace for the container, without adding any UID mappings.
#   `host`    Share host USER Namespace with the container. Root in the container is mapped to the host user UID.
#   `auto`    Automatically create a USER namespace with a unique mapping.
#   `keep-id` Like `private`, but container UIDs are mapped to the host user's subordinate UIDs listed in `/etc/subuid`, and the current user's `UID:GID` are mapped to the same values in the container.
#   `no-map`  Like `keep-id`, but the current user's `UID:GID` does not map to any `UID:GID` inside the container.
#userns = "host"

# Default way to to create a UTS namespace for the container
# Options are:
# `private`        Create private UTS Namespace for the container.
# `host`    Share host UTS Namespace with the container.
#
#utsns = "private"

# List of volumes. Specified as
# "<directory-on-host>:<directory-in-container>:<options>", for example:
# "/db:/var/lib/db:ro".
# If it is empty or commented out, no volumes will be added
#
#volumes = []

#[engine.platform_to_oci_runtime]
#"wasi/wasm" = ["crun-wasm"]
#"wasi/wasm32" = ["crun-wasm"]
#"wasi/wasm64" = ["crun-wasm"]

[secrets]
#driver = "file"

[secrets.opts]
#root = "/example/directory"

[network]

# Network backend determines what network driver will be used to set up and tear down container networks.
# Valid values are "cni" and "netavark".
# The default value is empty which means that it will automatically choose CNI or netavark. If there are
# already containers/images or CNI networks preset it will choose CNI.
#
# Before changing this value all containers must be stopped otherwise it is likely that
# iptables rules and network interfaces might leak on the host. A reboot will fix this.
#
network_backend = "netavark"

# Path to directory where CNI plugin binaries are located.
#
#cni_plugin_dirs = [
#  "/usr/local/libexec/cni",
#  "/usr/libexec/cni",
#  "/usr/local/lib/cni",
#  "/usr/lib/cni",
#  "/opt/cni/bin",
#]

# List of directories that will be searched for netavark plugins.
#
#netavark_plugin_dirs = [
#  "/usr/local/libexec/netavark",
#  "/usr/libexec/netavark",
#  "/usr/local/lib/netavark",
#  "/usr/lib/netavark",
#]

# The firewall driver to be used by netavark.
# The default is empty which means netavark will pick one accordingly. Current supported
# drivers are "iptables", "nftables", "none" (no firewall rules will be created) and "firewalld" (firewalld is
# experimental at the moment and not recommend outside of testing).
#
firewall_driver = "nftables"


# The name of the default network as seen in `podman network ls`. This option only effects the network assignment when
# the bridge network mode is selected, i.e. `--network bridge`. It is the default for rootful containers but not as
# rootless. To change the default network mode use the **netns** option under the `[containers]` table.
#
# Note: This should not be changed while you have any containers using this network.
#
default_network = "podman"

# The default subnet for the default network given in default_network.
#
# Note: This should not be changed if any containers are currently running on the default network.
#
#default_subnet = "10.88.0.0/16"

# DefaultSubnetPools is a list of subnets and size which are used to
# allocate subnets automatically for podman network create.
# It will iterate through the list and will pick the first free subnet
# with the given size. This is only used for ipv4 subnets, ipv6 subnets
# are always assigned randomly.
#
#default_subnet_pools = [
#  {"base" = "10.89.0.0/16", "size" = 24},
#  {"base" = "10.90.0.0/15", "size" = 24},
#  {"base" = "10.92.0.0/14", "size" = 24},
#  {"base" = "10.96.0.0/11", "size" = 24},
#  {"base" = "10.128.0.0/9", "size" = 24},
#]



# Configure which rootless network program to use by default. Valid options are
# `slirp4netns` and `pasta` (default).
#
default_rootless_network_cmd = "slirp4netns"

# Path to the directory where network configuration files are located.
# For the CNI backend the default is "/etc/cni/net.d" as root
# and "$HOME/.config/cni/net.d" as rootless.
# For the netavark backend "/etc/containers/networks" is used as root
# and "$graphroot/networks" as rootless.
#
#network_config_dir = "/etc/cni/net.d/"

# Port to use for dns forwarding daemon with netavark in rootful bridge
# mode and dns enabled.
# Using an alternate port might be useful if other dns services should
# run on the machine.
#
#dns_bind_port = 53

#	A list of default pasta options that should be used running pasta.
# It accepts the pasta cli options, see pasta(1) for the full list of options.
#
#pasta_options = []

[engine]
# Index to the active service
#
#active_service = "production"

#List of compression algorithms. If set makes sure that requested compression variant
#for each platform is added to the manifest list keeping original instance intact in
#the same manifest list on every `manifest push`. Supported values are (`gzip`, `zstd` and `zstd:chunked`).
#`zstd:chunked` is incompatible with encrypting images, and will be treated as `zstd` with a warning
#in that case.
#
#add_compression = ["gzip", "zstd", "zstd:chunked"]

# Enforces using docker.io for completing short names in Podman's compatibility
# REST API. Note that this will ignore unqualified-search-registries and
# short-name aliases defined in containers-registries.conf(5).
#compat_api_enforce_docker_hub = true

# Specify one or more external providers for the compose command.  The first
# found provider is used for execution.  Can be an absolute and relative path
# or a (file) name.
#compose_providers=[]

# Emit logs on each invocation of the compose command indicating that an
# external compose provider is being executed.
#compose_warning_logs = true

# The compression format to use when pushing an image.
# Valid options are: `gzip`, `zstd` and `zstd:chunked`.
# This field is ignored when pushing images to the docker-daemon and
# docker-archive formats. It is also ignored when the manifest format is set
# to v2s2.
# `zstd:chunked` is incompatible with encrypting images, and will be treated as `zstd` with a warning
# in that case.
#
#compression_format = "gzip"

# The compression level to use when pushing an image.
# Valid options depend on the compression format used.
# For gzip, valid options are 1-9, with a default of 5.
# For zstd, valid options are 1-20, with a default of 3.
#
#compression_level = 5

# Cgroup management implementation used for the runtime.
# Valid options "systemd" or "cgroupfs"
#
cgroup_manager = "systemd"

# Environment variables to pass into conmon
#
#conmon_env_vars = [
#  "PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin"
#]

# Paths to look for the conmon container manager binary
#
#conmon_path = [
#  "/usr/libexec/podman/conmon",
#  "/usr/local/libexec/podman/conmon",
#  "/usr/local/lib/podman/conmon",
#  "/usr/bin/conmon",
#  "/usr/sbin/conmon",
#  "/usr/local/bin/conmon",
#  "/usr/local/sbin/conmon"
#]

# Enforces using docker.io for completing short names in Podman's compatibility
# REST API. Note that this will ignore unqualified-search-registries and
# short-name aliases defined in containers-registries.conf(5).
#compat_api_enforce_docker_hub = true

# The database backend of Podman.  Supported values are "" (default), "boltdb"
# and "sqlite". An empty value means it will check whenever a boltdb already
# exists and use it when it does, otherwise it will use sqlite as default
# (e.g. new installs). This allows for backwards compatibility with older versions.
# Please run `podman-system-reset` prior to changing the database
# backend of an existing deployment, to make sure Podman can operate correctly.
#
#database_backend = ""

# Specify the keys sequence used to detach a container.
# Format is a single character [a-Z] or a comma separated sequence of
# `ctrl-<value>`, where `<value>` is one of:
# `a-z`, `@`, `^`, `[`, `\`, `]`, `^` or `_`
# Specifying "" disables this feature.
#detach_keys = "ctrl-p,ctrl-q"

# Determines whether engine will reserve ports on the host when they are
# forwarded to containers. When enabled, when ports are forwarded to containers,
# ports are held open by as long as the container is running, ensuring that
# they cannot be reused by other programs on the host. However, this can cause
# significant memory usage if a container has many ports forwarded to it.
# Disabling this can save memory.
#
#enable_port_reservation = true

# Environment variables to be used when running the container engine (e.g.,
# Podman, Buildah). For example "MYVAR=value". These environment variables
# will not be used within the container. Set the env section under the
# [containers] table, if you want to set environment variables for the
# container.
# Note when using this to set http proxy variables then they might get
# leaked into the container depending on if `http_proxy` (under the
# [containers] table) is set to to true (default) or false.
#
#env = []

# Define where event logs will be stored, when events_logger is "file".
#events_logfile_path=""

# Sets the maximum size for events_logfile_path.
# The size can be b (bytes), k (kilobytes), m (megabytes), or g (gigabytes).
# The format for the size is `<number><unit>`, e.g., `1b` or `3g`.
# If no unit is included then the size will be read in bytes.
# When the limit is exceeded, the logfile will be rotated and the old one will be deleted.
# If the maximum size is set to 0, then no limit will be applied,
# and the logfile will not be rotated.
#events_logfile_max_size = "1m"

# Selects which logging mechanism to use for container engine events.
# Valid values are `journald`, `file` and `none`.
#
events_logger = "journald"

# Creates a more verbose container-create event which includes a JSON payload
# with detailed information about the container.
#events_container_create_inspect_data = false

# Whenever Podman should log healthcheck events.
# With many running healthcheck on short interval Podman will spam the event
# log a lot as it generates a event for each single healthcheck run. Because
# this event is optional and only useful to external consumers that may want
# to know when a healthcheck is run or failed allow users to turn it off by
# setting it to false. The default is true.
#
#healthcheck_events = true

# A is a list of directories which are used to search for helper binaries.
#
#helper_binaries_dir = [
#  "/usr/local/libexec/podman",
#  "/usr/local/lib/podman",
#  "/usr/libexec/podman",
#  "/usr/lib/podman",
#]

# Path to OCI hooks directories for automatically executed hooks.
#
#hooks_dir = [
#  "/usr/share/containers/oci/hooks.d",
#]

# Directories to scan for CDI Spec files.
#
#cdi_spec_dirs = [
#  "/etc/cdi",
#  "/var/run/cdi",
#]

# Manifest Type (oci, v2s2, or v2s1) to use when pulling, pushing, building
# container images. By default image pulled and pushed match the format of the
# source image. Building/committing defaults to OCI.
#
#image_default_format = ""

# Default transport method for pulling and pushing for images
#
#image_default_transport = "docker://"

# Maximum number of image layers to be copied (pulled/pushed) simultaneously.
# Not setting this field, or setting it to zero, will fall back to containers/image defaults.
#
#image_parallel_copies = 0

# Tells container engines how to handle the built-in image volumes.
#   * anonymous: An anonymous named volume will be created and mounted
#     into the container.
#   * tmpfs: The volume is mounted onto the container as a tmpfs,
#     which allows users to create content that disappears when
#     the container is stopped.
#   * ignore: All volumes are just ignored and no action is taken.
#
#image_volume_mode = ""

# Default command to run the infra container
#
#infra_command = "/pause"

# Infra (pause) container image name for pod infra containers.  When running a
# pod, we start a `pause` process in a container to hold open the namespaces
# associated with the  pod.  This container does nothing other than sleep,
# reserving the pod's resources for the lifetime of the pod. By default container
# engines run a built-in container using the pause executable. If you want override
# specify an image to pull.
#
#infra_image = ""

# Default Kubernetes kind/specification of the kubernetes yaml generated with the `podman kube generate` command.
# The possible options are `pod` and `deployment`.
#kube_generate_type = "pod"

# Specify the locking mechanism to use; valid values are "shm" and "file".
# Change the default only if you are sure of what you are doing, in general
# "file" is useful only on platforms where cgo is not available for using the
# faster "shm" lock type. You may need to run "podman system renumber" after
# you change the lock type.
#
#lock_type = "shm"

# MultiImageArchive - if true, the container engine allows for storing archives
# (e.g., of the docker-archive transport) with multiple images.  By default,
# Podman creates single-image archives.
#
#multi_image_archive = false

# Default engine namespace
# If engine is joined to a namespace, it will see only containers and pods
# that were created in the same namespace, and will create new containers and
# pods in that namespace.
# The default namespace is "", which corresponds to no namespace. When no
# namespace is set, all containers and pods are visible.
#
#namespace = ""

# Path to the slirp4netns binary
#
network_cmd_path = "/usr/bin/slirp4netns"

# Default options to pass to the slirp4netns binary.
# Valid options values are:
#
# - allow_host_loopback=true|false: Allow the slirp4netns to reach the host loopback IP (`10.0.2.2`).
#   Default is false.
# - mtu=MTU: Specify the MTU to use for this network. (Default is `65520`).
# - cidr=CIDR: Specify ip range to use for this network. (Default is `10.0.2.0/24`).
# - enable_ipv6=true|false: Enable IPv6. Default is true. (Required for `outbound_addr6`).
# - outbound_addr=INTERFACE: Specify the outbound interface slirp should bind to (ipv4 traffic only).
# - outbound_addr=IPv4: Specify the outbound ipv4 address slirp should bind to.
# - outbound_addr6=INTERFACE: Specify the outbound interface slirp should bind to (ipv6 traffic only).
# - outbound_addr6=IPv6: Specify the outbound ipv6 address slirp should bind to.
# - port_handler=rootlesskit: Use rootlesskit for port forwarding. Default.
#   Note: Rootlesskit changes the source IP address of incoming packets to a IP address in the container
#   network namespace, usually `10.0.2.100`. If your application requires the real source IP address,
#   e.g. web server logs, use the slirp4netns port handler. The rootlesskit port handler is also used for
#   rootless containers when connected to user-defined networks.
# - port_handler=slirp4netns: Use the slirp4netns port forwarding, it is slower than rootlesskit but
#   preserves the correct source IP address. This port handler cannot be used for user-defined networks.
#
network_cmd_options = [ "--mtu=1500" ]

# Whether to use chroot instead of pivot_root in the runtime
#
#no_pivot_root = false

# Number of locks available for containers, pods, and volumes. Each container,
# pod, and volume consumes 1 lock for as long as it exists.
# If this is changed, a lock renumber must be performed (e.g. with the
# 'podman system renumber' command).
#
#num_locks = 2048

# Set the exit policy of the pod when the last container exits.
#pod_exit_policy = "continue"

# Whether to pull new image before running a container
#
pull_policy = "missing"

# Indicates whether the application should be running in remote mode. This flag modifies the
# --remote option on container engines. Setting the flag to true will default
# `podman --remote=true` for access to the remote Podman service.
#
#remote = false

# Number of times to retry pulling/pushing images in case of failure
#
#retry = 3

# Delay between retries in case pulling/pushing image fails.
# If set, container engines will retry at the set interval,
# otherwise they delay 2 seconds and then exponentially back off.
#
#retry_delay = "2s"

# Default OCI runtime
#
#runtime = "crun"

# List of the OCI runtimes that support --format=json. When json is supported
# engine will use it for reporting nicer errors.
#
#runtime_supports_json = ["crun", "runc", "kata", "runsc", "youki", "krun"]

# List of the OCI runtimes that supports running containers with KVM Separation.
#
#runtime_supports_kvm = ["kata", "krun"]

# List of the OCI runtimes that supports running containers without cgroups.
#
#runtime_supports_nocgroups = ["crun", "krun"]

# Default location for storing temporary container image content. Can be overridden with the TMPDIR environment
# variable. If you specify "storage", then the location of the
# container/storage tmp directory will be used.
# image_copy_tmp_dir="/var/tmp"

# Number of seconds to wait without a connection
# before the `podman system service` times out and exits
#
#service_timeout = 5

# Directory for persistent engine files (database, etc)
# By default, this will be configured relative to where the containers/storage
# stores containers
# Uncomment to change location from this default
#
#static_dir = "/var/lib/containers/storage/libpod"

# Number of seconds to wait for container to exit before sending kill signal.
#
#stop_timeout = 10

# Number of seconds to wait before exit command in API process is given to.
# This mimics Docker's exec cleanup behaviour, where the default is 5 minutes (value is in seconds).
#
#exit_command_delay = 300

# map of service destinations
#
# [engine.service_destinations]
#  [engine.service_destinations.production]
#     URI to access the Podman service
#     Examples:
#       rootless "unix:///run/user/$UID/podman/podman.sock" (Default)
#       rootful "unix:///run/podman/podman.sock (Default)
#       remote rootless ssh://engineering.lab.company.com/run/user/1000/podman/podman.sock
#       remote rootful ssh://root@10.10.1.136:22/run/podman/podman.sock
#       tcp/tls remote tcp://10.10.1.136:9443
#
#    uri = "ssh://user@production.example.com/run/user/1001/podman/podman.sock"
#    Path to file containing ssh identity key
#    identity = "~/.ssh/id_rsa"
#    Path to PEM file containing TLS client certificate
#    tls_cert = "/path/to/certs/podman/tls.crt"
#    Path to PEM file containing TLS client certificate private key
#    tls_key = "/path/to/certs/podman/tls.key"
#    Path to PEM file containing TLS certificate authority (CA) bundle
#    tls_ca = "/path/to/certs/podman/ca.crt"

# Directory for temporary files. Must be tmpfs (wiped after reboot)
#
#tmp_dir = "/run/libpod"

# Directory for libpod named volumes.
# By default, this will be configured relative to where containers/storage
# stores containers.
# Uncomment to change location from this default.
#
#volume_path = "/var/lib/containers/storage/volumes"

# Default timeout (in seconds) for volume plugin operations.
# Plugins are external programs accessed via a REST API; this sets a timeout
# for requests to that API.
# A value of 0 is treated as no timeout.
#volume_plugin_timeout = 5

# Paths to look for a valid OCI runtime (crun, runc, kata, runsc, krun, etc)
[engine.runtimes]
#crun = [
#  "/usr/bin/crun",
#  "/usr/sbin/crun",
#  "/usr/local/bin/crun",
#  "/usr/local/sbin/crun",
#  "/sbin/crun",
#  "/bin/crun",
#  "/run/current-system/sw/bin/crun",
#]

#crun-vm = [
#  "/usr/bin/crun-vm",
#  "/usr/local/bin/crun-vm",
#  "/usr/local/sbin/crun-vm",
#  "/sbin/crun-vm",
#  "/bin/crun-vm",
#  "/run/current-system/sw/bin/crun-vm",
#]

#kata = [
#  "/usr/bin/kata-runtime",
#  "/usr/sbin/kata-runtime",
#  "/usr/local/bin/kata-runtime",
#  "/usr/local/sbin/kata-runtime",
#  "/sbin/kata-runtime",
#  "/bin/kata-runtime",
#  "/usr/bin/kata-qemu",
#  "/usr/bin/kata-fc",
#]

#runc = [
#  "/usr/bin/runc",
#  "/usr/sbin/runc",
#  "/usr/local/bin/runc",
#  "/usr/local/sbin/runc",
#  "/sbin/runc",
#  "/bin/runc",
#  "/usr/lib/cri-o-runc/sbin/runc",
#]

#runsc = [
#  "/usr/bin/runsc",
#  "/usr/sbin/runsc",
#  "/usr/local/bin/runsc",
#  "/usr/local/sbin/runsc",
#  "/bin/runsc",
#  "/sbin/runsc",
#  "/run/current-system/sw/bin/runsc",
#]

#youki = [
#  "/usr/local/bin/youki",
#  "/usr/bin/youki",
#  "/bin/youki",
#  "/run/current-system/sw/bin/youki",
#]

#krun = [
#  "/usr/bin/krun",
#  "/usr/local/bin/krun",
#]

# Default flags for a valid OCI runtime (crun, runc, kata, runsc, krun, etc)
# Note: Do not pass the leading -- to the flag. To pass the runc flag --log-format json, the option given is log-format=json.
[engine.runtimes_flags]
#crun = []

#crun-vm = []

#kata = []

#runc = []

#runsc = []

#youki = []

#krun = []

[engine.volume_plugins]
#testplugin = "/run/podman/plugins/test.sock"

[machine]
# Number of CPU's a machine is created with.
#
#cpus=1

# The size of the disk in GB created when init-ing a podman-machine VM.
#
#disk_size=10

# Default Image used when creating a new VM using `podman machine init`.
# Can be specified as registry with a bootable OCI artifact, download URL, or a local path.
# Registry target must be in the form of `docker://registry/repo/image:version`.
# Container engines translate URIs $OS and $ARCH to the native OS and ARCH.
# URI "https://example.com/$OS/$ARCH/foobar.ami" would become
# "https://example.com/linux/amd64/foobar.ami" on a Linux AMD machine.
# If unspecified, the default Podman machine image will be used.
#
#image = ""

# Memory in MB a machine is created with.
#
#memory=2048

# The username to use and create on the podman machine OS for rootless
# container access.
#
#user = "core"

# Host directories to be mounted as volumes into the VM by default.
# Environment variables like $HOME as well as complete paths are supported for
# the source and destination. An optional third field `:ro` can be used to
# tell the container engines to mount the volume readonly.
#
#volumes = [
#  "$HOME:$HOME",
#]

# Virtualization provider used to run Podman machine.
# If it is empty or commented out, the default provider will be used.
# Linux:
#    qemu    - Open source machine emulator and virtualizer. (Default)
# Windows: there are currently two options:
#    wsl     - Windows Subsystem for Linux (Default)
#    hyperv  - Windows Server Virtualization
# Mac: there are currently two options:
#    applehv - Default Apple Hypervisor (Default)
#    libkrun - Launch virtual machines using the libkrun platform, optimized
#              for sharing GPU with the machine.
#provider = ""

# Rosetta supports running x86_64 Linux binaries on a Podman machine on Apple silicon.
# The default value is `true`. Supported on AppleHV(arm64) machines only.
#
#rosetta=true

# The [machine] table MUST be the last entry in this file.
# (Unless another table is added)
# TOML does not provide a way to end a table other than a further table being
# defined, so every key hereafter will be part of [machine] and not the
# main config.

[farms]
#
# the default farm to use when farming out builds
# default = ""
#
# map of existing farms
#[farms.list]

[podmansh]
# Shell to spawn in container. Default: /bin/sh.
#shell = "/bin/sh"
#
# Name of the container the podmansh user should join.
#container = "podmansh"
#
# Default timeout in seconds for podmansh logins.
# Favored over the deprecated "podmansh_timeout" field.
#timeout = 30

$ tree ~/.config/containers/
/home/markus/.config/containers/
├── auth.json
├── networks
├── podman
│   └── machine

[tgeruzov]

I think it's just the option format.

network_cmd_options isn't a raw list of slirp4netns CLI arguments. Podman's config expects key=value there.

So instead of:

network_cmd_options = ["--mtu=1500"]

try:

network_cmd_options = ["mtu=1500"]

with:

[network]
default_rootless_network_cmd = "slirp4netns"

The containers.conf docs actually list mtu=MTU as the supported form, with 65520 as Podman's default.

That also explains the error: Podman is treating --mtu=1500 as an unknown config option rather than passing it through unchanged to slirp4netns.

[upmcplanetracker]

for my 'piaware' container I have mtu set in the container file with this line:

ExecStartPost=/usr/bin/podman exec piaware ip link set dev eth0 mtu 1500

for immich I have it set this way, in my immich.network file:

[Network]
NetworkName=immich.network
IPv6=true
Internal=false
Options=mtu=1420
Subnet=10.89.31.0/24
Gateway=10.89.31.1
Subnet=fd89:9000:a1b2:0031::/64
Gateway=fd89:9000:a1b2:0031::1

and when I podman network inspect immich.network I can see it passed through to the network itself-

     {
          "name": "immich.network",
          "id": "657a003ef17f31743c24e23fafe1165f74f7d14cebcf88ee1926d693d84b74b1",
          "driver": "bridge",
          "network_interface": "podman1",
          "created": "2026-08-06T01:11:29.866435604Z",
          "subnets": [
               {
                    "subnet": "10.89.31.0/24",
                    "gateway": "10.89.31.1"
               },
               {
                    "subnet": "fd89:9000:a1b2:31::/64",
                    "gateway": "fd89:9000:a1b2:31::1"
               }
          ],
          "ipv6_enabled": true,
          "internal": false,
          "dns_enabled": true,
          "options": {
               "mtu": "1420"
          },
          "ipam_options": {
               "driver": "host-local"
          },
          "containers": {

[psehgaft]

---

The 65520 isn't coming from slirp4netns — its own default is 1500. That 65520 is podman: when podman launches slirp4netns for the default rootless network it sets the MTU to 65520 itself (it's documented in containers.conf: mtu=MTU: Specify the MTU to use for this network. (Default is 65520)). It's a deliberate performance choice — fewer, bigger frames inside the user-mode stack — but if anything on your real path has a smaller MTU (VPN, tunnel, PPPoE) and PMTUD can't get through, large packets get silently blackholed. That's the classic symptom behind the connection drops you're seeing, and why forcing 1500 fixes them.

Your network_cmd_options attempt fails because those aren't CLI flags — they're key=value pairs. The valid values (from the containers.conf reference) are mtu=MTU, cidr=CIDR, enable_ipv6=true|false, port_handler=..., etc. — no dashes. Podman rejects --mtu=1500 before slirp4netns ever sees it, hence "unknown option". The fix:

[engine]
network_cmd_options = ["mtu=1500"]

Two side notes:

• /etc/containers/containers.conf is read for rootless too, so your edit location works — but ~/.config/containers/containers.conf is the cleaner per-user spot (it overrides /etc).
• Alternative: switch the default rootless stack to pasta with [network] default_rootless_network_cmd = "pasta" — it's the default on recent distros and doesn't do the 65520 thing. Or per-network: podman network create -o mtu=1500 mynet.

Verify with your ip link show check — eth0 should report mtu 1500 afterwards.