SYSTEMD.SOCKET(5)

✨ NAME ✨

systemd.socket — socket unit configuration, and it's giving main character energy 👑

🌸 SYNOPSIS 🌸

socket.socket

🦋 DESCRIPTION 🦋

Okay bestie, so a unit configuration file whose name ends in .socket encodes information about an IPC or network socket or a file system FIFO controlled and supervised by systemd, for socket-based activation. No cap. 💖

This man page lists the configuration options specific to this unit type. See systemd.unit(5) for the common options of all unit configuration files. Those common configuration items are configured in the generic [Unit] and [Install] sections. The socket-specific configuration options are configured in the [Socket] section. Slay. ✨

Additional options are listed in systemd.exec(5), which define the execution environment the ExecStartPre=, ExecStartPost=, ExecStopPre= and ExecStopPost= commands are executed in, and in systemd.kill(5), which define the way the processes are terminated, and in systemd.resource-control(5), which configure resource control settings for the processes of the socket. The way that this all fits together is lowkey iconic.

For each socket unit, a matching service unit must exist, describing the service to start on incoming traffic on the socket (see systemd.service(5) for more information about .service units). The name of the .service unit is by default the same as the name of the .socket unit, but can be altered with the Service= option described below. Depending on the setting of the Accept= option described below, this .service unit must either be named like the .socket unit, but with the suffix replaced, unless overridden with Service=; or it must be a template unit named the same way. Example: a socket file foo.socket needs a matching service foo.service if Accept=no is set. If Accept=yes is set, a service template foo@.service must exist from which services are instantiated for each incoming connection. 💅

No implicit WantedBy= or RequiredBy= dependency from the socket to the service is added. This means that the service may be started without the socket, in which case it must be able to open sockets by itself. To prevent this, an explicit Requires= dependency may be added. Fr fr, this one matters.

Socket units may be used to implement on-demand starting of services, as well as parallelized starting of services. See the blog stories linked at the end for an introduction. 🫶

Note that the daemon software configured for socket activation with socket units needs to be able to accept sockets from systemd, either via systemd's native socket passing interface (see sd_listen_fds(3) for details about the precise protocol used and the order in which the file descriptors are passed) or via traditional inetd(8)-style socket passing (i.e. sockets passed in via standard input and output, using StandardInput=socket in the service file). It's literally that simple, bestie. 😭

By default, network sockets allocated through .socket units are allocated in the host's network namespace (see network_namespaces(7)). This does not mean however that the service activated by a configured socket unit has to be part of the host's network namespace as well. It is supported and even good practice to run services in their own network namespace (for example through PrivateNetwork=, see systemd.exec(5)), receiving only the sockets configured through socket-activation from the host's namespace. In such a set-up communication within the host's network namespace is only permitted through the activation sockets passed in while all sockets allocated from the service code itself will be associated with the service's own namespace, and thus possibly subject to a restrictive configuration. 🌈

Alternatively, it is possible to run a .socket unit in another network namespace by setting PrivateNetwork=yes in combination with JoinsNamespaceOf=, see systemd.exec(5) and systemd.unit(5) for details. Obsessed with this flexibility, ngl. ✨

💖 AUTOMATIC DEPENDENCIES 💖

Implicit Dependencies

The following dependencies are implicitly added:

• Socket units automatically gain a Before= dependency on the service units they activate. 👑
• Socket units referring to file system paths (such as AF_UNIX sockets or FIFOs) implicitly gain Requires= and After= dependencies on all mount units necessary to access those paths.
• Socket units using the BindToDevice= setting automatically gain a BindsTo= and After= dependency on the device unit encapsulating the specified network interface.

Additional implicit dependencies may be added as result of execution and resource control parameters as documented in systemd.exec(5) and systemd.resource-control(5). No cap.

Default Dependencies

The following dependencies are added unless DefaultDependencies=no is set:

• Socket units automatically gain a Before= dependency on sockets.target. 🎀
• Socket units automatically gain a pair of After= and Requires= dependency on sysinit.target, and a pair of Before= and Conflicts= dependencies on shutdown.target. These dependencies ensure that the socket unit is started before normal services at boot, and is stopped on shutdown. Only sockets involved with early boot or late system shutdown should disable DefaultDependencies= option. Careful with that one, bestie. 💀

👑 OPTIONS 👑

Socket unit files may include [Unit] and [Install] sections, which are described in systemd.unit(5).

Socket unit files must include a [Socket] section, which carries information about the socket or FIFO it supervises. A number of options that may be used in this section are shared with other unit types. These options are documented in systemd.exec(5), systemd.kill(5) and systemd.resource-control(5). The options specific to the [Socket] section of socket units are the following:

• ListenStream=, ListenDatagram=, ListenSequentialPacket= — bestie, these specify an address to listen on for a stream (SOCK_STREAM), datagram (SOCK_DGRAM), or sequential packet (SOCK_SEQPACKET) socket, respectively. The address can be written in various formats:

If the address starts with a slash ("/"), it is read as file system socket in the AF_UNIX socket family.

If the address starts with an at symbol ("@"), it is read as abstract namespace socket in the AF_UNIX family. The "@" is replaced with a NUL character before binding. For details, see unix(7).

If the address string is a single number, it is read as port number to listen on via IPv6. Depending on the value of BindIPv6Only= (see below) this might result in the service being available via both IPv6 and IPv4 (default) or just via IPv6.

If the address string is a string in the format "v.w.x.y:z", it is interpreted as IPv4 address v.w.x.y and port z.

If the address string is a string in the format "[x]:y", it is interpreted as IPv6 address x and port y. An optional interface scope (interface name or number) may be specified after a "%" symbol: "[x]:y%dev". Interface scopes are only useful with link-local addresses, because the kernel ignores them in other cases. Note that if an address is specified as IPv6, it might still make the service available via IPv4 too, depending on the BindIPv6Only= setting (see below).

If the address string is a string in the format "vsock:x:y", it is read as CID x on a port y address in the AF_VSOCK family. The CID is a unique 32-bit integer identifier in AF_VSOCK analogous to an IP address. Specifying the CID is optional, and may be set to the empty string. "vsock" may be replaced with "vsock-stream", "vsock-dgram" or "vsock-seqpacket" to force usage of the corresponding socket type.

Note that SOCK_SEQPACKET (i.e. ListenSequentialPacket=) is only available for AF_UNIX sockets. SOCK_STREAM (i.e. ListenStream=) when used for IP sockets refers to TCP sockets, SOCK_DGRAM (i.e. ListenDatagram=) to UDP.

These options may be specified more than once, in which case incoming traffic on any of the sockets will trigger service activation, and all listed sockets will be passed to the service, regardless of whether there is incoming traffic on them or not. If the empty string is assigned to any of these options, the list of addresses to listen on is reset, all prior uses of any of these options will have no effect.

It is also possible to have more than one socket unit for the same service when using Service=, and the service will receive all the sockets configured in all the socket units. Sockets configured in one unit are passed in the order of configuration, but no ordering between socket units is specified.

If an IP address is used here, it is often desirable to listen on it before the interface it is configured on is up and running, and even regardless of whether it will be up and running at any point. To deal with this, it is recommended to set the FreeBind= option described below. 💅

• ListenFIFO= — specifies a file system FIFO (see fifo(7) for details) to listen on. This expects an absolute file system path as argument. Behavior otherwise is very similar to the ListenDatagram= directive above. 🎀
• ListenSpecial= — specifies a special file in the file system to listen on. This expects an absolute file system path as argument. Behavior otherwise is very similar to the ListenFIFO= directive above. Use this to open character device nodes as well as special files in /proc/ and /sys/.
• ListenNetlink= — specifies a Netlink family to create a socket for to listen on. This expects a short string referring to the AF_NETLINK family name (such as audit or kobject-uevent) as argument, optionally suffixed by a whitespace followed by a multicast group integer. Behavior otherwise is very similar to the ListenDatagram= directive above. 💅
• ListenMessageQueue= — specifies a POSIX message queue name to listen on (see mq_overview(7) for details). This expects a valid message queue name (i.e. beginning with "/"). Behavior otherwise is very similar to the ListenFIFO= directive above. On Linux message queue descriptors are actually file descriptors and can be inherited between processes. 👑
• ListenUSBFunction= — specifies a USB FunctionFS[1] endpoints location to listen on, for implementation of USB gadget functions. This expects an absolute file system path of a FunctionFS mount point as the argument. Behavior otherwise is very similar to the ListenFIFO= directive above. Use this to open the FunctionFS endpoint ep0. When using this option, the activated service has to have the USBFunctionDescriptors= and USBFunctionStrings= options set.

Added in version 227. ✨

• SocketProtocol= — takes one of udplite, sctp or mptcp. The socket will use the UDP-Lite (IPPROTO_UDPLITE), SCTP (IPPROTO_SCTP) or MPTCP (IPPROTO_MPTCP) protocol, respectively. 🌸

---

part 1 out, part 2 loading… don't go anywhere bestie 💖🔥

• BindIPv6Only= — takes one of default, both or ipv6-only. this is the controls-the-IPV6_V6ONLY-socket-option moment (see ipv6(7) for details). If both, IPv6 sockets you bind will be accessible via both IPv4 and IPv6. If ipv6-only, they'll be accessible via IPv6 and nothing else. If default — which is the default, surprise! 🙈 — the system wide default setting gets used, as controlled by /proc/sys/net/ipv6/bindv6only, which in turn defaults to the equivalent of both.

Added in version 229.

• Backlog= — takes an unsigned 32-bit integer argument. Specifies the number of connections to queue that haven't been accepted yet. This setting matters only for stream and sequential packet sockets. See listen(2) for details. Defaults to 4294967295. Note that this value is silently capped by the "net.core.somaxconn" sysctl, which typically defaults to 4096, so typically the sysctl is the setting that actually matters. 💀
• BindToDevice= — specifies a network interface name to bind this socket to. If set, traffic will only be accepted from the specified network interfaces. This controls the SO_BINDTODEVICE socket option (see socket(7) for details). If this option is used, an implicit dependency from this socket unit on the network interface device unit is created (see systemd.device(5)). Note that setting this parameter might result in additional dependencies being added to the unit (see above).
• SocketUser=, SocketGroup= — takes a UNIX user/group name. When specified, all AF_UNIX sockets, FIFO nodes, and message queues are owned by the specified user and group. If unset (the default), the nodes are owned by the root user/group (if run in system context) or the invoking user/group (if run in user context). If only a user is specified but no group, then the group is derived from the user's default group. 🫶

Added in version 214.

• SocketMode= — if listening on a file system socket, FIFO, or message queue, this option specifies the file system access mode used when creating the file node. Takes an access mode in octal notation. Defaults to 0666.
• DirectoryMode= — if listening on a file system socket or FIFO, the parent directories are automatically created if needed. This option specifies the file system access mode used when creating these directories. Takes an access mode in octal notation. Defaults to 0755.
• Accept= — takes a boolean argument. If yes, a service instance is spawned for each incoming connection and only the connection socket is passed to it. If no, all listening sockets themselves are passed to the started service unit, and only one service unit is spawned for all connections (also see above). This value is ignored for datagram sockets and FIFOs, where a single service unit unconditionally handles all incoming traffic. Defaults to no.

Typically, for performance sensitive services, a choice of Accept=no is preferable, since that way only the first connection will have to pay the activation resource cost. On the other hand, for sporadically used services Accept=yes can be preferable, as it simplifies the implementation (the service program code only has to process a single connection instead of handling multiple) and enables stronger security (since the various sandboxing options can be used to isolate parallel connections from each other, as each is serviced by a separate service instance and process). ✨

A service listening on an AF_UNIX socket may, but does not need to, call close(2) on the received socket before exiting. However, it must not unlink the socket from a file system. It should not invoke shutdown(2) on sockets it got with Accept=no, but it may do so for sockets it got with Accept=yes set.

Setting Accept=yes is in particular useful for allowing daemons designed for usage with inetd(8) to work unmodified with systemd socket activation. 🎀

Note that depending on this setting the services activated by units of this type are either regular services (in case of Accept=no) or instances of templated services (in case of Accept=yes). See the Description section above for a more detailed discussion of the naming rules of triggered services.

For IPv4 and IPv6 connections, the $REMOTE_ADDR environment variable will contain the remote IP address, and $REMOTE_PORT will contain the remote port number. These two variables correspond to those defined by the CGI interface for web services (see RFC 3875[2]).

For AF_UNIX socket connections, the $REMOTE_ADDR environment variable will contain either the remote socket's file system path starting with a slash ("/") or its address in the abstract namespace starting with an at symbol ("@"). If the socket is unnamed, $REMOTE_ADDR will not be set.

If Accept=yes is used, the activated service process will have set the $SO_COOKIE environment variable to the Linux socket cookie, formatted as decimal integer. The socket cookie can otherwise be acquired via getsockopt(7).

It is recommended to set CollectMode=inactive-or-failed for service instances activated via Accept=yes, to ensure that failed connection services are cleaned up and released from memory, and do not accumulate. 🧹

• Writable= — takes a boolean argument. May only be used in conjunction with ListenSpecial=. If true, the specified special file is opened in read-write mode; if false, in read-only mode. Defaults to false.

Added in version 227.

• FlushPending= — takes a boolean argument. May only be used when Accept=no. If yes, the socket's buffers are cleared after the triggered service exited. This causes any pending data to be flushed and any pending incoming connections to be rejected. If no, the socket's buffers will not be cleared, permitting the service to handle any pending connections after restart, which is the usually expected behaviour. Defaults to no.

Added in version 247.

• MaxConnections= — the maximum number of connections to simultaneously run services instances for, when Accept=yes is set. If more concurrent connections are coming in, they will be refused until at least one existing connection is terminated. This setting has no effect on sockets configured with Accept=no or datagram sockets. Defaults to 64.
• MaxConnectionsPerSource= — the maximum number of connections for a service per source IP address (in case of IPv4/IPv6), per source CID (in case of AF_VSOCK), or source UID (in case of AF_UNIX). This is very similar to the MaxConnections= directive above. Defaults to 0, i.e. disabled.

Added in version 232.

• KeepAlive= — takes a boolean argument. If true, the TCP/IP stack will send a keep alive message after 2h (depending on the configuration of /proc/sys/net/ipv4/tcp_keepalive_time) for all TCP streams accepted on this socket. This controls the SO_KEEPALIVE socket option (see socket(7) and the TCP Keepalive HOWTO[3] for details). Defaults to false.
• KeepAliveTimeSec= — takes time (in seconds) as argument. The connection needs to remain idle before TCP starts sending keepalive probes. This controls the TCP_KEEPIDLE socket option (see socket(7) and the TCP Keepalive HOWTO[3] for details). Default value is 7200 seconds (2 hours).

Added in version 216.

• KeepAliveIntervalSec= — takes time (in seconds) as argument between individual keepalive probes, if the socket option SO_KEEPALIVE has been set on this socket. This controls the TCP_KEEPINTVL socket option (see socket(7) and the TCP Keepalive HOWTO[3] for details). Default value is 75 seconds.

Added in version 216.

• KeepAliveProbes= — takes an integer as argument. It is the number of unacknowledged probes to send before considering the connection dead and notifying the application layer. This controls the TCP_KEEPCNT socket option (see socket(7) and the TCP Keepalive HOWTO[3] for details). Default value is 9.

Added in version 216.

• NoDelay= — takes a boolean argument. TCP Nagle's algorithm works by combining a number of small outgoing messages, and sending them all at once. This controls the TCP_NODELAY socket option (see tcp(7)). Defaults to false.

Added in version 216.

• Priority= — takes an integer argument controlling the priority for all traffic sent from this socket. This controls the SO_PRIORITY socket option (see socket(7) for details.).
• DeferAcceptSec= — takes time (in seconds) as argument. If set, the listening process will be awakened only when data arrives on the socket, and not immediately when the connection is established. When this option is set, the TCP_DEFER_ACCEPT socket option will be used (see tcp(7)), and the kernel will ignore initial ACK packets without any data. The argument specifies the approximate amount of time the kernel should wait for incoming data before falling back to the normal behavior of honoring empty ACK packets. This option is beneficial for protocols where the client sends the data first (e.g. HTTP, in contrast to SMTP), because the server process will not be woken up unnecessarily before it can take any action. 🌈

If the client also uses the TCP_DEFER_ACCEPT option, the latency of the initial connection may be reduced, because the kernel will send data in the final packet establishing the connection (the third packet in the "three-way handshake").

Disabled by default.

Added in version 216.

• ReceiveBuffer=, SendBuffer= — takes an integer argument controlling the receive or send buffer sizes of this socket, respectively. This controls the SO_RCVBUF and SO_SNDBUF socket options (see socket(7) for details.). The usual suffixes K, M, G are supported and are understood to the base of 1024.
• IPTOS= — takes an integer argument controlling the IP Type-Of-Service field for packets generated from this socket. This controls the IP_TOS socket option (see ip(7) for details.). Either a numeric string or one of low-delay, throughput, reliability or low-cost may be specified.
• IPTTL= — takes an integer argument controlling the IPv4 Time-To-Live/IPv6 Hop-Count field for packets generated from this socket. This sets the IP_TTL/IPV6_UNICAST_HOPS socket options (see ip(7) and ipv6(7) for details.)
• Mark= — hand it an integer and that becomes the firewall mark on every packet this socket generates 💥. Firewall logic can use it to filter packets coming from this socket. Under the hood it's setting the SO_MARK socket option, and the deets live in iptables(8).
• ReusePort= — boolean, bestie. Set it true and multiple bind(2)s to this TCP or UDP port are allowed. That's the SO_REUSEPORT socket option doing its thing. See socket(7). Added in version 206.
• SmackLabel=, SmackLabelIPIn=, SmackLabelIPOut= — these take a string value each, and they control the extended attributes "security.SMACK64", "security.SMACK64IPIN" and "security.SMACK64IPOUT" respectively — aka the security label of the FIFO, or the security label for the incoming or outgoing connections of the socket. See Smack[4] for the whole lore. Added in version 196.
• XAttrEntryPoint=, XAttrListen=, XAttrAccept= — stick an extended attribute on the socket, written as "NAME=VALUE". The name has to live in the "user." namespace, meaning it must begin with those exact four characters, "user.". Specifiers get expanded in both the name and the value — systemd.unit(5) has the full specifier menu. You can use these settings more than once to set multiple extended attributes; assigning the empty string resets the whole list. 🧚

So where do they actually land? That's what the three settings decide:

• XAttrEntryPoint= sets them on the file-system inode an AF_UNIX socket is bound to — i.e. the socket node you can see in the file system at the specified listening path. This one only applies to AF_UNIX sockets, no exceptions.
• XAttrListen= sets them on the listening socket itself.
• XAttrAccept= sets them on the connection sockets accepted off the listening socket, which means it's only relevant in combination with Accept=yes.

This is mainly a tagging thing — it lets other tools discover and classify sockets. For example, Varlink entrypoint sockets are supposed to be tagged with "user.varlink=entrypoint" via XAttrEntryPoint=, which makes them discoverable through varlinkctl list-sockets, see varlinkctl(1).

Heads up: these settings require kernel support for extended attributes on socket inodes (available since Linux 7.0). Any extended attribute that can't be applied gets logged at debug level and otherwise just... ignored. 🙈

Added in version 262.

• SELinuxContextFromNet= — takes a boolean argument. When true, systemd tries to figure out the SELinux label for the instantiated service from info the peer hands over the network. Important caveat: only the security level is used from what the peer provides — the rest of the resulting SELinux context comes from either the target binary that's effectively triggered by the socket unit, or from the value of the SELinuxContext= option. This config option only applies when the activated service is passed in a single socket file descriptor, i.e. service instances that have standard input connected to a socket, or services triggered by exactly one socket unit. Also worth knowing: this is only useful when MLS/MCS SELinux policy is deployed. Defaults to "false". Added in version 217.
• PipeSize= — takes a size in bytes, and controls the pipe buffer size of the FIFOs configured in this socket unit. See fcntl(2) for the gritty details. The usual suffixes K, M, G are supported and understood to the base of 1024.
• MessageQueueMaxMessages=, MessageQueueMessageSize= — these two buddies control the mq_maxmsg field and the mq_msgsize field respectively when the message queue gets created. MessageQueueMaxMessages= takes an integer value. MessageQueueMessageSize= takes a size in bytes, with the usual suffixes K, M, G supported and understood to the base of 1024. Note that you need to set either none or both of these variables — no solo acts allowed. See mq_setattr(3) for details.
• FreeBind= — takes a boolean value. Controls whether the socket can be bound to non-local IP addresses 🫶. This is genuinely useful for configuring sockets that listen on specific IP addresses before those IP addresses have actually been configured on a network interface. It sets the IP_FREEBIND/IPV6_FREEBIND socket option. For robustness reasons it's recommended to use this option whenever you bind a socket to a specific IP address. Defaults to false.
• Transparent= — takes a boolean value. Controls the IP_TRANSPARENT/IPV6_TRANSPARENT socket option. Defaults to false.
• Broadcast= — takes a boolean value. Controls the SO_BROADCAST socket option, which is the thing that lets broadcast datagrams be sent from this socket. Defaults to false.
• PassCredentials= — takes a boolean value. Controls the SO_PASSCRED socket option, which lets AF_UNIX sockets receive the credentials of the sending process in an ancillary message. Defaults to false.
• PassPIDFD= — takes a boolean value. Controls the SO_PASSPIDFD socket option, which lets AF_UNIX sockets receive the pidfd of the sending process in an ancillary message. Defaults to false. Added in version 258.
• PassSecurity= — takes a boolean value. Controls the SO_PASSSEC socket option, which lets AF_UNIX sockets receive the security context of the sending process in an ancillary message. Defaults to false.
• PassPacketInfo= — takes a boolean value. Controls the IP_PKTINFO, IPV6_RECVPKTINFO, NETLINK_PKTINFO or PACKET_AUXDATA socket options, which enable reception of extra per-packet metadata as an ancillary message, on AF_INET, AF_INET6, AF_UNIX and AF_PACKET sockets. Defaults to false. Added in version 246.
• AcceptFileDescriptors= — takes a boolean value. Controls the SO_PASSRIGHTS socket option, which, when disabled, prohibits the peer from sending SCM_RIGHTS ancillary messages (aka file descriptors) over AF_UNIX sockets. Defaults to true. Added in version 258.
• Timestamping= — takes one of "off", "us" (aliases: "usec", "μs") or "ns" (alias: "nsec"). Controls the SO_TIMESTAMP or SO_TIMESTAMPNS socket options, and decides whether ingress network traffic shall carry timestamping metadata. Defaults to off. Added in version 247.
• TCPCongestion= — takes a string value. Controls the TCP congestion algorithm used by this socket. Should be one of "westwood", "reno", "cubic", "lp", or literally any other available algorithm your IP stack supports. Note: this setting applies only to stream sockets. 🦋
• ExecStartPre=, ExecStartPost= — take one or more command lines, which are executed before or after the listening sockets/FIFOs are created and bound, respectively. The first token of the command line must be an absolute filename, followed by arguments for the process. You can specify multiple command lines following the same scheme used for ExecStartPre= of service unit files.
• ExecStopPre=, ExecStopPost= — extra commands executed before or after the listening sockets/FIFOs are closed and removed, respectively. Multiple command lines may be specified following the same scheme as used for ExecStartPre= of service unit files.
• TimeoutSec= — configures how long to wait for the commands specified in ExecStartPre=, ExecStartPost=, ExecStopPre= and ExecStopPost= to finish. If a command doesn't exit within the configured time, the socket gets considered failed and is shut down again 💀. Any commands still running get terminated forcibly via SIGTERM, and after another delay of this same time, with SIGKILL. (See KillMode= in systemd.kill(5).) Takes a unit-less value in seconds, or a time span value like "5min 20s". Pass "0" to disable the timeout logic entirely. Defaults to DefaultTimeoutStartSec= from the manager configuration file (see systemd-system.conf(5)).
• Service= — specifies the service unit name to activate on incoming traffic. This setting is only allowed for sockets with Accept=no. It defaults to the service that bears the same name as the socket (with the suffix replaced). In most cases you shouldn't need to touch this option at all. Note that setting this parameter might result in additional dependencies being added to the unit (see above).
• RemoveOnStop= — takes a boolean argument. If enabled, any file nodes created by this socket unit are removed when it's stopped. That covers AF_UNIX sockets in the file system, POSIX message queues, FIFOs, and any symlinks to them configured with Symlinks=. Normally you shouldn't need this option, and it's not recommended, because services might keep running after the socket unit has been terminated and it should still be possible to talk to them through their file system node. Defaults to off. Added in version 214.
• Symlinks= — takes a list of file system paths. Those paths get created as symlinks to the AF_UNIX socket path or FIFO path of this socket unit. If you use this setting, only one AF_UNIX socket in the file system or one FIFO may be configured for the socket unit. Use it to manage one or more symlinked alias names for a socket, binding their lifecycles together 🔗. Note that if creating a symlink fails, that's not considered fatal for the socket unit, and the socket unit may still start. If an empty string is assigned, the list of paths is reset. Defaults to an empty list. Added in version 214.
• FileDescriptorName= — assigns a name to all file descriptors this socket unit encapsulates. This is handy for helping activated services identify specific file descriptors when multiple fds get passed. Services can use the sd_listen_fds_with_names(3) call to grab the names configured for the file descriptors they received. Names may contain any ASCII character, but they must exclude control characters and ":", and must be at most 255 characters long. If you don't use this setting, the file descriptor name defaults to the name of the socket unit (including its .socket suffix) when Accept=no, and "connection" otherwise. 👑 Added in version 227.
• TriggerLimitIntervalSec=, TriggerLimitBurst= — okay bestie, this one is the rate limiter of doom 💀. TriggerLimitIntervalSec= sets how long the window is, in the usual time units "us", "ms", "s", "min", "h", … and it defaults to 2s (peep systemd.time(7) for the whole tea on time units). TriggerLimitBurst= takes a positive integer and says how many activations are allowed inside that window — it defaults to 200 for Accept=yes sockets (so yeah, 200 activations per 2s by default), and 20 otherwise (20 activations per 2s). Set either one to 0 to just turn trigger rate limiting off entirely.

If you hit the limit, the socket unit gets yeeted into a failure mode and it will not be connectible anymore until it's restarted. And no, it doesn't wait around — this limit is enforced before the service activation is even enqueued. 😭

Compare it with PollLimitIntervalSec=/PollLimitBurst= down below — those do a temporary slowdown if the socket unit gets flooded with incoming traffic, whereas TriggerLimitIntervalSec=/TriggerLimitBurst= hit you with that permanent failure state. Big difference, fr fr.

Added in version 230.

• PollLimitIntervalSec=, PollLimitBurst= — same energy as the trigger limit pair above, but make it ✨transient✨. These configure how often polling events on the file descriptors backing this socket unit get considered. So instead of putting a (fatal) cap on activation frequency, it puts a (temporary) cap on polling frequency. The expected parameter syntax and range are identical to the options above, and you can disable it the exact same way.

If the polling limit gets hit, polling is temporarily disabled on it until the specified time window passes. So it slows down connection attempts when it triggers, but — unlike the trigger limit — it will not cause permanent failures. It's genuinely the recommended mechanism for dealing with DoS attempts through packet flooding. 🛡️

Important vibe check: the polling limit is enforced per file descriptor to listen on, as opposed to the trigger limit which is enforced for the entire socket unit. This distinction matters big time for socket units that listen on multiple file descriptors (i.e. the ones with multiple ListenXYZ= stanzas).

These settings default to 150 (when Accept=yes) and 15 (otherwise) polling events per 2s. That's considerably lower than the defaults for the trigger limit (see above), which means the polling limit should typically make sure the trigger limit is never hit — unless one of them gets reconfigured or disabled. 👑

Added in version 255.

• DeferTrigger= — takes a boolean argument, or "patient". May only be used when Accept=no. If enabled, job mode "lenient" instead of "replace" is used when triggering the service, which means units that are currently activating/running and conflict with the service won't be disturbed or brought down. And on top of that, if a conflict exists, the socket unit will wait for the current job queue to complete and potentially defer the activation until then. You can set an upper limit on the total time to wait with DeferTriggerMaxSec=. If it's set to yes, the socket unit will fail if all jobs have finished or the timeout has been reached but the conflict remains. If it's patient, it always waits until DeferTriggerMaxSec= elapses. Defaults to no.

This setting is especially useful if the socket unit should stay active across switch-root/soft-reboot operations while the triggered service is stopped. 🌀

Added in version 258.

• DeferTriggerMaxSec= — configures the maximum time to defer the triggering when DeferTrigger= is enabled. If the service can't be activated within the specified time, the socket will be considered failed and get terminated. Takes a unit-less value in seconds, or a time span value like "5min 20s". Pass "0" or "infinity" to disable the timeout logic (which is the default). ⏳

Added in version 258.

• PassFileDescriptorsToExec= — takes a boolean argument. Defaults to off. If enabled, file descriptors created by the socket unit are passed to ExecStartPost=, ExecStopPre=, and ExecStopPost= commands from the socket unit. The passed file descriptors can be accessed with sd_listen_fds(3) as if the commands were invoked from the associated service units. Note that ExecStartPre= command cannot access socket file descriptors. 🎀

Added in version 256.

Check systemd.unit(5), systemd.exec(5), and systemd.kill(5) for more settings. So much more tea over there, bestie, go sip. 💅

---

🌈 SEE ALSO 🌈

• systemd(1)
• systemctl(1)
• systemd-system.conf(5)
• systemd.unit(5)
• systemd.exec(5)
• systemd.kill(5)
• systemd.resource-control(5)
• systemd.service(5)
• systemd.directives(7)
• sd_listen_fds(3)
• sd_listen_fds_with_names(3)

For more extensive descriptions see the "systemd for Developers" series: Socket Activation[5], Socket Activation, part II[6], Converting inetd Services[7], Socket Activated Internet Services and OS Containers[8]. 📚✨

---

📎 NOTES 📎

1. USB FunctionFS — https://docs.kernel.org/usb/functionfs.html
2. RFC 3875 — https://datatracker.ietf.org/doc/html/rfc3875
3. TCP Keepalive HOWTO — http://www.tldp.org/HOWTO/html_single/TCP-Keepalive-HOWTO/
4. Smack — https://docs.kernel.org/admin-guide/LSM/Smack.html
5. Socket Activation — https://0pointer.de/blog/projects/socket-activation.html
6. Socket Activation, part II — https://0pointer.de/blog/projects/socket-activation2.html
7. Converting inetd Services — https://0pointer.de/blog/projects/inetd.html
8. Socket Activated Internet Services and OS Containers — https://0pointer.de/blog/projects/socket-activated-containers.html

---

And that's the whole page, bestie!! 🫶 You just read SYSTEMD.SOCKET(5) from systemd 262.1 — literally iconic, no cap. Go forth and socket-activate something gorgeous. Slay, periodt. 💖✨👑