ranch_tcp
The ranch_tcp
module implements a TCP Ranch transport.
Note that due to bugs in OTP up to at least R16B02, it is
recommended to disable async threads when using the
sendfile
function of this transport, as it can make
the threads stuck indefinitely.
Types
opt() = {backlog, non_neg_integer()}
| {buffer, non_neg_integer()}
| {delay_send, boolean()}
| {dontroute, boolean()}
| {exit_on_close, boolean()}
| {fd, non_neg_integer()}
| {high_msgq_watermark, non_neg_integer()}
| {high_watermark, non_neg_integer()}
| inet
| inet6
| {ip, inet:ip_address()}
| {keepalive, boolean()}
| {linger, {boolean(), non_neg_integer()}}
| {low_msgq_watermark, non_neg_integer()}
| {low_watermark, non_neg_integer()}
| {nodelay, boolean()}
| {port, inet:port_number()}
| {priority, integer()}
| {raw, non_neg_integer(), non_neg_integer(), binary()}
| {recbuf, non_neg_integer()}
| {send_timeout, timeout()}
| {send_timeout_close, boolean()}
| {sndbuf, non_neg_integer()}
| {tos, integer()}
Listen options.
This does not represent the entirety of the options that can
be set on the socket, but only the options that may be
set independently of protocol implementation.
opts() = [opt()]
List of listen options.
Option descriptions
None of the options are required.
Please consult the gen_tcp
and inet
manuals for a more
thorough description of these options. This manual only aims
to provide a short description along with what the defaults
are. Defaults may be different in Ranch compared to gen_tcp
.
Defaults are given next to the option name.
- backlog (1024)
- Max length of the queue of pending connections.
- buffer
- Size of the buffer used by the Erlang driver. Default is system-dependent.
- delay_send (false)
- Always queue packets before sending, to send fewer, larger packets over the network.
- dontroute (false)
- Don't send via a gateway, only send to directly connected hosts.
- exit_on_close (true)
- Disable to allow sending data after a close has been detected.
- fd
- File descriptor of the socket, if it was opened externally.
- high_msgq_watermark (8192)
- Limit in the amount of data in the socket message queue before the socket queue becomes busy.
- high_watermark (8192)
- Limit in the amount of data in the ERTS socket implementation's queue before the socket becomes busy.
- inet
- Set up the socket for IPv4.
- inet6
- Set up the socket for IPv6.
- ip
- Interface to listen on. Listen on all interfaces by default.
- keepalive (false)
- Enable sending of keep-alive messages.
- linger ({false, 0})
- Whether to wait and how long to flush data sent before closing the socket.
- low_msgq_watermark (4096)
- Amount of data in the socket message queue before the socket queue leaves busy state.
- low_watermark (4096)
- Amount of data in the ERTS socket implementation's queue before the socket leaves busy state.
- nodelay (true)
- Whether to enable TCP_NODELAY.
- port (0)
- TCP port number to listen on. 0 means a random port will be used.
- priority (0)
- Priority value for all packets to be sent by this socket.
- recbuf
- Minimum size of the socket's receive buffer. Default is system-dependent.
- send_timeout (30000)
- How long the send call may wait for confirmation before returning.
- send_timeout_close (true)
- Whether to close the socket when the confirmation wasn't received.
- sndbuf
- Minimum size of the socket's send buffer. Default is system-dependent.
- tos
- Value for the IP_TOS IP level option. Use with caution.
In addition, the raw
option can be used to set system-specific
options by specifying the protocol level, the option number and
the actual option value specified as a binary. This option is not
portable. Use with caution.
Exports
None.