Previous topic |
Next topic |
Contents |
Contact z/OS |
Library |
PDF
TCP_KeepAlive socket option z/OS Communications Server: IP Programmer's Guide and Reference SC27-3659-02 |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Some TCP⁄IP users require a keep alive function with better timing granularity (in seconds) than that provided by the existing SO_KeepAlive socket option, which uses a stack-wide time value provided by configuration data. The Posix.1g standard defines an alternative keep alive function, TCP_KeepAlive, which provides a value in seconds that is specific to a particular socket. The value of TCP_KeepAlive, which is used for the current
connection in place of the configuration default keep alive time (when
keep alive timing is made active by the SO_KeepAlive socket option),
can be in the range 1 – 2 147 460 seconds. If a value greater than
2 147 460 is specified, 2 147 460 is used. If the TCP_KeepAlive value
0 is specified for a specific socket, keep alive timing for that socket
is disabled.
Unlike the algorithm that is used to issue probes during an SO_KeepAlive cycle, the TCP_KeepAlive function varies the number of probes that are issued before terminating the connection. Probe retry intervals are scaled in proportion to the interval specified, as shown in Table 1.
The TCP_KeepAlive option value can range from 1 – 2 147 460 seconds. For values greater than 2 hours (7200 seconds), the probe interval and number of probes are adjusted as the specified interval increases until they coincide with the default algorithm. If no response is received from the remote partner after the listed number of probes, the connection is terminated. Tips:
|
Copyright IBM Corporation 1990, 2014
|