[OpenSIPS-Users] tcp_max_msg_time question
liviu at opensips.org
Wed Dec 2 11:15:21 CET 2015
Do not confuse TCP connection _lifetime & keepalives_ with _fragmented
SIP messages over TCP_. There is no connection between them.
Lifetime & keepalives
Here you have /tcp_connection_lifetime/  and /tcp_keepalive/ .
Pinging is done at kernel level, with 4-byte pings which look like
"\0\0\0\0". If these TCP pings do not reach your UACs (e.g. due to some
firewall), the nathelper module may perform its own OPTIONS pinging at
SIP level if you enable /natping_tcp/  (a bit slower, but gets the
job done, as you may have already realised!)
By default, OpenSIPS has built-in resilience against fragmented SIP
messages over TCP. If the TCP core reads only a chunk of a SIP message,
the connection will be closed unless the whole message is delivered
within /tcp_max_msg_time/  seconds since the first chunk, in order to
On 01.12.2015 12:56, Чалков Артём wrote:
> I use opensips v2.1.1 and cant clearly understand how to use
> tcp_max_msg_time better with high amount of sip-users behind NAT. I
> dont want to use natping_tcp, only TCP-keepalives, but opensips ends
> all TCP connections after tcp_max_msg_time number of seconds despite
> the fact that i use TCP-keepalives and i have to use natping_tcp to
> hold ongoing TCP-connections.
> Maybe you can provide me with some recomendations about how to tune
> tcp_max_msg_time value for my tasks?
> Users mailing list
> Users at lists.opensips.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Users