<html>
<head>
<meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<div class="moz-cite-prefix"><tt>George,<br>
<br>
If the TCP conn closes, it is not much you can do from the
server side...you cannot open a TCP conn behind NAT. You have to
wait for the UAC to re-connect again.<br>
<br>
About your manual UDP defrag - I do not think it is a good it -
the UDP stack must to that for you, you cannot simulate it on
top of datagram.<br>
<br>
Regards,<br>
</tt>
<pre class="moz-signature" cols="72">Bogdan-Andrei Iancu
OpenSIPS Founder and Developer
<a class="moz-txt-link-freetext" href="http://www.opensips-solutions.com">http://www.opensips-solutions.com</a></pre>
On 17.10.2014 14:05, george wu wrote:<br>
</div>
<blockquote
cite="mid:6c87981c.c7d9.1491dc8637c.Coremail.aihuawu2012@163.com"
type="cite">
<div
style="line-height:1.7;color:#000000;font-size:14px;font-family:Arial">
<div>Bogdan-Andrei:<br>
<br>
Thank you very much for so many helps.<br>
Yes, tcp_connection_lifetime.<br>
As you said before, udp scale up very well. But tcp is not.<br>
However, I have no choice other than tcp. If the tcp get
closed often, is it similar to udp?<br>
<br>
Recently I modify the udp_server.c so that I defragment the
udp package manually.<br>
(For ice, the packet is about 1700+ bytes.)<br>
This approach works 80% of time.<br>
<br>
However the udp packages sometimes come out of order<br>
which makes the linphone parser gets a coredump.<br>
I googled a lot, some router's mtu is 1500 (payload 1472),
some 1400. the only requirement is minimum 576 bytes.<br>
</div>
<div>So I finally give up the udp. I think tcp is the only
reliable way to deliver big packet.<br>
Unless your packet is less than 576 bytes. But my packet is
1700 bytes.<br>
<br>
George<br>
</div>
<br>
<br>
<br>
<br>
At 2014-10-17 17:18:43, "Bogdan-Andrei Iancu"
<a class="moz-txt-link-rfc2396E" href="mailto:bogdan@opensips.org"><bogdan@opensips.org></a> wrote:<br>
<blockquote id="isReplyContent" style="PADDING-LEFT: 1ex;
MARGIN: 0px 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid">
<div class="moz-cite-prefix"><tt>Hi Geroge,<br>
<br>
You want to change the keepalive or the lifetime of the
TCP conn ?<br>
<br>
As it sounds to me you are more looking for lifetime (to
have closed asap):<br>
<a moz-do-not-send="true"
class="moz-txt-link-freetext"
href="http://www.opensips.org/Documentation/Script-CoreParameters-1-11#toc91">http://www.opensips.org/Documentation/Script-CoreParameters-1-11#toc91</a><br>
<br>
Regards,<br>
</tt>
<pre class="moz-signature" cols="72">Bogdan-Andrei Iancu
OpenSIPS Founder and Developer
<a moz-do-not-send="true" class="moz-txt-link-freetext" href="http://www.opensips-solutions.com">http://www.opensips-solutions.com</a></pre>
On 17.10.2014 02:44, george wu wrote:<br>
</div>
<blockquote
cite="mid:478de7f3.666.1491b5932da.Coremail.aihuawu2012@163.com"
type="cite">
<div
style="line-height:1.7;color:#000000;font-size:14px;font-family:Arial">
<div>Hi, guys:<br>
<br>
How can I adjust the global tcp keep alive time? what's
the default in opensips?<br>
I want it to close as soon as possible so that tcp
deployment can scale.<br>
<br>
George Wu<br>
<br>
</div>
</div>
<br>
<br>
<span title="neteasefooter"><span id="netease_mail_footer"></span></span><br>
<fieldset class="mimeAttachmentHeader"></fieldset>
<br>
<pre wrap="">_______________________________________________
Users mailing list
<a moz-do-not-send="true" class="moz-txt-link-abbreviated" href="mailto:Users@lists.opensips.org">Users@lists.opensips.org</a>
<a moz-do-not-send="true" class="moz-txt-link-freetext" href="http://lists.opensips.org/cgi-bin/mailman/listinfo/users">http://lists.opensips.org/cgi-bin/mailman/listinfo/users</a>
</pre>
</blockquote>
<br>
</blockquote>
</div>
<br>
<br>
<span title="neteasefooter"><span id="netease_mail_footer"></span></span>
</blockquote>
<br>
</body>
</html>