Thank you, on the second point it seems so far that the problem is fixed. Regarding the first problem: a) the problem affects even NAT endpoints which have fixed IP on the public IP network, although another NATed endpoint with dynamic IP may have made some calls. Could a persistent problem arise just because a call is twice made from a NAted endpoint, each time with a different public IP? Or were you thinking of the possibility that the IP actually changed during the course of a call (surely fatal, though I hope only to the call in progress...)? I can assure that this did not happen.... I have looked at the output of 'netstat -apn | grep gnugk' when the problem is occurring, and I can see a TCP session on port 1720 to the last calling terminal in 'CLOSE WAIT', long after the call is finished, which does not happen when everything is working well. By the way, if another call is then made (which doesn't work of course!), then there is still left such a session in 'CLOSE WAIT', but it is not the same session as before... Would it be of any use to make a level 5 trace when the calls are not being properly connected? thanks again... Quoting Zygmuntowicz Michal <m.zygmuntowicz@xxxxxxx>: > 1. I would use netstat to see how many sockets are in use by the gk. > Another option may be that the IP is dynamic and changing. > 2. This issue has been solved in the 2.0.8 version. It is available from the > cvs. > > ----- Original Message ----- > From: "g cantor" <mfgx09@xxxxxxxxxxxxxx> > Sent: Friday, February 27, 2004 5:56 PM > > > > Help! > > I am running 2.0.7/linux, binary downloaded from gnugk.org. > > Using only for test calls (no volume), gatekeeper on public IP, call > originating > > from small gateway from behind NAT routers, routed to another gateway on > public > > IP and operating in full proxy mode, I have seen two problems: > > > > 1) After some short time (<24 hours) the gatekeeper stops forwarding media > > traffic from the gateway on public IP to the NAT system, i.e. call > signalling is > > OK, call comes up including h.245 capability negotiation etc. however no > RTP > > traffic is seen at all towards NAT device. > > Only way to restore operation is to restart gnugk. > > > > 2) on same system, gnugk crashes with following message in logfile: > > 2004/02/27 16:06:46.105 0 assert.cxx(105) PWLib Assertion > fail: > > Null pointer reference, file /home/zvision/src/openh323/src/h225_1.cxx, > line > > 350, Error=22 > > > > Can anyone suggest how to go about solving these problems? Is there any > useful > > way I can dump useful diagnostic data from the running gnugk after it > stops > > forwarding RTP to the NAT device? > > > > thanks > > -- > > Georg > > > > ------------------------------------------------------- > SF.Net is sponsored by: Speed Start Your Linux Apps Now. > Build and deploy apps & Web services for Linux with > a free DVD software kit from IBM. Click Now! > http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click > _______________________________________________ > List: Openh323gk-users@xxxxxxxxxxxxxxxxxxxxx > Archive: http://sourceforge.net/mailarchive/forum.php?forum_id=8549 > Homepage: http://www.gnugk.org/ > ------------------------------------------------------- SF.Net is sponsored by: Speed Start Your Linux Apps Now. Build and deploy apps & Web services for Linux with a free DVD software kit from IBM. Click Now! http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click _______________________________________________ List: Openh323gk-users@xxxxxxxxxxxxxxxxxxxxx Archive: http://sourceforge.net/mailarchive/forum.php?forum_id=8549 Homepage: http://www.gnugk.org/
