On 2003.09.16 09:52, Ramin Dousti wrote:
On Tue, Sep 16, 2003 at 09:42:20AM -0600, Jim Redman wrote:
> Ramin, > > On 2003.09.16 08:24, Ramin Dousti wrote: > >Do the packets belong to one UDP session? If not, this number of > >packets > >might overflow your connection-tracking table. > > The packets are all individual entities. Any solution if this is the > problem? Any way to test? (There are no indications in the syslog > that I've noticed).
I'm sure there are better ways of doing this but a simple
cat /proc/net/ip_conntrack | wc
Seems I must have told you an untruth since this is all I have in the file relating to the particular connection. 192.168.2.189 is the device sending data to me and 192.168.2.97 is me. I suspect the first line is a connection from my app back to the device, since it only appears when the app is running.
udp 17 179 src=192.168.2.97 dst=192.168.2.189 sport=32815 dport=995 src=192.168.2.189 dst=192.168.2.97 sport=995 dport=32815 [ASSURED] use=1
udp 17 179 src=192.168.2.189 dst=192.168.2.97 sport=995 dport=995 src=192.168.2.97 dst=192.168.2.189 sport=1995 dport=995 [ASSURED] use=1
Today, however, everything is working (at least at the moment)! It's very strange. Nothing has changed. I took the notebook home and brought it back. Same software running here, same device still sending the values, etc. etc. I still don't know that it's a netfilter problem. I see the packet with tcpdump and don't receive them in the application. NMAP says the port is open and the code is just open connection & receive (in Java).
Ugh. Well, better get on with the debugging while I'm receiving data.
Ji
--
Jim Redman (505) 662 5156 x85 http://www.ergotech.com
