On the demand pppd setup I am working on I have a box that I used a browser (lynx) to view the configuration and statistics for a switch with a web interface. The switch is on the other side of my demand pppd link. after I closed lynx and the link shut down, it kept redialing and then hanging back up and then about a minute,later hanging up again. using tcp dump and netstat I was able to track th problem back to the machine I viewed the switch on. The send queue had about 20 or so bytes stacked up and now that I am keeping the link up via ping it is slowly going down. I am assuming that if and when it gets to 0 in the send Q then the connection will close and it won't be keeping my pppd link up. Is this normal behavior? I know I did at least part of it to my self by letting the link only be idle for 20 secinds before shutting down (testing ya' know) but it seems really curious that the last_ack count continued to climb even though the link was broken. BTW I do believe one of the stats page I was looking at has a refresh coded into it every 40 seconds or so. Here is the relevent line from netstat Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name tcp 0 13 192.168.0.13:3735 192.168.1.2:80 LAST_ACK - I've got that learning feeling again. Can someone tell me what is going on? Bret -- To unsubscribe: mail [EMAIL PROTECTED] with "unsubscribe" as the Subject.
