![]() | |||||||||||||
Messages in this thread |
On Wed, Jun 17, 1998 at 12:11:29AM +0100, you [Alan Cox] claimed: > > I can't imagine how this could be a bug in ssh (the connect() > > parameters are perfectly correct according to strace), but > > I never get these stalls with other programs doing connect(). > > Further more, I've not seen this on any other RH 5.1 or kernel > > 2.0.34. > > Its probably happening in a case where the socket in question is > in TIME_WAIT from a previous connect. 2.0.34 ought to be replying > to that however. Can you send me a tcpdump. I think I know the fix > especially if 2.0.33 doesnt show it. > > Alan I can't say 2.0.33 doesn't do that, since I've not tested 2.0.33 on the same platform. I can say that I've never seen this on the other 2.0.33-machines I've used (much more than 2.0.34), though. The machine I've tried to connect is 2.0.33, though, but as I said I could even telnet to the same port from the same machine while ssh was stalling in connect(). BTW: The local port allocated for the lagging connection always seems to be the same (1023). In netstat it says tcp 0 2 client:1023 server:22 SYN_SENT From tcpdump I get [all the packets related to the local port connect()] reports having had via strace. 06:42:07.856907 client.1023 > server.22: S 2665709351:2665709351(0) win 512 <mss 1460> 06:42:10.856907 client.1023 > server.22: S 2665709351:2665709351(0) win 32120 <mss 1460> 06:42:16.856907 client.1023 > server.22: S 2665709351:2665709351(0) win 32120 <mss 1460> Hope I did that too much, ask for more if so. -- v -- v@iki.fi - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu | ||||||||||||
| Last update: 2005-03-22 12:43 [W:0.163 / U:0.030 seconds] ©2003-2008 Jasper Spaans | |||||||||||||