Messages in this thread | | | Date | Sun, 31 May 1998 19:18:31 +0200 (MET DST) | From | Carsten Gross <> | Subject | Re: 2.1.103 won't connect to ftp.kernel.org |
| |
Hello Andi and Developer List
[Cc'ed to the List]
On Sun, 31 May 1998 ak@muc.de wrote: > That's very strange. Could you try to turn off RFC1323 > (echo 0 > /proc/sys/net/ipv4/tcp_rfc1323), SACK (the same for > tcp_sack) and pmtu discovery (echo 1 > ..../ip_no_pmtu_disc)
Yes, I tried to turn everything as you pointed it. BTW: It seems that in 2.1.103 some of the configuration 'files' are named in a different way. Is it correct to use "tcp_timestamps" and "tcp_window_scaling" as replacement for "rfc1323", which doesn't exists?
I've set python:/proc/sys/net/ipv4>cat tcp_sack tcp_timestamps tcp_window_scaling ip_no_pmtu_disc 0 0 0 1
This is as you like it?
Okay, lets try again, tcpdump on one console, ncftp on the other:
18:38:20.206606 python.wohnheim.uni-ulm.de.1174 > linux.kernel.org.ftp: S 2442966210:2442966210(0) win 32696 <mss 536>
18:38:22.105328 linux.kernel.org.ftp > python.wohnheim.uni-ulm.de.1174: S 2317914197:2317914197(0) ack 2442966211 win 32696 <mss 536> (DF)
18:38:22.105502 python.wohnheim.uni-ulm.de.1174 > linux.kernel.org.ftp: . ack 1 win 32696
18:38:22.822946 linux.kernel.org.ftp > python.wohnheim.uni-ulm.de.1174: . ack 1 win 32696 (DF) [tos 0x10]
18:38:24.211198 linux.kernel.org.ftp > python.wohnheim.uni-ulm.de.1174: . ack 1 win 32696 (DF) [tos 0x10]
18:38:40.273347 linux.kernel.org.ftp > python.wohnheim.uni-ulm.de.1174: . ack 1 win 32696 (DF) [tos 0x10]
[Yes, the 30 seconds are oooover...]
18:38:52.098709 python.wohnheim.uni-ulm.de.1174 > linux.kernel.org.ftp: F 1:1(0) ack 1 win 32696
18:38:53.959903 linux.kernel.org.ftp > python.wohnheim.uni-ulm.de.1174: . ack 2 win 32696 (DF) [tos 0x10]
18:38:53.962660 linux.kernel.org.ftp > python.wohnheim.uni-ulm.de.1174: . ack 2 win 32696 (DF) [tos 0x10]
I played around a little bit, and suddenly after a while it worked! But this was not reproducible :/
tcpdump from the one correct working connect: 18:48:45.425304 python.wohnheim.uni-ulm.de.1179 > linux.kernel.org.ftp: S 3083732050:3083732050(0) win 32696 <mss 536>
18:48:46.160564 linux.kernel.org.ftp > python.wohnheim.uni-ulm.de.1179: S 2946181376:2946181376(0) ack 3083732051 win 32696 <mss 536> (DF)
18:48:46.160784 python.wohnheim.uni-ulm.de.1179 > linux.kernel.org.ftp: . ack 1 win 32696
18:48:46.900507 linux.kernel.org.ftp > python.wohnheim.uni-ulm.de.1179: P 1:108(107) ack 1 win 32696 (DF) [tos 0x10]
[THIS is the main difference between working or not, there are 107 bytes of data in this paket]
18:48:46.906552 python.wohnheim.uni-ulm.de.1179 > linux.kernel.org.ftp: P 1:17(16) ack 108 win 32696
18:48:47.715742 linux.kernel.org.ftp > python.wohnheim.uni-ulm.de.1179: P 108:176(68) ack 17 win 32696 (DF) [tos 0x10]
18:48:47.718300 python.wohnheim.uni-ulm.de.1179 > linux.kernel.org.ftp: P 17:58(41) ack 176 win 32696
[ and so on.. it's not interesting for you to read further the log of a normal connect]
It seems, that ftp.kernel.org has the error/problem, not my system. The working connect 'pushes' 107 Bytes of data in the first data-paket. In the case of non working connect, ftp.kernel.org just sends an empty 'ACK' at this point, without any data.
BTW: finger @ftp.kernel.org works. telnet ftp.kernel.org has the same problem as ftp.
Regards
Carsten
-- unzip ; touch ; finger ; mount ; gasp ; yes ; umount ; sleep Carsten Gross carsten@sol.wohnheim.uni-ulm.de Wohnheim Heilmeyersteige Sebastian Kneipp Weg 6, 89075 Ulm
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu
| |