Sockets stuck in FIN_WAIT1 (2.2.14)

From: Mike A. Harris (mharris@meteng.on.ca)
Date: Thu Jan 20 2000 - 00:37:00 EST

I don't know what the problem is here but it is driving me
absolutely crazy to the extreme.

My kernel is 2.2.14, and this happened with 2.2.13 as well. FTP
connections to ANY site in the redhat.com domain (ftp.redhat.com,
updates.redhat.com, contrib.redhat.com) as well as
metalab.unc.edu, all HANG in FIN_WAIT1 after attempting to
download a file.

If I download any file, or try to view a file in any of the
aforementioned ftp sites, Midnight Commander (my FTP client) sits
there frozen until I kill it from another VC. Prior to killing
it, I do a netstat -nt, and see:

2 root@asdf:/usr/src# netstat -nt
Active Internet connections (w/o servers)
Proto Recv-Q Send-Q Local Address Foreign Address State
tcp 0 0 206.248.78.167:1293 152.19.254.81:18982 CLOSE_WAIT
tcp 0 0 206.248.78.167:1278 152.19.254.81:21 ESTABLISHED
tcp 0 1 206.248.78.43:1272 152.19.254.81:22819 LAST_ACK
tcp 0 1 206.248.78.43:1262 152.19.254.81:21 FIN_WAIT1
tcp 0 7 206.248.78.23:1211 209.10.41.242:21 FIN_WAIT1
tcp 1 0 206.248.78.192:1170 206.168.112.90:80 CLOSE_WAIT
tcp 1 0 206.248.78.192:1169 206.168.112.90:80 CLOSE_WAIT
tcp 1 0 206.248.78.192:1168 206.168.112.90:80 CLOSE_WAIT
tcp 0 0 127.0.0.1:7101 127.0.0.1:1028 ESTABLISHED
tcp 0 0 127.0.0.1:1028 127.0.0.1:7101 ESTABLISHED

Every time this freeze occurs, I check netstat -nt and see the
"state" is FIN_WAIT1. If I sit here waiting, it is an indefinite
wait. This occurs 100% of the time when I connect to any of the
sites mentioned above. I can change dirs easily on these sites,
but the second I download a SINGLE file, I get the file intact
and it is saved, and complete, but the FTP client hangs, and the
connection shows FIN_WAIT1 in netstat.

I've tried other ftp clients such as ncftp, ftp, and get the same
results often.

Is anyone else having trouble like this? If I go to a redhat
mirror site the problem goes away. Come back to redhat.com and
it is back. My guess is that something in the configuration of
redhat.com, and metalab, and possibly other sites is screwed up.

Are there tcp/ip settings that can be played with in /proc on a
linux system which could have a negative result like I'm
experiencing? I'm refering to the REMOTE computer that is as I'm
using a default setup with no tweaks..

Any help publically or privately to help narrow down this
incredibly annoying glitch would be greatly appreciated in the
extreme!!!

Is it possible the 2.2.x kernel still contains some TCP/IP
glitches that could be causing this? What about the remote
sites? Does anyone know what kernel redhat.com, metalab, etc..
uses, and what FTP server software?

Thanks in advance.
TTYL 

--
Mike A. Harris                                     Linux advocate     
Computer Consultant                                  GNU advocate  
Capslock Consulting                          Open Source advocate

Join the FreeMWare project - the goal to produce a FREE program in
which you can run Windows 95/98/NT, and other operating systems.

                    http://www.freemware.org

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

This archive was generated by hypermail 2b29 : Sun Jan 23 2000 - 21:00:21 EST