connections still hang in CLOSED

[steve at cdp.UUCP]

This bug report is addressed primarily to Interactive Support, in
particular, Marty C. Stewart, Support Team Leader.

...........................................................


BUG REPORT
==========

I. Summary

The problem with tcp/ip connections hanging in CLOSED state,
and accepting no more connections on a particular tcp port
continues in recent versions of interactive UNIX and interactive
TCP/IP.  The bug was reported many times in this
group.  Interactive's SSU.41a TCP/IP patch decreased the
frequency of the bug.


II. Hardware/software

I run Interactive UNIX Release 2.2.1 with the SSU.41a tcp/ip patch.
Machine has a mylex MWS-20 motherboard, adaptek 1542B
disk controller, and Western Digital 16 bit Ethercard Plus.


III. Description of problem

Ocasionally it becomes impossible to 'rcp' to the UNIX host.
rcp requests return with a timeout error message.  A reboot
solves the problem.


IV. Detail

I try the following from the machine cdp.igc.org  :

   rcp /tmp/steve sf.igc.org:/tmp/steve

It returns the error condition :

   sf.igc.org: Connection timed out

Other network services, such as telnet and ftp continue to work fine.

'netstat -a' on sf.igc.org indicates that 'shell' is stuck in CLOSED
state :

 Active Internet connections (including servers)
 Proto Recv-Q Send-Q  Local Address          Foreign Address        (state)
 tcp        0      0  sf.igc.org.login       cdp.igc.org.1023       ESTABLISHED
 tcp        0   5120  sf.igc.org.shell       cdp.igc.org.1023       CLOSED
 tcp        0      0  *.time                 *.*                    LISTEN
 tcp        0      0  *.daytime              *.*                    LISTEN
 tcp        0      0  *.chargen              *.*                    LISTEN
 tcp        0      0  *.discard              *.*                    LISTEN
 tcp        0      0  *.echo                 *.*                    LISTEN
 tcp        0      0  *.exec                 *.*                    LISTEN
 tcp        0      0  *.login                *.*                    LISTEN
 tcp        0      0  *.shell                *.*                    LISTEN
 tcp        0      0  *.telnet               *.*                    LISTEN
 tcp        0      0  *.ftp                  *.*                    LISTEN
 tcp        0      0  *.*                    *.*                    CLOSED
 udp        0      0  *.3131                 *.*                   
 udp        0      0  *.1025                 *.*                   
 udp        0      0  *.1024                 *.*                   
 udp        0      0  *.syslog               *.*                   
 udp        0      0  *.*                    *.*                   

'ps -ef' on sf.igc.org shows that there are no unusual network-related
processes :

  UID   PID  PPID  C    STIME TTY      TIME COMMAND
 root    72     1  0  Feb 13  ?       10:47 netsched
 root    74     1  0  Feb 13  ?        0:00 /etc/netd 
 root    85     1  0  Feb 13  ?        3:17 /etc/rwhod 
 root    87     1  0  Feb 13  ?        0:45 /etc/syslogd 
 root    84     1  0  Feb 13  ?        0:03 /etc/inetd 
 root  6385    84 10 12:14:45 ?        0:01 rlogind 

A reboot returns 'rcp' to a working state for a few days.


V. Resolution

I am happy to assist interactive's support and engineering staffs
in any way practical, to repair the bug.

Could you please give me some information :

  1. Is this a new or standing bug report ?

  2. What is the schedule for fixing it ?

I'm fine with, "we don't consider this an important bug, 
and so it may not be fixed for many months."  I would not
be well served by an no acknowledgement or an ambiguous answer.


Thank you for the service,


Steven Fram
Technical Director
Institute for Global Communications (IGC)
steve at igc.org
cdp!steve