I received 2 replies for this which both are indeed helpful:

a. Setting the time_wait interval to a shorter period (in this sample,
60 secs):
ndd -set /dev/tcp tcp_time_wait_interval 60000
b. Increasing the number of ports in the /etc/system file if you're
using Solaris 7.
set pty_cnt=3Dnumber (default in Solaris 7 is 48, max. number
is 3000)

Many thanks to (a) Casper Dik [mailto:Casper.Dik at Sun.COM] and (b)
Arvinder_Dadhwal [mailto:Arvinder_Dadhwal at satyam.com]=20

Rgds,

Hendra

-----Original Message-----
From: Hendra=20
Sent: Tuesday, 3 July 2001 06:18
To:codeprof at codeprof.com
Subject: "socket: All ports in use" Error
Importance: High

Gurus,

One of my nodes does lots of rsh'ing to other unix boxes which works
fine most of the time. But during periods of heavy use, I
sometimes get:

socket:All ports in use

and the rsh does not work. Is there any way to check the availability of
the ports? How to free up ports?

TIA,

Hendra