According to the archives, if you insist on running inetd with
discard active, then you need to set tcp_keepalive_default=1
under inet subsystem, which I did, using the dxkerneltuner.
But, it's back. An obvious question would be: If you change
the current parameter from 0 to 1 using the dxkerneltuner and
it supposedly immediately changes the kernel parameter and you
also change the boottime parameter to 1, do you really have to
reboot the system in order to make discard stop acting this way?
Looking for other suggestions.
Lynn Bogovich
Sr UNIX System Administrator
Space Telescope Science Institute
# sysconfig -q inet
inet:
inifaddr_hsize = 32
ipdefttl = 64
ipdirected_broadcast = 0
ipforwarding = 0
ipfragttl = 60
ipgateway = 0
ipport_userreserved = 5000
ipport_userreserved_min = 1024
ipqmaxlen = 512
ipqs = 1
ipsendredirects = 1
ipsrcroute = 1
pmtu_decrease_intvl = 1200
pmtu_enabled = 1
pmtu_increase_intvl = 240
pmtu_rt_check_intvl = 20
subnetsarelocal = 1
tcbhashnum = 1
tcbhashsize = 512
tcbquicklisten = 1
tcp_compat_42 = 1
tcp_cwnd_segments = 2
tcp_dont_winscale = 0
tcp_keepalive_default = 1
<snip>
# ps -ef |grep discard
root 32591 17077 0.0 19:30:19 ttyp5 0:00.00 grep discard
#
# telnet machinename 9
Trying xxx.xxx.xx.xx...
Connected to machinename.domain.edu.
Escape character is '^]'.
^]
telnet> Connection closed.
# ps -ef |grep discard
root 5165 13606 90.6 19:30:27 ?? 0:08.33 -discard [130.16 (inetd)
root 25514 17077 0.0 19:30:48 ttyp5 0:00.00 grep discard
#
Received on Fri Mar 31 2000 - 19:34:27 NZST