> requests. They seem sleeping! The telnet deamon is ok, but a little bit
> slow. However, when I reboot the system, smbd wakes up but wu-ftpd is
> still so slow.
> I've comment out the "log_on_failure" in "telnet" and "wu-ftpd" in
> /etc/xinetd.d/ but seems in vain.
I think commenting out the "log_on_failure" string simply tells xinetd
to use its default log mode (I may be wrong on this - but your case
seems to imply this). Besides, disabling failed log on messages will
make your life hard in case your system gets pummelled by brute-force
There are two reasons that this pause occurs - both of which are
frequently asked questions. During the authentication phase of an
xinetd service, it does use the "log_on_failure" string to determine
what it should put in the log.
USERID relies on the remote machine to be running the RPC identd daemon
- this is unsupported in Windows, and many machines have it turned off
by default to protect privacy (of user accounts). If there is no such
daemon enabled on the remote machine, the RPC call must first timeout
(hence the long pause) before the service can continue.
The best method to correct this is to eliminate "USERID" and replace it
with "HOSTID" so you at least derive the IP address of the machine.
Once this is done, restart the xinetd service to activate your changes.
If it still takes a long time to authenticate, it's a host name resolve
issue in which case the IP address reverse-name lookup is taking a long
time. You can correct this by adding IP addresses to your /etc/hosts
file (usually, name resolving comes from this first) or add a DNS server
to your box and assign names to machines on your network.
Software Developer, Engineering Services
Red Hat, Inc.
-> For NASA, space is still a high priority.
-> - Dan Quayle