I’m experiencing very long connect (actually login) times on a QNX 6.1 box. Once the login prompt appears, things seem to proceed normally (that is, reasonable response time to commands). But the login prompt sometimes takes as long as 30 seconds to appear. Slinger’s running, and I can fetch web pages, and I can ping the box successfully with no excessively long round-trip times reported and no packets lost. netstat -ni shows lots of inbound/outbound packets, and no collisions or errors. Any ideas what’s causing this behavior?
It sounds like a name lookup problem. Can the QNX box find the name of the machine you are telneting from?
If you don’t have control over the dns server, the easiest “fix” is to add the IP address/hostname of your Windows box to the /etc/hosts file on the QNX box.
Almost immediately after posting I started thinking about DNS problems, and it appears that may be the culprit. Unfortunately “patching” addresses into /etc/hosts won’t solve the problem, since this is a machine that my students are using to develop their solutions to my little assignments 
and they’re telnetting in from lots of different places. Once they’ve got what they think is a solution, then we run it on the non-networked machine with the real device (since they have to do port I/O in superuser mode).
Is there a hidden flag I can use with telnetd (I presume it’s telnetd) to tell it not to lookup the reverse hostname lookup, like the -n flag for netstat?
–Stan Wileman
I don’t think there is a flag to telnetd that you can use to disable lookup. One thing you can do though is to completely turn off dns on this qnx box.
mv /etc/resolv.conf /etc/resolv.conf.sav
setconf RESOLVE “”
This should take care of it.