> this loop, It entered but never got out. Another machine, No problem, runs
> straight through this loop. So after having thought a lot over what on
> earth this loop really does I finally concluded that it doesn't do
> anything, cept maybe pauses something (Not that I can see what). So I
> finally decided to give in and kill the loop, thereby changing the
> function to:
>
> -- *SNIP* --
> if (select(0, (fd_set *) 0, (fd_set *) 0, (fd_set *) 0, &timeout) < 0) {
> if (errno != EINTR) {
> perror("Select sleep");
> exit(1);
> }
> }
> -- *SNIP* --
We had this very same problem, running a very modified bpl11 on RH5.0. I
couldn't fathom what the problem was, but a look at netstat -a showed LOTS
of sockets that were in time waits, or closed state attached to our port,
but that wouldn't go away, even after killing all instances of the mud.
We've been running on this server for 2 years, and hadnt seen this
problem, hadn't changed comm.c, and weren't experiencing exeptional
connection numbers or uptime. I was flummoxed.
A quick email to our sysadmin showed that nothing on the box had changed,
but they detected some misconfiguration with our SMP support, and a few
other little things (nothing related to networking) and rebooted us,
after recompiling the kernel. This was about a month ago, maybe more, and
we've seen no repeat of this problem.
I realize this doesn't offer much in the way of help, but maybe someone
that understands it all better will be able to use these clues.
Ghost Shaidan
Implementor of Questionable Sanity
qs.mudservices.com 4000
+------------------------------------------------------------+
| Ensure that you have read the CircleMUD Mailing List FAQ: |
| http://qsilver.queensu.ca/~fletchra/Circle/list-faq.html |
+------------------------------------------------------------+
This archive was generated by hypermail 2b30 : 12/15/00 PST