On Thu, Jun 30, 2005 at 01:53:11AM +0200, Damjan wrote:
> Apache's log was full of
> [Wed Jun 29 13:55:15 2005] [warn] (101)Network is unreachable: connect to
listener
> messages, and when I tried to restart it, I couldn't. All of the apache
> children processes were in status Z (zombie), which suggests that the
> master process didn't react on SIGCLD signals.

It does seem possible that the problem could be a fd leak.  However,
after looking at the code I don't see how that could happen.

Do you have still have the logs?  Are there any suspicious error
messages before the "Network is unreachable: connect to listener"
messages start to appear?

  Neil