Re: Logging changes in 0.7.2f

[Randy Terbush <ra...@zyzzyva.com>]

>  
> > Rob, some major changes in http_log.c seemed to have broken logging
> > for virtual hosts. It appears that you are *very* familiar with this
> > code. :-)  Any idea what might have broken this?
> 
> What's broken ?

Oh...Duh. (sorry)

Nothing is being written to the specified files. Nor is it
being written to the "default" access_log.  This seemed to be working
fine in 0.7.2d.  I'll restart things and verify the problem.

One other thing. After some children have been respawned to replace
older ones, killing the parent will not take the children down with
it. I have to kill each separately.

> I just finished removing all (I hope) the file parameters that were
> being used in almost every routine.

Just for my edification, why did you do that?

> I pounded the result from another local machine with 10,000 connections.
> (15 children, 20 requests max each)
> All appeared to work without a hitch. The server load was >0.5, while the
> client was at 30.
> The nice thing was that I was able to browse my web pages during the
> test with no noticable delay.

This thing is noticibly faster!