Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 29 Mar 1996 12:43:54 -0800 (PST)
From:      Bryan Ogawa at Work <bogawa@netvoyage.net>
To:        dwhite@resnet.uoregon.edu
Cc:        Steve <shovey@buffnet.net>, freebsd-questions@FreeBSD.org
Subject:   Re: Apache still and timeouts
Message-ID:  <Pine.NEB.3.92.960329121302.1474F-100000@digital.netvoyage.net>
In-Reply-To: <Pine.BSF.3.91.960328194216.1542A-100000@riley-net170-164.uoregon.edu>

next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 28 Mar 1996, Doug White wrote:

> On Thu, 28 Mar 1996, Steve wrote:
>
> >
> > Im still having apache troubles (1.0.0) under freebsd 2.1R and I now
> > think it has to do with timeouts under apache.
>
> Try 1.0.3.  New port should be in ports-current.
>
> Have it on a 2.1R box and it is running great.

I've run it (quick tests ONLY -- we use NCSA 1.5, which I'm running now on
our 2.0.5R machine) with few problems (actually, fewer problems w/ Apache
than NCSA).  The only reason I'm still running NCSA is due to minor
incompatibilities between them (and we're kinda planning on switching
everything all at once).

One thing that I read somewhere (on this list, perhaps) is that setting :

DNSMode Minimum

in your httpd.conf will help.  What it does is stop reverse DNS lookups on
IP to find the hosts that are hitting you.

So, your access_log only has IPs, which may be irritating if you're trying
to do analysis.  This can be overcome with a little perl script.

I dunno if I can distribute my version to do the lookups and rewrite the
access_log (this is at work), but it's really easy, so I'll:

(a) describe the algorithm briefly

	and

(b) offer to rewrite it at home (email bkogawa@netvoyage.net to remind
me).

The basic algorithm:

	for each line in the access_log:
	strip out the IP
	look it up.  If you can find it, rewrite the line with the name.

:)

One trick is that you can use is an associative array to cache the
lookups.  This can lower it to one ns request per IP, at the cost of
memory.  This is how I was doing it when I was testing my software, but I
suspect performance could get really horrible if the logs were too big and
memory ran out.

bryan

>
> Doug White                              | University of Oregon
> Internet:  dwhite@resnet.uoregon.edu    | Residence Networking Assistant
> http://gladstone.uoregon.edu/~dwhite    | Computer Science Major
>
>

Bryan K. Ogawa
Questions or Problems with NetVoyage?  help@netvoyage.net
Check out the NetVoyage HelpWeb at..   <URL: http://www.netvoyage.net/~help/>;




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.NEB.3.92.960329121302.1474F-100000>