Alexander Hall wrote:
Jeremy Bowen wrote:
Doh! <facepalm>
After sleeping on this overnight I've discovered a misconfigured server.
All working as desired now.
Too embarrassing to share for the archives? :-9
Well, almost. But here we go :-)
I think I know what the problem was but I'm not 100% sure of the failure
mechanism.
I've just shifted my webserver to a newly installed OpenBSD v4.4 box.
I'm doing this behind NAT (pf rdr) and using Virtual hosting.
For various reasons, I'm running an internal DNS which I had forgotten
to update to point to the new webserver.
The clue was that telnetting to port 80 on localhost showed activity in
the logs/mod_gzip.log file but specifying the Host: header for a virtual
host in an HTTP/1.1 request, didn't show any gzip activity, but did show
in the logs/access_log file.
Updating the misconfigured DNS records seems to have cured the mod_gzip
problem. I can't quite explain why this would have affected mod_gzip as
the new server was definitely serving up the pages (verified by checking
logs/access_log) but that was the fix.
In addition, the "http://www.whatsmyip.org/mod_gzip_test/"; site I was
using to verify my setup appears (I'm guessing) to be geared towards
apache v2.x as having finally seen info in my logfiles, it isn't sending
the expected GET request first. Instead it sends a HEAD which mod_gzip
v1.3.x doesn't handle. So even when I had mod_gzip working, it would lie
to me and say that it wasn't. Testing with perl/telnet verified that the
server is now behaving as desired.
I doubt this will be of any use to anyone else (except maybe for some
light entertainment value) but there you have it.
Cheers
