httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Christoph Anton Mitterer <cales...@scientia.net>
Subject Re: [users@httpd] reverse proxy to http 1.0 origin server
Date Mon, 05 Aug 2013 22:25:15 GMT
Hi Nick.



On Mon, 2013-08-05 at 15:38 +0100, Nick Kew wrote: 
> > AFAIU, strict HTTP 1.0 has neither persistent connections / keep-alives
> > - a connection ends after a single request has been responded.
> > Neither does it have Host: headers.
> 
> No, it has keepalives and Host headers.  But the former are more
> limited than in 1.1 because there's no chunked encoding, and the
> latter are technically optional (though in practice pretty-much
> universal since about 1995 or 96).

Sure? I'm talking about strict RFC compliant HTTP 1.0, i.e. RFC 1945
There is no Host Header or keep alives named... at least I wouldn't have
found it.


> > a) Do I need to tell the reverse-proxy about this?
> Try it and see what your backend is happy with!
> Chances are, the only thing you'll need to worry about
> is to avoid sending it chunked encoding.

Well it's happy with anything... that doesn't really help me though.

I'm trying to find out the ideal settings with best performance so that
we could give this as howto for people using SKS
(https://bitbucket.org/skskeyserver/sks-keyserver)

And there we have a single threaded webserver, which speaks HTTP1.0.
Therefore the question about max=1 and how far Apache
auto-detects/auto-sets the other env flags when speaking to HTTP1.0
backends.

Also as per HKP (http://tools.ietf.org/html/draft-shaw-openpgp-hkp-00)
clients MUST be able to speak HTTP 1.0 and receive 1.0... (but they MAY
use 1.1 as well)... so therefore the question about force-response-1.0.


So... could you perhaps have a closer look at my original questions if
possible and trying to answer what you know? :)


Cheers,
Chris.

Mime
View raw message