Re: [naviserver-devel] Naviserver as reverse proxy - header fields

[Gustaf N. <ne...@wu...>]

Both are valid requests and should go to the source base. david, could 
you issue a pull request on bitbucket?
-g

Am 26.06.15 um 11:51 schrieb David Osborne:
> Hi,
>
> We've been looking at using Naviserver as a reverse proxy so it can 
> serve up fastpath requests, and also forward requests to an upstream 
> app server when required.
>
> There was a couple of issues we came across that I'd like to ask about.
> The proxy server code follows the same basic idea as Gustaf's post 
> here: http://sourceforge.net/p/naviserver/mailman/message/31063859/
>
> 1. Host header fields.
> When we forward a request upstream using ns_ssl I couldn't find a way 
> of getting the Host header field to be forwarded untouched by ns_ssl.
>
> I pass the query headers into ns_ssl which include a Host field 
> containing the host originally request by the client, and ns_ssl 
> always overwrites it with the value of the upstream Host. I don't 
> think this behaviour is wrong by any means, but, in Apache for 
> example, there is an option which turns on the behaviour we were 
> looking for - proxyPreserveHost
> http://httpd.apache.org/docs/2.2/mod/mod_proxy.html#proxypreservehost
>
> Could there be scope for such an argument for ns_ssl (and ns_http as 
> well I guess) or is there a better way to deal with this?
> Something like this:
> https://bitbucket.org/davidqc/nsssl/branches/compare/davidqc/nsssl:tip%0Dnaviserver/nsssl:default#diff
>
>
> 2. Duplicate headers added by ns_respond.
> Once upstream has replied to the proxy, the response is passed back to 
> the client using ns_respond with the headers we received from 
> upstream. What we are seeing is duplicate content-type and 
> content-length headers being received by the client.
> It seems like this may be because Ns_ConnUpdateHeaders is using a case 
> sensitive Ns_SetUpdate to update the content-type header. So if 
> content-type already exists, it will add Content-Type as well.
>
> Should this behaviour be that a case insensitive Ns_SetIUpdate is used 
> like the following or are there other ways around this?:
> https://bitbucket.org/davidqc/naviserver/branches/compare/davidqc/naviserver:tip%0Dnaviserver/naviserver:default#diff
>
>
> Thanks
> -- 
> David Osborne
> Qcode Software Limited
> http://www.qcode.co.uk
>
>
>
> ------------------------------------------------------------------------------
> Monitor 25 network devices or servers for free with OpManager!
> OpManager is web-based network management software that monitors
> network devices and physical & virtual servers, alerts via email & sms
> for fault. Monitor 25 devices for free with no restriction. Download now
> http://ad.doubleclick.net/ddm/clk/292181274;119417398;o
>
>
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel


-- 
Univ.Prof. Dr. Gustaf Neumann
WU Vienna
Institute of Information Systems and New Media
Welthandelsplatz 1, A-1020 Vienna, Austria