Good thought. I don't think it's the case this time, though. The error
occurs when I switch WKServer to port 8087 and put the load-balancer into
the mix. Same context, same code, same database backend, etc. With
load-balancer, the error shows up; without load-balancer, no error. I'm
stopping and starting Apache and WebKit before each test
So obviously the load-balancer is part of the problem. But in other TCP
applications I've load-balanced with it, it's as if it's not even there
(that is, it seems like a direct connection to the destination server).
Are there other conditions (besides munging the headers after they're
already streamed) that are known to cause the error? The place that raises
it in mod_webkit.c looks like the following (but I wouldn't know where to
start debugging it):
// log_message("scanning for headers",r);
//pull out headers
if ((ret=ap_scan_script_header_err_buff(r, buffsocket, NULL))) {
if( ret>=500 || ret < 0) {
log_message("cannot scan servlet headers ", r);
return 2;
}
r->status_line = NULL;
}
Cheers!
--
David Hancock | [EMAIL PROTECTED] | 410-266-4384
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
Sent: Friday, August 29, 2003 8:56 PM
To: Hancock, David (DHANCOCK); [EMAIL PROTECTED]
Subject: Re: [Webware-discuss] Problem using WebKit through a load-balancer
--> Friday, August 29, 2003, 7:35:39 PM, [EMAIL PROTECTED] wrote:
> I'm experimenting with using pythondirector for load-balancing several
> WebKit servers, but I get the following error messages in the Apache log:
> [Fri Aug 29 23:59:34 2003] [error] [client 144.243.111.8] Premature
> end of script headers: /home/adc/adc/wwwroot/ADC
> [Fri Aug 29 23:59:34 2003] [error] cannot scan servlet headers
> [Fri Aug 29 23:59:34 2003] [error] error transacting with app server
> -- giving up
I know nothing of your specific implementation, but this error is fairly
common.
If I remember right, this occurs when the headers have already been
streamed,
but the dynamic engine (webware/php/etc) tries to add to them.
I suppose a problem like this could show up by circumstance when you add
threads/processes to a previously working system. Have you tried very simple
no code no content tests? If so, maybe WebWare itself is trying to do
something to headers incorrectly...
If you think about what the error means aside from common concurrency
problems, perhaps something is trying to start a new response on top of one
which is currently being sent in the same thread.
That's my 2 cents. (or less...)
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
Webware-discuss mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/webware-discuss
