>More importantly, it's the exact same cookie domain >and path if I use the server IP or name. It appears the cookie manager won't >*send* it if I'm using the IP. and that behavior is correct. It is exactly how your browser would behave as well. Two DNS names , even if they map to the same IP address arent the same for cookies. Technically your application should not be setting cookies based on the host header , it should be based on what the user has typed in their browser.
Other than a beanshell script that would copy the cookie to the IP address , I dont see how you could get this to work without actually requesting the DNS name from Jmeter (or an entry in the hosts file on the jmeter machine) regards deepak On Wed, Dec 23, 2009 at 11:27 AM, Trenton Lipscomb <trent...@gmail.com>wrote: > > I guesss the cookie is tied to the domain it receives (if you access a > site > > using Internet explorer with the IP address and get a cookie and then > > access > > it again but this time type the DNS , the browser will not send the > > cookie). > > > > Agreed, but that's slightly different that what I'm doing. That scenario > would be analogous for using an IP in the Host header. > > > > if you add a view results tree listener and check the Set-Cookie you > > receive > > does it have a domain? > > > > It does. It's the exact same as the Host I'm requesting > > domain=.beta.mysite.com > > That would make sense, since it's Apache setting the cookie and it knows > which vhost I requested. More importantly, it's the exact same cookie > domain > and path if I use the server IP or name. It appears the cookie manager > won't > *send* it if I'm using the IP. Perhaps it's using the server name as the > key > and not the Host header? > > > > Also if you dont add a host header and just request using IP address what > > happens? > > > > Won't work because I have the site configured as a virtual host. It must > get > that Host header to know which host/docroot I want. >
