When analyzing traffic with a packet sniffer, we are seeing an http response from a weblogic server prior to the completion of the http post to that server.
In this case, the jsp page on the server is basically a static page, no logic to do anything with the contents of the post at this time.
But why would the server send the response prior to completion of the post?
I found Weblogic documentation about how to configure the server to ignore a denial-of-service attack using Http post. Maybe that is what is happening?
No one I know has seen this behaviour before. Maybe some weblogic-savvy person will know what is going on.
Thanks
I don't think that Weblogic is analyzing the JSP to determine whether it is static or not.
My guess is that either
someone else was accessing the server at the same time
you saw the answer to a previous request
[EDIT] To determine what is going on, I suggest to set a breakpoint in the JSP. If you still get an answer without hitting the breakpoint, something further up the stack must be intercepting the request (for example, a cache).
Related
I need to know which requests a webpage sends. Basically the site i call, calls another service/api/url whatever and receives the data (probably within javascript) and show me this. Can i see all the calls it make?
Edit: concrete example:
From this site (http://www.flickriver.com/lenses/nikon/) you can choose a lens, at that moment, the page sends a request to flickr, and get all the data. But in chrome developer tools i could not see this request.
Here is a screenshot of get requests. I have looked through them but could not see any request to flickr.
The first is request to the page. And the sixth one is the picture request already, where it requests the picture by its id. So in between other 4 requests should contain a request to the external source which gives the picture id in return or do i miss sth?
And what if the backend makes this request? Do i still need to see this request in developer tools?
No, of course you cannot see the calls made by some server to another server. Why would you expect to be able to do that? Those calls have nothing to do with the browser. The browser knows nothing about those requests. The browser knows only about requests that it itself initiated. Devtools can only report on requests made by the browser. If in fact there were some way to spy on the requests made by a server to another server, it would be gaping security hole.
Is there any way to recognize (by process http packet or filtering tcp connections) does several requests belong to one opening url or another?
Try to explain in more detail.
When we open any page in browser it also initializes different requests to download images, resources, scripts. I d like to get know that some scope of requests was invoked by opening site (call it main site).
I can get referer property but in that case how to distinguish request to resorce from request to different site link on which was clicked on main site. In both cases referer will be the same.
I suspect that this problem could not be resolved, but I hope that I'm mistaken. Or you can offer some workaround.
If you are in control of the site, set a cookie or a URL parameter and check if it exists in subsequent requests.
I have an ASP MVC3 website with a rest API service.
When a user passes in an invalid API or they have been blacklisted i wish to ignore the response.
I know I could send back a 404 or pass back an 503 but if someone keeps polling me then I would ideally like to ignore the response causing a time-out their end. Thus delaying the hammering my server gets.
Is this possible within ASP.net MVC3? If so any help would be most appreciated.
Thank you
For what you want, you still need to parse the request, so it will always consume server resources, specially if you have an annoying user sending a query every 500ms...
In this situations you would block the IP / Header of the request for a period of, for example 10 minutes, but it would be a very good idea to block it on your load balancer and prevent that request that even reach your application, this is easily accomplish if you're using Amazon Services to run your Service, but all other cloud provider do support this as well, if by any means you are using a cloud hosting.
if you can only use your web application, and this is a solution that is not tested, you could add an ignored route to your routing mechanism like:
routes.IgnoreRoute("{*allignore}", new {allignore=#".*\.ignore(/.*)?"});
and upon check that the IP is banned, simple redirect using for example Response.Redirect() to your site, to a .ignore path... or, why not redirecting that request to google.com just for the fun of it?
Strange one here folks.
I'm studiying a web application's inner workings using Fiddler and have become a bit stumped. I'm requesting /account via the browser and Fiddler shows in the "Host" column website.local as expected since this is the URL the browser is pointing at.
Immediately after this a second request is made, however this time the host is services.website.com. I also cannot find any script in /account that makes this request.
So how can the Host change? Where is the response being sent to? Where can this be getting called from?
I'd expect that the server is sending a redirect to services.website.com: Fiddler isn't showing any redirects?
It depends on what really is in the first response that you receive. When you see a second request in Fiddler, does the page change too (in the browser)?
It will help a great deal if you could share the part of the Fiddler trace.
I have a asp.net 3.5 site with a *.asmx that serves several webservice methods. The only client that should be calling these methods is one I wrote, and it calls them using a POST request. However, my error logs show many InvalidOperationException errors due to these methods being called with a GET request.
Question: What might be causing these GET requests? Might proxies convert POST requests to GET requests without the client making the request knowing about it?
to expand on rusanu's answer, bots and crawlers and/or hackers?
bots and crawlers?
It's always possible there is a bug in your client app. Why not get hold of an HTTP sniffer so you can see exactly what requests are being sent.