I tried JMeter and it's great but fails in what I need.
Basically I'm testing an ASP script that returns frameset to the client. When frameset renders it calls a couple of other asp scripts which render in separate frames. When I test this with JMeter I only get results for the output of the main script that renders first. JMeter is not behaving as browser would and is not trying to render the html that it receives, and therefore not requesting separate asp scripts and doesn't render them.
The application I'm trying to test is "Aquaforest Tiff server". And I need to know how much time it needs to load tiffserver, along with tiff images from the storage under large number of concurrent connections. Is there a way to actually simulate browser behavior and to render the script that is returned as a response and measure response times all together from start to end???
PS. If I was not clear enough, please tell me... I'll try to elaborate more if you don't understand what I'm talking about here...
Try Selenium or Google WebDriver. You can write unit tests to remote control Firefox and IE. We're doing that for integration tests and it works fine. Occasionally, i pimp one of the integration tests to run multi-threaded in loops for benchmarking.
Visual Studio Test Edition should do the trick for you. A Visual Studio web test recording will record that the page requested the sub pages as dependant http requests.
However, you can still simulate the load correctly, you just have to instruct JMeter to do the http requests for the dependant requests. To record the complete array of http requests made, try using fiddler2 to record. Fiddler2 also works to record visual studio web tests.
I found the 'Retrieve All Embeded Resources from HTML Files' option under HTTP Request and it does download resources that are more deeply nested...
But I don't think this solves my problem, because I have some resources that are downloaded with use of javascript. And I don't see them downloaded in the Results Tree. Am I right?
Related
I am doing load testing in Soap ui and working on cab booking application so set api like this for example
userSingin ,GetlocationDetails,bookCab,getDriverDetails,acceptRide and completeRide so on.
It will become on loop(on completing of one ride successfully ) so this seniors may happened like 1000 or 10000 times.
The question is how to interlink api do it a single entity and perform load test using soap ui, jmeter or any load testing tools,or by coding please help me sort out problem with simple demo.
SoapUI has some load testing capabilities so given you have a functional test ready you can run it with increased load.
If SoapUI's load testing is not enough (i.e. you cannot create the desired load or find results not too verbose/informative) you can build the same test plan using JMeter's HTTP Request samplers. If you need to send custom Content-Type and SOAPAction headers - it can be done via HTTP Header Manager. See Building a SOAP WebService Test Plan article for more details.
Last but not the least, you can either execute existing SoapUI project using JMeter using Taurus tool, see SoapUI Support for more details on the specific feature and Navigating your First Steps Using Taurus to get familiarized with the tool and its concept.
Using Fiddler, I want to send multiple requests in one hit, to check the response time from the server, if too many requests are sent at one time. Basically, I want to perform a, kind of, load testing on my service. Is there any way to perform this action? I want to repeat the process of hitting the server, again and again.
In Fiddler, you can repeat a request as many times as you like by hitting SHIFT+R on the selected Web Session. You'll be prompted for a repeat count and then Fiddler will issue the specified number of requests.
Caveat: Having said that, generally speaking, you'd want to use a tool like Telerik Test Studio's Load Test tool for a task like this. Alternatively, you could use Fiddler's Export architecture to generate a script for VS WebTest or Microsoft's free WCAT tool and use those tools to generate the load. You can then run these scripts on multiple machines from multiple networks and generate a more-realistic set of load than you could by simply running on a single client.
I've been load testing with StresStimulus today. Overall, I'm quite impressed.
It's now a standalone application (it used to be a fiddler extension). There's a 7 day free trial which allows up to 50 virtual users. Also, the setup wizard is great for beginners.
For basic load testing the trial should be fine. Consider upgrading for extensive/professional use.
I am nearing the release of an iOS app that syncs and otherwise interacts with a server. I am struggling with a testing procedure that can cover most/all possible situations. I don't have any experience with automated testing so I have been doing everything manually so far with the iPhone simulator and a physical device.
How would I start designing automated tests that can help me get better coverage of possible situations and also serve me well in the future as I make changes and add new features?
You probably need to be more specific in your question. ie. outline how you communicate with your server, what technology is being employed etc.
But as a general approach the first thing I would be doing is looking to find a way to get reproducable results from the server. For example if I send a message asking for a record with an id of 'x' then the server will alwasy return the same record with the same data. There are severa ways to do this, one would be to load a set of test data into your server. Another would be to create a local test server and talk to that instead. Another option is to avoid the server all together in your automaticed tests and mock out the communication classes in your app. It totally depends on what you are trying to test and how.
Once you have your back end dealt with you can then look into automating the tests. This very much depends on how you have dealt with the server. For example, if you are performing an integration style test where you actually talk to a server, then the test might take the form:
Reset or clear the server data.
Load it with predictable data.
Run the iOS app using some testing framework and verify any data sent from the server.
Access the server and verify any changes made there.
I followed these instructions
http://code.google.com/webtoolkit/usingeclipse.html
There appears to be no entry point function for the server? How do I run background threads or code not related to the rpc services that the server exports?
For example, what if some embedded database needs to be updated every 5 minutes. So then a background thread would fetch this new data to do the updating
GWT is client-side technology and has nothing to do with server-side. You can use any servers-side technology with it. If you use java/servlets then you can optionally use GWT-RPC, which is nice, but not required.
Web applications are based around request-reply paradigm: when there is a request, they handle it and send back the reply. Servlets are designed around this paradigm. They are used at some of the largest sites and are not just a toy (as you noted in other comment).
When you need something to run periodically, then this is usually the job for Job Schedulers. I'd recommend Quartz, which has great documentation. There is also an example how to initialize it in servlet environment.
thats not how web applications are supposed to work. Read http://code.google.com/intl/de-AT/webtoolkit/doc/latest/tutorial/clientserver.html
If you want to run some processing when request comes and potentially include some dynamic parts, you can just make your pages to be JSP or servlets. GWT does not need to be used in HTML files. Just the page served by server need to be HTML. So something like server side entry point is either JSP or servlet. Otherwise you need to use PRC. But if you needed to run RPC for every page loaded, you could consider this tip of embedding RPC in the base response.
The application is simple, an HTML form that posts to a Perl script. The problem is we sometimes have our customers upload very large files (gt 500mb) and their internet connections can be unreliable at times.
Is there any way to resume a failed transfer like in WinSCP or is this something that can't be done without support for it in the client?
AFAIK, it must be supported by the client. Basically, the client and the server need to negotiate which parts of the file (likely defined as parts in "multipart/form-data" POST) have already been uploaded, and then the server code needs to be able to merge newly uploaded data with existing one.
The best solution is to have custom uploader code, usually implemented in Java though I think this may be possible in Flash as well. You might be even able to do this via JavaScript - see 2 sections with examples below
Here's an example of how Google did it with YouTube: http://code.google.com/apis/youtube/2.0/developers_guide_protocol_resumable_uploads.html
It uses "308 Resume Incomplete" HTTP response which sends range: bytes=0-408 header from the server to indicate what was already uploaded.
For additional ideas on the topic:
http://code.google.com/p/gears/wiki/ResumableHttpRequestsProposal
Someone implemented this using Google Gears on calient side and PHP on server side (the latter you can easily port to Perl)
http://michaelshadle.com/2008/11/26/updates-on-the-http-file-upload-front/
http://michaelshadle.com/2008/12/03/updates-on-the-http-file-upload-front-part-2/
It's a shame that your clients can't use ftp uploading, since this already includes abilities like that. There is also "chunked transfer encoding" in HTTP. I don't know what Perl modules might support it already.