I've been using the HTTP(S) Test Script Recorder in Jmeter to record some 
amazon S3 actions. Specifically a file upload (PUT). Problem is, the captured 
HTTP sampler has a body data that has some extra data inserted into it. 

My original sample file I uploaded is 25630 bytes. A small image file (jpg). 
The one that gets uploaded with the captured HTTP sampler is 25723 bytes. 




I performed a hex compare between the original and the one from the sampler. (I 
downloaded normally what the captured sampler uploaded). What I found 
immediately was a single byte inserted very near the top of the file, which had 
the effect of offsetting everything else after that by one byte and making it 
all look different. Since the difference in size is just under 100 bytes, I'm 
guessing this is happening in just under 100 other places in the file as well. 




Look at the different Content-Length headers below... 




Here are the headers from when I upload the file manually: 
User-Agent: S3 Browser 4-6-1 http://s3browser.com
x-amz-acl: private
x-amz-date: Thu, 23 Oct 2014 17:25:08 GMT
Content-Type: image/jpeg
Content-Length: 25630 


Here are the headers that Jmeter sends based on the captured sample: 
Connection: keep-alive
Content-Type: image/jpeg
x-amz-date: Thu, 23 Oct 2014 16:57:31 GMT
x-amz-acl: private
User-Agent: S3 Browser 4-6-1 http://s3browser.com
Content-Length: 25723 


Does anyone know why this is and how I might modify my recorder so that it 
records the file data exactly? 

Allen Granum 
[email protected] 


1309 Carling Avenue 
Ottawa, Ontario, Canada K1Z 7L3 
(613) 722-3232 
x263 

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to