HomeBUGS & ISSUES

Think you found a bug? Report it here.

v0.9.4197 Gives more timeout than v0.9.4189 Messages in this topic - RSS

Huong Nguyen
Huong Nguyen
Posts: 69


7/10/2011
Huong Nguyen
Huong Nguyen
Posts: 69
***Resolved in v0.9.4197***
 Hi Vadim,
 
I ran the same test, on the same computer, same custom script to same server, using v0.9.4189, got 0 errors, 0 timeout.
The same exact test, when ran using an upgraded SS v0.9.4197 gave me 0 errors, 467 timeout.
Can you look into this and let me know what other debugging I should perform?
I'm sending you some user iteration files and summary in another email.
0 link
Fabien Lennertz
Fabien Lennertz
Posts: 12


7/26/2011
Fabien Lennertz
Fabien Lennertz
Posts: 12
Hi Huong,

How long was the duration time of your tests? Deviations are possible, even under the same circumstances.  I remember that I got different results when I ran the same stress test (detecting whenever the system returned the first error) for the second time. It could be caused by a lot of reasons: duration time, type of test, use of external host etc.
 
Also when the duration time is too short you can get different figures. For example the average load time can differ when there are a few peaks and only got a few requests. So you can say that you have a test with volatile results. Increase the duration time to have less variance and to make your results more reliable.
 
Last but not least I noticed that whenever an error is produced, they grow proprotionally. So an error 'produces'  errors. Maybe this applies to your time outs as well?
 
Regards,
 
Fab
0 link
Huong Nguyen
Huong Nguyen
Posts: 69


7/26/2011
Huong Nguyen
Huong Nguyen
Posts: 69
Hi Fab,
 
I run all of my tests for a duration of 1 hour under the same environment, except for the SS upgrade. Vadim told me that the increase in timeout was due to a bug fix and that those timeouts are real timeouts that wasn't recorded in previous versions. This is confirmed by using the timeline in Fiddler.
0 link
Fabien Lennertz
Fabien Lennertz
Posts: 12


7/26/2011
Fabien Lennertz
Fabien Lennertz
Posts: 12
Hi Huong,
 
Ok, fair enough. Then consider my reply as a best practice for other users who struggle with deviations.
 
Cheers,
 
Fab
0 link
Huong Nguyen
Huong Nguyen
Posts: 69


7/26/2011
Huong Nguyen
Huong Nguyen
Posts: 69
smile. Thank you.
0 link






Copyright © 2017 Stimulus Technology