Home > Archive > Software Testing > October 2005 > Stability tests
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
|
|
| hjorgensen27@yahoo.fr 2005-10-12, 8:08 am |
| In our company we run a set of scripts in a loop on the application for
72 hours - if the process is still running after that laps of time, we
say that the application passed the stability test.
My question is: do you know if this a standard way of doing it? Do you
know if the 72 hours is a reference or an arbitrary (but good) value
set by my company?
thanks
Hanne
| |
| xtremetester 2005-10-12, 7:03 pm |
|
hjorgensen27@yahoo.fr wrote:
> In our company we run a set of scripts in a loop on the application for
> 72 hours - if the process is still running after that laps of time, we
> say that the application passed the stability test.
>
> My question is: do you know if this a standard way of doing it? Do you
> know if the 72 hours is a reference or an arbitrary (but good) value
> set by my company?
>
> thanks
> Hanne
What I know - that IBM is running the stability test for DB2 during a
w .
It all depends of application.
Regards
Alex
http://www.geocities.com/xtremetesting/
| |
| gregmg@my-deja.com 2005-10-12, 7:03 pm |
| What you are describing as stability testing is often called load
testing. In my experience, these types of tests are frequently executed
for 48 to 72 hours, but I've never seen a duration specified in any
texts on the subject.
I suggest rather than looking for any standard test duration, instead
consider how your application will be used. Do you expect it to be
under extremely heavy use for days or ws at a time? If so, 72 hours
may be too short. If you expect it to be under heavy use in short
bursts, then a 72 hour test is probably fine.
You might also want to consider running a stress test. While load
testing involves testing with a heavy, but reasonable system load,
stress testing is an attempt at breaking the system. Basically, the
goal is to put so much of a load on the system or application that it
will fail, and you can then confirm that it fails gracefully.
Greg G.
http://www.testengineering.info
| |
| dumitru.corobceanu@gmail.com 2005-10-20, 7:58 am |
| Everything is depending on application U're testing. It may be enough
3hours and could not be enough 3 ws. In my company a bug was
discovered after 4 ws of working (not under test, normal work). And
I suggest not start the application and wait for it to crush. But look
on what it's doing. Is it increasing memory consume, does the log
file grows to fast (bug: a log file in 30 seconds become 20GB size),
what happens with the DB, with process that it starts(maybe app. Starts
processes and 'forget' to close them).
|
|
|
|
|