[cvsnt] Stress Tests results for CVSNT/CVS/Subversion

[Sunil Guzman]

This has been a very interesting thread. What follows is my analysis of some of the highlights of this discussion. Sadly, I have concluded that I can not recommend CVSNT to my managers, since the quality of the product matters at least as much as the feature-set, if not more. At this point, we are trying to decide between CVS and Subversion.
  From Tony’s response:
-----
I've run successful tests with over 1000 clients (the OS gets real slow 
under that kind of load so the cvs processes end up not being the 
bottleneck)...  50 is not a stress test.  50 what happens when an office 
starts up in the morning and everyone runs cvs update... it's routine 
behaviour.
----
However, when Rahul asked for the CVSNT version that was subjected to this kind of stress, Tony’s response was:
----
That was probably around the 2.5.01 era... I can't imagine it's got much 
slower since although I could fairly easily setup the test again.
----
Hmm – so, releases are not routinely qualified with any stress-testing?!
  Arthur Barrett, in a rather defensive response to Rahul’s post, said:
----
Of significance to me is that build 2221 is one we rejected to send to
commercial support customers, though not specifically for that
(performance testing) reason.
----
Are builds that you “rejected to send to commercial support customers” routinely foisted on non-paying customers? What is the intent of this process? To destabilize installations at non-paying customer sites and “encourage” them to become “commercial support customers”?


			
---------------------------------
 Yahoo! Mail
 Use Photomail to share photos without annoying attachments.