> -----Original Message-----
> From: Testbed Support for GridPP member institutes [mailto:TB-
>
>
> Any of the sysadmins out there reading this can comment if they feel
> their time is being wasted and the approach needs to change.
> Suggestions are always welcomed. Let's also see how things go today/
> tomorrow.
>
It's useful. There's value in it for both ATLAS and for the sites; we've
certainly made improvements to things (notably SE pool network links) in
response to the test behaviour, and it will probably inform our future
purchasing decisions (10GbE is rather further up the wish list than it
was before, for example).
I also suspect that these large tests are more effective at finding
problems
in the ATLAS infrastructure than smaller tests are, so there's probably
some
value there even if it were a complete waste of time from a site POV.
What worries me isn't wasting time on ATLAS tests - it's how badly LHCb
are going to fall over at switch-on given that they're apparently not
doing
anything similar.
Ewan
|