I just figured out how to do it:
close the long queues, leave the short queues open til Sunday evening :)
however i also just realized that my ERT stuff is not going to report a
sensible ERT for a closed queue :( but perhaps that's OK, one should not
be inspecting an ERT for a closed queue.
J "kind of like checking if the brakes work on a car with no wheels" T
Burke, S (Stephen) wrote:
> LHC Computer Grid - Rollout
>
>>[mailto:[log in to unmask]] On Behalf Of Jeff Templon
>
> said:
>
>>The site is not down, it's just temporarily not accepting new
>>jobs. I
>>can understand that I need downtime while upgrading, because services
>>will truly be inaccessible, and any jobs still running at the time of
>>the upgrade will be lost.
>
>
> But the SFTs won't run so the monitoring will think there's a problem.
> Also it will affect accounting in some way, although what you want isn't
> so obvious: one way you are nominally available but your CPU usage is
> low, the other way you are nominally unavailable but you accumulate some
> usage anyway. The second one probably looks better :)
>
> Stephen
|