On 11/16/2011 01:26 PM, Jon Peatfield wrote:
>
> Thanks, the errors seem to have gone away.
:) Happy to help.
>
> <snip>
>
> How large do you suggest making the random variation?

I'd say about 30 minutes of randomness should be enough swing to keep it 
interesting while not being too all over the place
>
> We currently run the scripts at about 03:37 (GMT0BST), so I suppose 
> that may correspond to a busy time elsewhere in the world - should we 
> move it to a different time?
>
It's always busy somewhere.  I know from the mirror side we are 
currently on CDT which puts us 6 hours later than UTC with 08:00-18:00 
CDT being when we've got people at the lab.  However I'd say about 99.9% 
of the people here have no interaction at all with our mirrors (we have 
an internal one for them).  In the end our biggest (by numerics) mirrors 
are in the US, the UK, Germany, Japan, and Russia.  By the time it is 
night for one of these it is day for another.  And those are, of course, 
the public mirrors - who knows on the private ones.

All this to say, I've no idea when a good time is, pick what feels good 
and go with it.  If it isn't working a slight adjustment may be necessary.

Pat

-- 
Pat Riehecky
Scientific Linux Developer