Hi,
Quotas are much nicer from some points of view because they're much easier
to change.
We've already upped the atlas partition size once since other VO's weren't
using their allocations but it's a fair ammount of work (even with LVM).
We're currently running with a partition per VO on our current SE but when I
bring up the new one I want one quota'd partition.
Is anyone doing this with a hacked info provider?
Yours,
Chris.
> -----Original Message-----
> From: LHC Computer Grid - Rollout
> [mailto:[log in to unmask]] On Behalf Of Burke, S (Stephen)
> Sent: 13 January 2005 14:37
> To: [log in to unmask]
> Subject: Re: [LCG-ROLLOUT] Quotas ... Full SEs
>
> LHC Computer Grid - Rollout
> > [mailto:[log in to unmask]] On Behalf Of Dr D J
> Colling said:
> > Does anybody know of any reason why we should not give different
> > experiment user groups different quotas on our SE? For
> > example, are these
> > then reported correctly by the information system? Does
> > anybody already impose quotas?
>
> As things stand I think the information provider just does a df on the
> storage area to get the free space. You could probably hack
> it to report a
> quotaed value, but it's not obvious that anyone is paying
> attention to the
> free space value anyway and it might confuse the monitoring.
> Alternatively
> you can give each VO a separate disk or partition. (In theory
> the schema has
> some support for quotas, but it has never been implemented.)
>
> Stephen
>
|