The tension is between putting in new fixes and testing and certifying
them. The CERN deployment people don't seem to have a procedure for
accepting and verifying patches.If they had that then just putting them
in the repository is, as you say, simple.
John
> -----Original Message-----
> From: Testbed Support for GridPP member institutes
> [mailto:[log in to unmask]] On Behalf Of Simon George
> Sent: 24 May 2005 11:55
> To: [log in to unmask]
> Subject: Re: Sites not publishing accounting data
>
> Hi John,
>
> Basically I agree with you. It's just frustrating me no end that there
> is such a simple way to address this but it isn't being done, at the
> cost of dozens of site admins having to do more work themselves. It's
> not just this one item which is probably easy to fix by hand. I would
> like to establish the general principle that I stated at the end of my
> last email and that I seem to keep banging on about at
> meetings. To get
> a working site, it shoudl not be necessary to do anything
> beyond what is
> written in the installation and (to be done) maintenance instructions.
> The amount of hot air we have expended on this is far more effort than
> it would have taken for someone to copy the new apel rpm into the lcg
> updates dir and run the necessary yum and apt commands to
> make it visible.
>
> Rant over, thanks to Jeremy for taking this up at the
> EGEE/LCG workshop.
>
> Anyway, in practice I can assure you that at RHUL we are not sitting
> back; we are really busy with other work, so the easier it is to have
> the right software installed and correctly configured, the sooner (and
> more often) we will be able to do it.
>
> Cheers,
> Simon
|