Sorry my mistake. I thought there were sites pointing to lcgbdii.gridpp.ac.uk which doesn't exist but they are actually pointing to lcgbdii.gridpp.rl.ac.uk I missed the rl. I could see that no-one had CERN as first on their list but I wrongly assumed that some had a non-existent site as first and CERN as second.
I understand the lcg-bdii.gridpp.ac.uk alias. I wasn't commenting on it.
About the suggestion below. CERN are asking that sites don't point to them so do we want to raise the issue with CERN and say we reject your request and insist on still using you? EGI are monitoring the availability of the combination RAL.OR.Imperial and it has been identically 100% for the few months they have been monitoring. I know some people get a warm and fuzzy feeling from relying on CERN (no-one ever got sacked for doing what CERN do:-) but one can understand that this isn't scalable.
John
> -----Original Message-----
> From: Testbed Support for GridPP member institutes [mailto:TB-
> [log in to unmask]] On Behalf Of J Coles
> Sent: 20 December 2011 11:51
> To: [log in to unmask]
> Subject: Re: Ops meeting today @ 11am
>
> Hi John
>
> I'll type up my notes later but the conclusion was:
>
> - No site is using CERN as the primary top-bdii
> - The issue about a typo: this is actually pointing to the GridPP alias
> to allow sites to add it and have another failover route. It was
> pointed out we do not have a good mechanism to communicate when the RAL
> bdii is down to ensure the alias is updated (if people use it as
> primary). We need to think about the order.
>
> The suggestion was: 1. RAL 2. Imperial 3. CERN.
>
> The meeting finished before I got to your new question so perhaps
> people will comment in response to these messages.
>
> Thanks,
> Jeremy
>
>
>
> On 20 Dec 2011, at 11:45, John Gordon wrote:
>
> > I see my issue on TL BDII was on the agenda. I'll wait for the
> minutes for feedback but a related issue has cropped up.
> >
> > EGI are planning to extend their availability/reliability calculation
> to include the BDII freshness test. You can see a comparison of the two
> measurements here
> https://www.egi.eu/indico/materialDisplay.py?contribId=5&materialId=0&c
> onfId=617
> >
> > No problem for most UK sites, miniscule effect for a few. In the
> light of this sites might consider whether they are running the right
> version of BDII - no I don't know what version this might be.
> >
> > John
> >
> >> -----Original Message-----
> >> From: Testbed Support for GridPP member institutes [mailto:TB-
> >> [log in to unmask]] On Behalf Of Jeremy Coles
> >> Sent: 20 December 2011 09:19
> >> To: [log in to unmask]
> >> Subject: Ops meeting today @ 11am
> >>
> >> Dear All
> >>
> >> There will be a short ops meeting today to cover a few things before
> >> everyone is away. We'll aim to finish by 11:30 - several people have
> >> indicated this would be useful.
> >>
> >> In case you are not attending let me thank you for a productive 2011
> >> and wish you a very Happy Christmas & New Year. A few people
> including
> >> myself will be watching things over the holiday period and will try
> to
> >> follow up any significant and urgent problems. We'll keep a list of
> >> more minor things to be followed up early next month.
> >>
> >> Best wishes,
> >> Jeremy
|