Markus Schulz wrote:
>
> Dear Resource Center Staff, Core Infrastructure and Regional Operation
> Centers,
> here is the curve showing how we progress towards moving our sites to
> LCG-2_4_0.
>
> Day 0 was the release date (06/04/2005).
> Based on a linear model, we will have on Wednesday the 11th May at 18:00
> 90% of the sites moved to the new version.
> This is a bit too slow to meet the target date 27/04/2005 (15 work days
> after the release).
>
> markus "I hope it is more exponential"
> schulz
Hi Markus, as you probably know for Italian ROC we are installing a lightly
customized of LCG 2.4.0 to simplify the deployment of some additional features
(e.g. gridice running on WN to check WN daemons, AFS on WN, etc...).
As you know, when the candidate 2.4.0 was out we asked to partecipate to the
first round of tests before the official starting of deployment.
We are a bit late because we found some little bugs (lcmaps and lcas not well
configured for voms, wrong torque configuration with multiple vo per queue,
configuration of edg-fmon-cleanspool, double-publishing of GlueServiceURI for
GridICE, edg-job-submit fails when there are more VOMS servers) and they of
course slowed down the addition and test of our customizations. We didn't do
special stress tests, just basic ones). We reported them to savannah and we
got very fast and efficient feedbacks (they were fixed in last yaim release,
the only almost-open problem is the last one, discovered only recently, but
Maarten already proposed a workaround).
So my hope for the future is in a more distributed certification/preproduction
process; middleware is complex, with more sites involved in this activity it is
likely they will discover more bugs/wrong configurations before starting the
deployment on production.
Best regards,
--
Enrico Ferro Ist. Naz. di Fisica Nucleare - Padova
Address: via F. Marzolo, 8 - 35131 Padova - ITALY
Email: [log in to unmask] Phone: +39-0498277154
|