Reports were not received from [Alice, Atlas, Biomed]
1. One site quotes e.g. it is *really* bad practise to push the gLite 3.0 rollout ('you have two weeks') although it is actually not ready. See GGUS ticket 8722. The relocatable tarball is not usable for a WN installation by now. An update was promised but did not make it on the ftp server until now. (G-SW ROC)
2. the *minimal requirement* to take part in SC4 for Tier2s should be e.g. an LCG 2.7.0 CE rather than an gLite 3.0 lcg-flavoured one, as there is no fundamental difference. More generally speaking we would appreciate, if participation in SC4 could be handled in a more pragmatic way, based on services.
The question should *not* be (for Tier2s): is gLite 3.0 completely installed at the specific site *but* are all required services for SC4 for all supported VOs at that site available. (G-SW ROC)
3. With the release of gLite 3.0 there are two flavors of CE. The gLite one and the LCG one. However in the information system both CEs are shown with version GLITE-3_0_0. It would be desirable to distinguish between the two different versions of CEs in the Information system and not rely only to data entered in the GOC DB regarding CE version.
Associated Ticket: (https://gus.fzk.de/ws/overview.php?quali=open&resultlist=1&ticket=9220)
Answer form Oliver: They are already distinguished by their UniqueID - the LCG ones have 'jobmanager' and the gLite ones have 'blah' in their service strings;
ce:2119/blah-pbs-dteam
vs
ce:2119/jobmanager-lcgpbs-dteam
(SE ROC)
4. Our notes regarding gLite migration where circulated earlier today. Where shall we store them?
(SE ROC)
5. LHCb wants to adopt the same software installation procedure (via SFT) at CERN but they are prevented because CERN provides the installation area as an AFS shared area. They need to either roll back to a normal NFS shared area or to have a mechanism in place and working at CERN (e.g. gssklog) for installing sofwtare via grid jobs.
(LHCb)
6. LHCb need - with high priority - a new special QoS log SE (now lxb2003.cern.ch which has been proved to be non adequate for this purpose). Last week thousand of jobs failed trying to upload logs to this storage that was overloaded and all lcg-cp commands were hanging. Unknown the cause. Instead of having this single access point (BTW very old hardware equipped) they need a reliable machine (managed by FIO) with alarms, monitors and 24X7 babysitting. This service is dramatically important.
(LHCb)
7. Transfers-
Data transfer is the one area that is not progressing as we would like. Of the FTS channels from CERN to Tier-1 centers it appears only transfers to PIC are routinely succeeding. We have entered a large variety of GGUS tickets.
(CMS)
8. Analysis Submission: We have had better success, as a fraction of sites, with the Tier-2 centers than we have had with the Tier-1 centers. Currently IN2P3 and PIC are not functional with CMS VO software installation issues. FZK and RAL have been unable to download the test sample. The PhEDEx instance as a VOBOX or alternative setup does not appear to be working. CERN has not announced the existence of the data sample.
(CMS)
