WLCG-OSG-EGEE Operations meeting

Monday 17 Jul 2006, 14:00 → 17:30 Europe/Zurich

28-R-15 (VRVS (Saturn room))

28-R-15

VRVS (Saturn room)

Maite Barroso

Description

VRVS "Saturn" room will be available 15:30 until 18:00 CET

- 14:00 → 17:25
  28-R-15
  
  28-R-15
  - 16:00
    
    Feedback on last meeting's minutes 5m
    
    Minutes
  - 16:05
    
    Grid-Operator-on-Duty handover 5m
  - From Russia (backup: CERN) to UK/Ireland (backup: Taiwan)

16:10

SC4 weekly report and upcoming activities 10m

See new and updated information at https://twiki.cern.ch/twiki/bin/view/LCG/SC4ExperimentPlans

Speaker: Harry Renshall

16:20

Security update 5m

Speaker: Ian Neilson

16:25

convention for naming SRM endpoints 5m

Site specific info should go before the VO directory path. The VO namespace should not be touched i.e. we shouldn't have (for example)
srm://site /path/<VO name>/barney
srm://site /path/<VO name>/rubble
but
srm://site /path/barney/<VO name>
srm://site /path/rubble/<VO name>
anything that comes after <VO name> should be specified by the VO. This would allow a VO to construct in a simple manner a SURL just by appending a LFN if they so desired.

16:30

Classic SE to DPM/dCache migration 5m

We are planning to stop the support for the Classic SE around autumn. The main reason for this is the lack of ACL's/VOMS support. The recommended strategy is to migrate to the Disk Pool Manager (DPM). The migration has been tested and already performed by some sites, and the needed migrating scripts are provided (see attached mail for more details). ------------------------------------------------------ For all sites running a Classic SE: We would like to inform you that the Disk Pool Manager (DPM) offers a good replacement solution for disk space storage. Indeed, on top of the Classic SE features, the DPM provides : - Logical Namespace, - Authorization and ACLs, - Manageable storage (easy to add/remove disk space), - Automatic garbage collection. The DPM also supports the SRM protocol, that is required by many VOs. More details about the DPM, in the administration point of view, are given in this presentation : http://indico.cern.ch/contributionDisplay.py?contribId=3&sessionId=s0&confId=a058483 When migrating from a Classic SE to the DPM, the existing physical files don't need to be moved. Only a metadata operation is required : the exisiting files have to be registered in the DPM Name Server. Of course, we provide a script to do this automatically. The migration procedure is describe here: https://twiki.cern.ch/twiki/bin/view/LCG/ClassicSeToDpm

Speaker: Markus

16:35

Issues to discuss from reports 25m

Reports were not received from [Tier-1s: BNL; NDGF; NIKHEF/SARA; PIC ]

1. What is the status of the OPS VO? When will we finally switch? (SouthEasternEurope)
The current statistics is the following (in number of CEs):
OK: 87
Job list-match (probably not supporting OPS at all): 39 Errors
(job submission or other critical tests): 63 Scheduled downtime: 20
You can see it here: https://lcg-sft.cern.ch/sft/lastreport.cgi?vo=ops Piotr

2. What is the difference between PPS and Production Operation Procedure (failing SFT, reaction time, relevance, suspension etc.) This was an open question at the COD meeting, because it is not clear if PPS sites should be handled similar to Production sites but with less priority and later deadlines. (SouthWesternEurope)

3. The official glite 3 updated versión in production is 3_0_1? Because it looks like that after doing the upgrade from the official repositiry http://glitesoft.cern.ch/EGEE/gLite/APT/R3.0/rhel30/RPMS.updates/ the sites appears at SFT to have versión 3.0.1 installed. We had this question at pic because we thought that 3.0.1 was only for preproduction. (SouthWesternEurope)

4. The main point at this moment are the T0-T1 transfers which are not properlyworking at this moment. (Alice) CERN-CNAF: They have been working for a while, butnow some problems (most probably because of the proxy inside the VOBOX) arebeing ovserved CERN-FZK: The AlieN SE has to be defined as SRM CERN-NIKHEF:Transfers failing, still to investigate the reason. CERN-RAL: No access todayto the VOBOX CERN-IN2P3: The AlieN SE has to be defined as SRM

5. lcg-cp doesn't support SURL as destination (LHCb) lcg-gt doesn't support list ofprotocols lcg-gt doesn't return ROOT compatible TURL and more in general different protocols (dcap,rfio gsidcap, castor?) do require different stringmangling. Using the gfal_plugin there wouldn't be any problem. The issueconsists to get to get a version of ROOT that can support all protocol in theGFAL plugin. This needs LCG AA release with the current TUNNEL library. On itsway hopefully.. lcg-cr doesn't allow user to copy and register a file byspecifying some other string for the storage_host field than the LCGone.

6. There's the carry on issue with the discussion of SRM endpoint. Ithink the major issue here is that VO namespace should be sacrosanct (LHCb)

7. The centres should bereminded a change in SRM endpoint name is major issue for the file catalogs. Soshouldn't be undertaken lightly & without necessary fore warning. (LHCb)

8. Thereseems to be continuing issues with CASTOR at CERN. (LHCb)

17:00

Review of action items 15m

17:20

AOB 5m

Choose timezone

WLCG-OSG-EGEE Operations meeting

28-R-15

VRVS (Saturn room)

28-R-15