Breakdown of SRMv2 endpoints before starting T0-T1 exercise.

This page summarizes the results of the test perfomed across all our T1 SRMv2 endpoints. The test consists in moving real MDF file (1.6 GB of size) from the CERN endpoints (LHCb_RAW space) to the corresponding space defined on each T1 using voms proxy /lhcb/role=production. The test relies on lcg-cp (version used by Flavia on her S2 suite) and the lcg-del of the same suite. The test also check the information published and the Access Control Roles defined on each space. Aim of this test is to avoid (or at least minimize) bad surprises from various SRM endpoints...

Site	Copy 1.6GB	Delete	LHCb_RAW (pub)	Access Control Base Rule	GlueSAPath
CNAF	64	OK	yes	VO:lhcb	/castor/cnaf.infn.it/grid/lcg/lhcb
RAL	279	Temporary down	yes	VO:lhcb	/castor/ads.rl.ac.uk/prod/lhcb/lhcb
SARA	692	OK	yes	VOMS:/lhcb/Role=production	/pnfs/grid.sara.nl/data/lhcb
PIC	59	OK	yes	VOMS:/lhcb/Role=production	/pnfs/pic.es/data/lhcb/CCRC08/RAW
FZK	351	OK	yes	VO:lhcb	/pnfs/gridka.de/lhcb/raw
IN2P3	104	OK	yes	none	/pnfs/in2p3.fr/data/lhcb/data/CCRC08

Remarks:

February 11th

FZK problem for Andrew was that he was not in the gridmap file.
Lyon: I tested the correct production endpoint (being not advertised correctly the SAPath at Lyon) with the right path (going now to be published correctly in the BDII too). It also works fine (see new logs in the table)

February 8th

Confirmation that all endpoints are OK using role=production for user santinel. Andrew reports problem in moving data to IN2P3 and FZK.

RAL CASTOR is down after the recent power cut, SARA is flickering. When it works it is very slow (~2MB/sec). The rest of the sites is pretty much stable.

February 6th

PIC: was publishing this morning LHCB_RAW. They have redone the reservations for LHCb_ so that the Space Manager was pushing into the BDII the right token description (LHCb_RAW). Not it is OK
GRIDKA: the lcg-cp command was hanging this morning Doris found a configuration file overwritten their side. Now it is OK
SARA: a problem with the authentication this morning despite the test are ran with lhcb production credentials. Ron fixed the problem (I was mapped onto the regular lhcb account rather than the production user account). Also experienced unacceptable slowness in the transfer though.Ron is looking at that
IN2P3: are not publishing correctly the block of information concerning LHCb_RAW space. This might affect the FTS transfers. Sysadmins notified. Frederic and Lioned have modified the information published by the LHCb_RAW and not it is OK

Dynamically up-to-dated information are now available at the S2 LHCb dedicated page kindly provided by Flavia.

http://lxdev25.cern.ch/s2test/vo/s2_logs/index-lhcb.html

-- RobertoSantinel - 06 Feb 2008