ADC Operation NL
Revision as of 10:12, 7 November 2008 by Hclee@nikhef.nl (talk | contribs) (→ATLAS SRMv2 space tokens deployed in NL cloud)
Introduction
This page is to summarize the current setup/configuration in the NL cloud for supporting the ATLAS Distributed Computing (ADC).
Sites
-
combined Tier1 center:
| Country | Institute | GOC site name |
|---|---|---|
| Netherlands | NIKHEF | NIKHEF-ELPROD |
| Netherlands | SARA | SARA-MATRIX |
| Country | Institute | GOC site name |
|---|---|---|
| Russia | ITEP | ITEP |
| Russia | IHEP | RU-PROTVINO-IHEP |
| Russia | JINR | JINR-LCG2 |
| Russia | PNPI | RU-PNPI |
| Russia | SINP | SINP |
| Russia | RRC-KI | RRC-KI |
| Ireland | CST | CSTCDIE |
| Turkey | ULAKBIM | TR-10-ULAKBIM |
Data Management Services
- Storage Resource Manager v2.2: deployed on each site as common interface to storage elements
- gLite File Transfer Service (FTS): deployed at SARA serving the T1-T1 and T1-T2 data transfers to the sites of the NL cloud
- LCG File Catalog Service (LFC): deployed at SARA for cataloging grid files stored on the sites of the NL cloud
ATLAS defined SRMv2 space tokens
The following table is taken from Stephane's talk on ATLAS Computing & Software Week, November 2008 showing the SRMv2 space tokens needed for ADC operations. More up-to-date information can be found at the DDM operations twiki.
ATLAS SRMv2 space tokens deployed in NL cloud
| Institute | Role | Tokens |
|---|---|---|
| NIKHEF | T1 & T2 | NIKHEF-ELPROD_DATADISK |
| NIKHEF-ELPROD_GROUPDISK | ||
| NIKHEF-ELPROD_ROFVAL | ||
| NIKHEF-ELPROD_USERDISK | ||
| NIKHEF-ELPROD_LOCALGROUPDISK | ||
| SARA | T1 | SARA-MATRIX_DATATAYPE |
| SARA-MATRIX_DATADISK | ||
| SARA-MATRIX_MCTAPE | ||
| SARA-MATRIX_MCDISK |
Data locations among SARA and NIKHEF
Monitoring pages
- NL cloud storage usage summary
- Data transfers (T0-T1, T1-T1) in last 24 hrs
- Data transfers (T1-T2) in last 24 hrs
- active PanDA tasks in NL cloud
- specific storage usage of essential space tokens at NL T1 (require CERN NICE account)
More information
- ADC eLog entries concerning NL cloud
- DDM operation wiki
- DDM browser
Daily operation logs
| Date | Actions | Remarks |
|---|---|---|
| 9 Oct. 2008 | requests cosmic and 1beam reprocessed ESD to NIKHEF-ELPROD_DATADISK | needed by NIKHEF physics group |
| 10 Oct. 2008 | NL Tier2s become on-line again for MC production | FTS performance issue at SARA fixed |
Trouble shooting logs
Sept. 2008 - huge transfer backlog from T2s to SARA
Problem fixed by running the FTS admin tool to slim down the FTS job history db table.
Sept. 2008 - SRM request timeout reading data from SARA
A cron job fixing orphan file issue of dCache loads PNFS server so it was stopped. Also observe a broken network switch involving 4 new dCache node. The problematic network switch has been replaced. A broken dCache node is also replaced.