Introduction to Storage Element HsinWei Wu Academia Sinica

Introduction to Storage Element Hsin-Wei Wu Academia Sinica Grid Computing Center, Taiwan

Outline � SRM � A union form standard for SEs � Storage Elements � Be. St. Man � Castor � d. Cash � DPM � Sto. RM � LFC (LHC File Catalog) � Demonstration � Data Management Commands 2

Overview of g. Lite Middleware CLI API Access Information & Monitoring Authorization Auditing Network Monitoring Authentication Security Services Metadata Catalog Storage Element File & Replica Catalog Data Movement Service Discovering Accounting Data Services Information & Monitoring Services Job Provenance Package Manager Computing Element Workload Management Job Mgmt. Services 3

Data service in g. Ltie � Storage(SE): The files are physical located � Storage URL or SURL � srm: //as-ds 01. euasiagrid. org/dpm/euasiagrid. org/data � Catalogs: High level hierarchical namespace, maps the “physical” files to a virtual “logical” filename � Logical File Name or LFN � lfn: /grid/twgrid/data � Movement: put/get files into grid SE's, move/replicate files between SE's � File Transfer Service or FTS � Transport URL or TRUL gsiftp: //sc 003. grid. sinica. edu. tw: 2811/data 4 Ref g. Lite-3. 2 User Guide

Storage Resource Management A Storage Element (SE) is an aggregate of Grid services that allows Grid users to store and manage files together with the space assigned to them � Data are stored on disk pool servers or Mass Storage Systems � Storage resource management needs to take into account � � � � Transparent access to files (migration to/from disk pool) File pinning Space reservation File status notification Life time management The SRM (Storage Resource Manager) takes care of all these details � The SRM is a single interface that takes care of local storage interaction and provides a Grid interface to the outside world 5

Berkeley Storage Manager (Be. St. Man) LBNL � � � • • • Java implementation Designed to work with unixbased disk systems As well as MSS to stage/archive from/to its own disk (currently HPSS) Adaptable to other file systems and storages (e. g. NCAR MSS, VU L-Store, TTU Lustre, NERSC GFS) Uses in-memory database (Berkeley. DB) Multiple transfer protocols Space reservation • Local Policy • Fair request processing • File replacement in disk • Garbage collection Directory management (no ACLs) Can copy files from/to remote SRMs or Grid. FTP Servers Can copy entire directory recursively • Large scale data movement of thousands of files • Recovers from transient failures (e. g. MSS maintenance, network down)

Castor-SRM CERN and Rutherford Appleton Laboratory � � CASTOR is the HSM in production at CERN Support for multiple tape robots � � Support for Disk-only storage recently added Designed to meet Large Hadron Collider Computing requirements � Maximize throughput from clients to tape (e. g. LHC experiments data taking) • • C++ Implementation Reuse of CASTOR software infrastructure • Derived SRM specific classes • • • Configurable number of thread pools for both front- and backends ORACLE centric Front and back ends can be distributed on multiple hosts

d. Cache-SRM FNAL and DESY � � � Strict name space and data storage separation Automatic file replication based on access patterns HSM Connectivity (Enstore, OSM, TSM, HPSS, DMF) Automated HSM migration and restore Scales to Peta-byte range on 1000’s of disks Supported protocols: � (gsi/krb)FTP, (gsi/krb)d. Cap, x. Root, NFS 2/3 � � � Separate I/O queues per protocol Resilient dataset management Command line and graphical admin interface Variety of Authorization mechanisms including VOMS Deployed in a large number of institutions worldwide • • • Support SRM 1. 1 and SRM 2. 2 Dynamic Space Management Request queuing and scheduling Load balancing Robust replication using srm. Copy functionality via SRM, (gsi)FTP and http protocols

� � � Disk Pool Manager (DPM) CERN Provide a reliable, secure and robust storage system Manages storage on disks only Security � � � � � GSI for authentication VOMS for authorization Standard POSIX permissions + ACLs based on user’s DN and VOMS roles Virtual ids � Accounts created on the fly � � Garbage collector Replication of hot files � � Grid. FTP (v 1 and v 2) Secure RFIO https Xroot Full SRMv 2. 2 implementation Standard disk pool manager capabilities Transfer protocols Works on Linux 32/64 bits machines Direct data transfer from/to disk server (no bottleneck) Support DICOM backend Requirement from Biomed VO Storage of encrypted files in DPM on the fly + local decryption � Use of GFAL/srm to get TURLs and decrypt the file � � • Supported database backends • My. SQL • Oracle • High availability • All servers can be load balanced (except the DPM one) • Resilient: all states are kept in the DB at all times

Storage Resource Manager (Sto. RM) INFN/CNAF - ICTP/EGRID � It's designed to leverage the advantages of high performing parallel file systems in Grid. � Different file systems supported through a driver mechanism: generic POSIX FS � GPFS � Lustre � XFS It provides the capability to perform local and secure access to storage resources (file: // access protocol + ACLs on data). � � Sto. RM architecture: • Frontends: Frontends C/C++ based, expose the SRM interface • • Backends: Backends Java based, execute SRM requests. • Each component can be replicated and instantiated on a dedicated machine. DB: DB based on My. SQL DBMS, stores requests data and Sto. RM metadata.

LHC File Catalog � � Logical File Name (LFN): user file name � � in VO namespace, aliases supported Glbally Unique IDentifier (GUID) � � LFC maps LFNs to SURLs LFN 1 LFN 2 unique string assigned by the system to the file Site URL (SURL): identifies a replica A Storage Element and the logical name of the file inside it GUID SURL 1 SURL 2 ACL GSI security: ACLs (based on VOMS) � � To each VOMS group/role corresponds a virtual group identifier Support for secondary groups Web Service query interface: Data Location Interface (DLI) � Hierarchical Namespace � Supports sessions and bulk operations � lfc-ls –l /grid/vo/ lfc-getacl /grid/vo/data 5 th NRENs and Grids Workshop, Paris - 11 -12 June 2007 LFC DLI /grid /vo /data file 11

Data Management Command SRM LFC DPM list lcg-ls SRM lfc-ls LFN dpns-ls DPNS upload file lcg-cp LOCAL SRM globus-url-copy LOCAL SRM lcg-cr -d SE -l LFN LOCAL rfcp LOCAL DPNS download file lcg-cp SRM LOCAL globus-url-copy SRM LOCAL lcg-cp LFN LOCAL rfcp DPNS LOCAL delete file lcg-del --nolfc SRM lcg-del -a LFN rfrm DPNS LFC_HOST LFC_HOME DPM_HOST DPNS_HOST DPMS_HOME environment variable Ref g. Lite-3. 2 User Guide 12

DPM Demonstration � Initial Proxy � Check DPM_HOST and DPNS_HOST 13 Ref g. Lite-3. 2 User Guide

DPM Demonstration � dpns-ls � Upload file to SE 14 Ref g. Lite-3. 2 User Guide

DPM Demonstration � Remove file 15 Ref g. Lite-3. 2 User Guide

SRM Demonstration � Initial Proxy � Check SE 16 Ref g. Lite-3. 2 User Guide

SRM Demonstration � Ls � Upload fire to SE 17 Ref g. Lite-3. 2 User Guide

SRM Demonstration � Remove file 18 Ref g. Lite-3. 2 User Guide

Thank you for your listening and patience 19 Ref g. Lite-3. 2 User Guide