Introduction to the current EDG Testbed Software Kraków, December 2002 Steve Fisher s. m. fisher@rl. ac. uk – RAL on behalf of The European Data. Grid Project Team http: //www. edg. org/
The European Data. Grid u Funded n by the European Union Jan 1, 2001 - Dec 31, 2003 u Develop, implement and exploit a large-scale data and CPU-oriented computational GRID. u Develop middleware, in collaboration with some of the leading centres of competence in GRID technology. u Complement, and help to coordinate at a European level, several ongoing national GRID projects. The EDG Testbed Software - 2
The EDG Main Partners Ø CERN – International (Switzerland/France) Ø CNRS - France Ø ESA/ESRIN – International (Italy) Ø INFN - Italy Ø NIKHEF – The Netherlands Ø PPARC - UK The EDG Testbed Software - 3
EDG Assistant Partners Industrial Partners • Datamat (Italy) • IBM-UK (UK) • CS-SI (France) Research and Academic Institutes • CESNET (Czech Republic) • Commissariat à l'énergie atomique (CEA) – France • Computer and Automation Research Institute, Hungarian Academy of Sciences (MTA SZTAKI) • Consiglio Nazionale delle Ricerche (Italy) • Helsinki Institute of Physics – Finland • Institut de Fisica d'Altes Energies (IFAE) - Spain • Istituto Trentino di Cultura (IRST) – Italy • Konrad-Zuse-Zentrum für Informationstechnik Berlin - Germany • Royal Netherlands Meteorological Institute (KNMI) • Ruprecht-Karls-Universität Heidelberg - Germany • Stichting Academisch Rekencentrum Amsterdam (SARA) – Netherlands • Swedish Research Council - Sweden The EDG Testbed Software - 4
EDG structure: work packages u WP 1: Work Load Management System u WP 2: Data Management u WP 3: Information and Monitoring u WP 4: Fabric Management u WP 5: Storage Element u WP 6: Testbed and demonstrators u WP 7: Network Monitoring u WP 8: High Energy Physics u WP 9: Earth Observation u WP 10: Biology u WP 11: Dissemination u WP 12: } Applications Management The EDG Testbed Software - 5
Current EDG Testbed Nordu. Grid Testbed 1 EDG sites Reference site: CERN Manchester NIKHEF RAL Lyon Lisboa Nordu. Grid: • Bergen • Copenhagen • Helsinki • Lund • Oslo • Stockholm • Uppsala Karlsruhe CERN Barcelona Madrid Italy: • Bologna • Cagliari • Catania • Milano • Napoli • Padova • Parma • Pisa • Roma • Torino The EDG Testbed Software - 6
Security: Authentication/Authorization u Authentication n Who you are s u users identified by certificates signed by a CA Authorization n What you are allowed to do s based on membership of Virtual Organizations (VO). The EDG Testbed Software - 7
Certificate Request grid-cert-request user cert-request once in every two. VO three years The EDG Testbed Software - 8
Requesting a Certificate u grid-cert-request A certificate request and private key is being created. [. . . ] Using configuration from /usr/local/grid/globus/etc/globus-user-ssleay. conf Generating a 1024 bit RSA private key [. . . ] A private key and a certificate request has been generated with the subject: /O=Grid/O=CERN/OU=cern. ch/CN=Akos Frohner [. . . ] Your private key is stored in. . . /. globus/userkey. pem Your request is stored in. . . /. globus/usercert_request. pem Please e-mail the certificate request to the CERN CA cat. . . /. globus/usercert_request. pem | mail cern-globus-ca@cern. ch Your certificate will be mailed to you within two working days. The EDG Testbed Software - 9
Certificate Signing CA grid-cert-request user cert signing cert-request certificate The EDG Testbed Software - 10
Registration/Authorization User registration in an EDG Virtual Organisation u convert n your certificate: openssl pkcs 12 –export –in ~/. globus/usercert. pem –inkey ~/. globus/userkey. pem –out user. p 12 –name ’Joe Smith’ u import your certificate in your browser u sign the usage guidelines: https: //marianne. in 2 p 3. fr/cgi-bin/datagrid/register/account. pl u ask an account from your VO administrator by email -> You are registered in the VO server and have a user account. The EDG Testbed Software - 11
Registration user certificate cert. pkcs 12 Account Registration convert registration VO once for the lifetime of the VO – you may change the certificate keys! Usage guidelines The EDG Testbed Software - 12
Starting a Session user certificate cert. pkcs 12 proxy-cert grid-proxy-init every 12/24 hours The EDG Testbed Software - 13
Usage You must have a valid certificate from a trusted CA! u “login”: grid-proxy-init short lifetime certificate: 24 hours Enter PEM pass phrase: . . . . +++++ u checking the proxy: grid-proxy-info -subject /O=Grid/O=CERN/OU=cern. ch/CN=Akos Frohner/CN=proxy -> use the grid services u “logout”: grid-proxy-destroy The EDG Testbed Software - 14
Configuration on the Server CA cert signing grid-cert-request service cert/crl update host-request host-cert ca-certificate crl automatically updated periodically The EDG Testbed Software - 15
Authorization Information service host-cert ca-certificates crls VO-server mkgridmap automatically updated periodically The EDG Testbed Software - 16
Using a Service service user host-certificate ca-certificates cert. pkcs 12 crls gridmap proxy-cert grid-proxy-init host/proxy certs exchanged The EDG Testbed Software - 17
EDG Logical Machine Types 1. User Interface (UI) 4. Computing Element (CE) n Gatekeeper (Front-end Node) n 2. Resource Broker (RB) 3. Information Service (IS) Worker Nodes (WN) 5. Storage Element (SE) 6. Replica Catalog (RC) The EDG Testbed Software - 18
Information Systems overview u. The aim of the Information and Monitoring Service is to deliver a flexible infrastructure that provides information on n the EU Data. Grid itself n grid applications EDG info systems are based upon Globus MDS (Metacomputing Directory Service or Monitoring and Discovery Service as it is now called) u n Based on Open. LDAP, a hierarchical database u. The information system is currently used mainly by the middleware. n You can use it to find out what is going on The EDG Testbed Software - 19
LDAP attributes u. A schema describes the attributes and the types of the attributes associated with data objects u Example - some attributes of Site. Info: n site. Name: RALDEV n sys. Admin. Contact: grid. sysadmin@rl. ac. uk n user. Support. Contact: grid. support@rl. ac. uk n site. Security. Contact: grid. security@rl. ac. uk n data. Grid. Version: 1. 2 n Installation. Date: 20020704142800 Z The EDG Testbed Software - 20
LDAP hierarchy u Lightweight Directory Assess Protocol (LDAP) offers a hierarchical view of information u The objects are arranged in a Directory Information Tree (DIT) u One or more attributes represent the Relative Distinguished Name (RDN) u An n object is identified by its Distinguished name This is its RDN with the Distinguished name of its parent The EDG Testbed Software - 21
RDNs and DNs site u RDN u DN n Site s n SE SE s n n se. Id=dev 02. hepgrid. clrc. ac. uk Protocols s s se. Id=dev 02. hepgrid. clrc. ac. uk, Mds. Vo-name=ral-dev, Mds-Voname=uk, o=Grid Protocols se. Protocol=rfio s supported protocols se. Protocol=gridftp SE s n Mds-Vo-name=ral-dev, Mds-Voname=uk, o=Grid se. Protocol=file s se. Protocol=gridftp, se. Id=dev 02. hepgrid. clrc. ac. uk, Mds. Vo-name=ral-dev, Mds-Voname=uk, o=Grid se. Protocol=rfio, se. Id=dev 02. hepgrid. clrc. ac. uk, Mds. Vo-name=ral-dev, Mds-Voname=uk, o=Grid se. Protocol=file, se. Id=dev 02. hepgrid. clrc. ac. uk, Mds. Vo-name=ral-dev, Mds-Voname=uk, o=Grid The EDG Testbed Software - 22
MDS GRISs & GIISs u Information providers are scripts which when invoked by the LDAP server make available the desired information n u u Within MDS the EDG information providers are invoked by a local LDAP server, the Grid Resource Information Server (GRIS) “Aggregate directories”, Grid Information Index Servers (GIIS), are used to group resources The GRISs use soft state registration to register with one or more GIISs The GIIS can then act as a single point of contact for a number of resources n u Information is cached by the server to improve performance A GIIS may represent a site, country, virtual organization, etc. In turn a GIIS may register with another GIIS The EDG Testbed Software - 23
EDG Information Providers & the Directory Information Tree site computing element storage element site information network information between this and other sites status storage elements that are close (not necessarily at the same site) supported protocols file statistics The EDG Testbed Software - 24
EDG GRIS/GIIS Hierarchy u There is a top level datagrid GIIS to which all of the country GIISs register datagrid u country. A country. B u site. A site. B site. C site. D u information providers Each country has a GIIS to which all of the site GIISs register Each Site has a Grid Information Index Server (GIIS) which acts as a single point of contact for all of the sites resources. The GRISs register with their site GIIS Information providers publish information to a local LDAP server known as a Grid Resource Information Server (GRIS) The EDG Testbed Software - 25
EDG Information Providers u The EDG have produced information providers: n Site information n The Computing Element n The Storage Element n Network Monitoring u All of the EDG data objects are dynamic, they have a time stamp and a time to live (used by the cache mechanism) associated with them The EDG Testbed Software - 26
Siteinfo in=siteinfo, Mds-Vo-name=ral-dev, Mds-Vo-name=uk, o=Grid object. Class: Site. Info object. Class: Data. Grid. Top object. Class: Dynamic. Object site. Name: RALDEV sys. Admin. Contact: grid. sysadmin@rl. ac. uk user. Support. Contact: grid. support@rl. ac. uk site. Security. Contact: grid. security@rl. ac. uk data. Grid. Version: 1. 2 installation. Date: 20020704142800 Z The EDG Testbed Software - 27
Computing Element ce. Id=dev 01. hepgrid. clrc. ac. uk: 2119/jobmanager-pbs -M, hn=dev 01. hepgrid. clrc. ac. uk, Mds-Voname=ral-dev, Mds-Vo-name=uk, o=Grid object. Class: Data. Grid. Top object. Class: Computing. Element CEId: dev 01. hepgrid. clrc. ac. uk: 2119/jobmanager-pbs -M Globus. Resource. Contact. String: dev 01. hepgrid. clrc. ac. uk: 2119/jobmanagerpbs: /O=Grid/O=UKHEP/CN=dev 01. hepgrid. clrc. a c. uk GRAMVersion: ? Architecture: intel Op. Sys: RH 6. 2 Min. Physical. Memory: 258 Min. Local. Disk. Space: 2048 Total. CPUs: 1 Free. CPUs: 1 Num. SMPs: 0 Min. SPUProcessors: 0 Max. SPUProcessors: 0 Total. Jobs: 0 Running. Jobs: 0 Idle. Jobs: 0 Max. Total. Jobs: 1 Max. Running. Jobs: 1 Worst. Traversal. Time: 108000 Estimated. Traversal. Time: 0 Active: TRUE Priority: 20 Max. CPUTime: 108000 Max. Wall. Clock. Time: 432000 Average. SI 00: 300 Min. SI 00: 300 Max. SI 00: 300 Authorized. User: /O=Grid/O=UKHEP/OU=hepgrid. cl rc. ac. uk/CN=Tim Eves Authorized. User: /O=Grid/O=UKHEP/OU=hepgrid. cl rc. ac. uk/CN=Tim Folkes Run. Time. Environment: RALDEV AFSAvailable: FALSE Outbound. IP: TRUE Inbound. IP: FALSE Queue. Name: M LRMSType: PBS LRMSVersion: Open. PBS_2. 3 The EDG Testbed Software - 28
Querying the Information & Monitoring Service u Queries can be posed to the current Information and Monitoring Service using LDAP search commands u An LDAP search consists of the following components $ldapsearch -x -H ldap: //lxshare 0225. cern. ch: 2135 -b 'Mds-Vo-name=datagrid, o=grid 'objectclass=Computing. Elment‘ CEId Free. CPUs -s base|one|sub “simple” authentication uniform resource identifier base distinguished name for search filter attributes to be returned scope of the search specifying just the base object, one-level or the complete subtree The EDG Testbed Software - 29
Querying the GRIS/GIIS Hierarchy u Mds-Vo-name=datagrid, o=grid n u Mds-Vo-name =datagrid Mds-Vo-name=site. B, Mds-Voname=country. A, Mds-Voname=datagrid, o=grid n u Mds-Vo-name =country. B u Mds-Vo-name =site. B Mds-Vo-name =site. C This will look at all the data from site. B Mds-Vo-name=site. B, o=grid n Mds-Vo-name =site. A This will look at all the data from country. A Mds-Vo-name=site. B, Mds-Voname=country. A, o=grid n u This will look at all the data from site. B Mds-Vo-name=country. A, o=grid n Mds-Vo-name =country. A This will look at all the data from site. B Mds-Vo-name =site. D The EDG Testbed Software - 30
The EDG WMS u u u The user interacts with GRID via a Workload Management System The Goal of WMS is the distributed scheduling and resource management in a GRID environment. What does it allow GRID users to do? To submit their jobs To execute them To get information about their status To retrieve their output u The WMS tries to optimize the usage of resources The EDG Testbed Software - 31
WMS Components u WMS is currently composed of the following parts: 1. User Interface (UI) : access point for the user to the GRID 2. Resource Broker (RB) : the broker of GRID resources, performing the match-making 3. Job Submission System (JSS) : provides a reliable submission system 4. Information Index (II) : a specialized Globus GIIS (LDAP server) used by the Resource Broker as a filter to the information service (IS) to select resources 5. Logging and Bookkeeping services (LB) : store Job Info available for users to query The EDG Testbed Software - 32
WMS UI Commands u dg-job-submits a job u dg-job-list-match lists resources matching a job description u dg-job-cancels a given job u dg-job-status displays the status of the job (submitted, waiting, ready, scheduled, running, outputready, aborted, cleared) u chkpt, done, dg-job-get-output returns the job-output to the user u dg-job-get-logging-info displays logging information about submitted jobs u dg-job-id-info is a utility for the user to display job info in a formatted style The EDG Testbed Software - 33
Example of UI Command Options u dg-job-submit –r <res_id> –n <user e-mail address> -c <config file> -o <output file> <job. jdl> -r the job is submitted by the RB directly to the computing element identified by <res_id> -n an e-mail message containing basic information regarding the job (status and identification) is sent to the specified <e-mail address> when the job enters one of the following status: DONE or ABORTED READY RUNNING -c the configuration file <config file> is pointed by the UI instead of the standard configuration file -o the generated dg_job. Id is written in the <output file> u dg-job-status –i <input file> (or dg_job. Id) -i the bookkeeping information about dg_job. Id contained in the <input file> are displayed The EDG Testbed Software - 34
Job Description Language (JDL) u Mandatory for every single JDL file: • • Executable (contains the command name) Other attributes: • Input. Sandbox • Output. Sandbox u Mandatory for JDL file dealing with Data Management: • Replica. Catalog (contains the Replica Catalog Identifier) • Data. Access. Protocol (contains the protocol or the list of protocols which the application is able to speak with for accessing Input. Data on a given SE) If Input. Data contains at least one PFN and no LFNs, only Data. Access. Protocol is mandatory. If Input. Data contains at least one LFN, both Replica. Catalog and Data. Access. Protocol are mandatory. The EDG Testbed Software - 35
Example JDL File Executable = “grid. Test”; Input. Data = “LF: testbed 0 -00019”; Replica. Catalog = “ldap: //sunlab 2 g. cnaf. infn. it: 2010/ rc=WP 2 INFN Test, dc=infn, dc=it”; Data. Access. Protocol = “gridftp”; Std. Error = “stderr. log”; Std. Output = “stdout. log”; Output. Sandbox = {“stderr. log”, “stdout. log”}; Input. Sandbox = {“home/joda/test/grid. Test”}; Rank = “other. Max. Cpu. Time”; Requirements = other. Architecture==“INTEL” && other. Op. Sys==“LINUX” && other. Free. Cpus >=4; The EDG Testbed Software - 36
A Job Submission Example UI JDL Replica Catalogue (RC) Information Service (IS) Resource Broker (RB) Logging & Book-keeping (LB) Job Submission Service (JSS) Storage Element (SE) Compute Element CE) The EDG Testbed Software - 37
A Job Submission Example UI JDL Input Sandbox Replica Catalogue (RC) Job Status Information Service (IS) submitted Job Submit Event Resource Broker (RB) Logging & Book-keeping (LB) Job Submission Service (JSS) Storage Element (SE) Compute Element (CE) The EDG Testbed Software - 38
A Job Submission Example UI JDL Replica Catalogue (RC) Job Status Information Service (IS) submitted waiting Resource Broker (RB) Logging & Book-keeping (LB) Job Submission Service (JSS) Storage Element (SE) Compute Element (CE) The EDG Testbed Software - 39
A Job Submission Example UI JDL Replica Catalogue (RC) Job Status Information Service (IS) submitted waiting ready Resource Broker (RB) Logging & Book-keeping (LB) Job Submission Service (JSS) Storage Element (SE) Compute Element (CE) The EDG Testbed Software - 40
A Job Submission Example UI JDL Replica Catalogue (RC) Information Service (IS) Job Status submitted waiting ready scheduled Resource Broker (RB) Logging & Book-keeping (LB) Broker. Info Storage Element (SE) Job Submission Service (JSS) Compute Element (CE) The EDG Testbed Software - 41
A Job Submission Example UI JDL Replica Catalogue (RC) Job Status Information Service (IS) submitted waiting ready Input Sandbox scheduled Resource Broker (RB) Logging & Book-keeping (LB) Job Submission Service (JSS) running Storage Element (SE) Compute Element (CE) The EDG Testbed Software - 42
A Job Submission Example Replica Catalogue (RC) UI JDL Job Status Information Service (IS) submitted waiting ready scheduled Resource Broker (RB) Logging & Book-keeping (LB) Job Submission Service (JSS) Job Status running Storage Element (SE) Compute Element (CE) The EDG Testbed Software - 43
A Job Submission Example Replica Catalogue UI JDL Job Status Information Service submitted waiting ready scheduled Resource Broker running Storage Element Logging & Book-keeping done Job Submission Service Job Status Compute Element The EDG Testbed Software - 44
A Job Submission Example Replica Catalogue UI JDL Job Status submitted Information Service waiting ready scheduled Resource Broker running Storage Element Logging & Book-keeping Job Submission Service Output Sandbox Job Status done outputready Compute Element The EDG Testbed Software - 45
A Job Submission Example Replica Catalogue (RC) UI JDL Job Status submitted Information Service (IS) waiting ready Output Sandbox Logging & Book-keeping (LB) scheduled Resource Broker (RB) Job Submission Service (JS) running Storage Element (SE) done outputready Compute Element (CE) cleared The EDG Testbed Software - 46
EDG Data Management Tools u Tools for n n Copying data n Managing and replicating data n u Locating data Meta Data management On EDG Testbed you have n EDG Replica Catalog n globus-url-copy (Grid. FTP) n EDG Replica Manager n Grid Data Mirroring Package (GDMP) The EDG Testbed Software - 47
EDG Replica Catalog u Based upon the Globus LDAP Replica Catalog (will be replaced by RLS) u Stores LFN/PFN mappings and additional information (e. g. filesize): n Physical File Name (PFN): host + full path & and file name n Logical File Name (LFN): logical name that may be resolved to PFNs n LFN : PFN = 1 : n u Only files on storage elements may be registered u Each VO has a specific storage dir on an SE u Example PFN: lxshare 0222. cern. ch/flatfiles/SE 1/iteam/file 1. dat host storage dir u LFN must be full path of file starting from storage dir LFN of above PFN: file 1. dat The EDG Testbed Software - 48
EDG Replica Catalog u API and command line tools n add. Logical. File. Name n get. Logical. File. Name n delete. Logical. File. Name n get. Physical. File. Name n add. Physical. File. Name n delete. Physical. File. Name n add. Logical. File. Attribute n get. Logical. File. Attribute n delete. Logical. File. Attribute http: //cmsdoc. cern. ch/cms/grid/userguide/gdmp-3 -0/node 85. html The EDG Testbed Software - 49
globus-url-copy u Low level tool for secure copying globus-url-copy <protocol>: //<source file> <protocol>: //<destination file> u Main Protocols: n gsiftp – for secure transfer, only available on SE and CE n file – for accessing files stored on the local file system on e. g. UI, WN globus-url-copy file: //`pwd`/file 1. dat gsiftp: //lxshare 0222. cern. ch/ flatfiles/SE 1/EDGTutorial/file 1. dat The EDG Testbed Software - 50
The EDG Replica Manager u Extends the Globus replica manager u Client side tool u Allows replication (copy) and registering of files in RC u Keeps RC consistent with stored data. The EDG Testbed Software - 51
The Replica Manager APIs u (un)register. Entry(Logical. File. Name lfn, File. Name source) n u Replica Catalogue operations only - no file transfer copy. File(File. Name source, File. Name destination, String protocol) n allows for third-party transfer n transfer between: s two Storage. Elements or s Computing. Element and Storage Element s Space management policies under development The EDG Testbed Software - 52
The Replica Manager APIs u copy. And. Register. File(Logical. File. Name lfn, File. Name source, File. Name destination, String protocol) n third-party transfer but : files can only be registered in Replica Catalogue if destination PFN contains a valid SE u replicate. File(Logical. File. Name lfn, File. Name source, File. Name destination, String protocol) u delete. File(Logical. File. Name lfn, File. Name source) The EDG Testbed Software - 53
u based on CMS requirements for replicating Objectivity files for High Level Trigger studies u production prototype project for evaluating Grid technologies (especially Globus) u http: //cern. ch/GDMP The EDG Testbed Software - 54
Overview of Components EDG Replica Catalogue GDMP client Site 1 Site 2 Site 3 The EDG Testbed Software - 55
Subscription Model n All the sites that subscribe to a particular site get notified whenever there is an update in its catalog. Site 1 Site 2 Subscriber list subscribe Site 3 The EDG Testbed Software - 56
Export / Import Catalogue n Export Catalog s s n Import Catalog s s n information about the new files produced. is published information about the files which have been published by other sites but not yet transferred locally As soon as the file is transferred locally, it is removed from the import catalogue. Site 1 Site 2 export catalog 1)register, publish new files 1) get info about new files import catalog 3) delete files Possible to pull the information about new files into your import catalogue. Site 3 2) transfer files The EDG Testbed Software - 57
Usage u gdmp_ping n Ping a GDMP server and get its status u gdmp_host_subscribe n first thing to be done by a site u gdmp_register_local_file n Registers a file in local file catalogue but NOT in Replica Catalogue (RC) u gdmp_publish_catalogue n send information of newly created files to subscribed hosts (no real data transfer) – update RC u gdmp_replicate_get n - gdmp_replicate_put get/put all the files from the import catalogue – update RC u gdmp_remove_local_file n Delete a local file and update RC u gdmp_get_catalogue n Get remote catalogue contents – for error recovery The EDG Testbed Software - 58
GDMP vs. EDG Replica Manager u GDMP u Replica Manager n Replicates sets of files n Replicates single files n Replication between SEs, CEs to SE. n Mass storage interface n File size as logical attribute n Subscription model n Event notification n CRC file size check n Support for Objectivity The EDG Testbed Software - 59
Скачать презентацию Introduction to the current EDG Testbed Software Kraków
7743664a807de3645896384b38e29c01.ppt