1 / 57

Globus Toolkit Execution Management

Globus Toolkit Execution Management. G T 4. Delegation Service. Community Scheduler Framework [contribution]. Python WS Core [contribution]. C WS Core. G T 3. Community Authorization Service. OGSA-DAI [Tech Preview]. WS Authentication Authorization. Reliable File Transfer.

Download Presentation

Globus Toolkit Execution Management

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Globus ToolkitExecution Management

  2. G T 4 Delegation Service Community Scheduler Framework [contribution] Python WS Core [contribution] C WS Core G T 3 CommunityAuthorization Service OGSA-DAI [Tech Preview] WS Authentication Authorization Reliable File Transfer Java WS Core Grid Resource Allocation Mgmt (WS GRAM) Monitoring & Discovery System (MDS4) G T 2 Pre-WS Authentication Authorization GridFTP Grid Resource Allocation Mgmt (Pre-WS GRAM) Monitoring & Discovery System (MDS2) C Common Libraries G T 3 Replica Location Service XIO G T 4 Credential Management Web ServicesComponents Non-WS Components Security Data Management Execution Management Information Services CommonRuntime I Foster

  3. Grid Resource Allocation Manager (GRAM)Job submission

  4. Resource Management • Job submission • Job status • Basic resource allocation

  5. Outline • GT2 job submission using RSL version 1 language. • GT 3.2 job submission using RSL version 2 language. • GT 4 job submission (can use RSL version language

  6. Resource Allocation Globus (2, 3.2, or 4.0) does not have its own job scheduler to find resources and automatically send jobs to suitable machines. For that, use a separate scheduler, e.g. Condor, Sun Grid Engine, LSF, PBS, … .

  7. Globus Version 2 (pre-2004) Pre-WS GRAM

  8. Globus version 2 From: “Introduction to Grid Computing with Globus,” IBM Redbooks, Fig. 7-3.

  9. GT 2 GRAM Job startup done using GRAM service. Consist of: • Gatekeeper • Job Manager Job manager can connect to a local resource manager (scheduler) GASS service -- provides access to remote files and for redirecting standard output streams.

  10. GRAM Commands • globusrun -- Runs a single executable on a remote site. • globus-job-run -- Allows you to run a job at one or several remote resources. Uses globusrun to submit job. • globus-job-submit -- For batch job submission (e.g. using a local scheduling job manager). Not recommended; use globus-job-run or globusrun instead, with job manager specified

  11. Scheduler • Can specify a job scheduler with globusrun, by adding scheduler name to hostname: <hostname>/jobmanager-lsf

  12. Specifying job • Command used a file to describe job in a language called Resource Specification Language, RSL • RSL Version 1 -- a metalanguage describing job and its required execution.

  13. Resource Specification LanguageRSL Provides a specification for: • Resource requirements - machine type, number of nodes, memory, etc. • Job description - directory, executable, arguments, environment

  14. RSL Version 1Constraints ExampleConjunction (AND): & • To create 3-5 instances of myProg, each on a machine with at least 64 Mbytes memory available to me for 1 hours: & (executable=myProg) (count>=3)(count<=5)(memory>=64) (max_time=60)

  15. RSL Version 1Constraints ExampleDisjunction (OR): | • To create 5 instances of myProg, each on a machine with at least 64 Mbytes memory or 7 instances of myProg, each on a machine with at least 32 Mbytes memory : &(executable=myProg) (|(&(count=5)(memory>=64)) (&(count=7)(memory>=32)))

  16. RSL version 1Requesting multiple resourcesmultirequest: + • To execute 5 instances of myProg1 on a machine with at least 64 Mbytes memory and execute 2 instances of myProg2: +(&(count=5)(memory>=64)) (executable=myProg1)) (&(count=2)(executable=myProg2))

  17. Can specify different resource managers on different machines using resourceManagerContact attribute.

  18. RSL creation with Globus version 2 • GT2 globus-job-run can be used to generate RSL from command line arguments with -dumprsl flag • -help gives options

  19. Globus 3.2

  20. GT 3.2 GRAM“Globus Resource Allocation Manager” A set of “OGSI” compliant services provided to start remote jobs. notably: • Master Managed Job Factory Service (MMJFS). Also a set of non-OGSI compliant services (Gatekeeper, Jobmanager) from pre-GT3.

  21. Globus GT 3.x

  22. Resource Specification Language, RSL, version 2 • GT3 and GT 4 use RSL version 2. (Some differences in RSL language specification in GT4, so not completely interchangeable.) • RSL Version 2 is an XML language.

  23. Resource Specification Language Version 2 (RSL -2) • Can specify everything from executable, paths, arguments, input/output, error file, number of processes, max/min execution time, max/min memory, job type etc. etc.

  24. RSL-2 • Much more elegant and flexible, and in keeping with systems using XML. • Can use XML parsers. • Allows more powerful mechanisms with job schedulers. • Resource scheduler/broker applies specification to local resources.

  25. RSL-2 ExampleSpecifying Executable(executable=/bin/echo) <gram:executable> <rsl:path> <rsl:stringElement value="/bin/echo"/> </rsl:path> </gram:executable> In GT 4 version of RSL-2, can simply write: <executable>/bin/echo</executable>

  26. RSL-2 ExampleSpecifying Directory(directory=“/bin”) <gram:directory> <rsl:path> <rsl:stringElement value="/bin/"/> </rsl:path> </gram:directory> In GT 4 version of RSL-2, can simply write: <directory>/bin/</directory>

  27. RSL-2 ExampleSpecifying Number(count=1) <gram:count> <rsl:integer value="1"/> </gram:count> In GT 4 version of RSL-2, can simply write: <count>1</count>

  28. RSL-2 ExampleSpecifying Arguments(arguments=“Hello”) <gram:arguments> <rsl:string> <rsl:stringElement value="Hello World"/> </rsl:string> </gram:arguments> In GT 4 version of RSL-2, can simply write: <argment>hello world</argument>

  29. <?xml version="1.0" encoding="UTF-8"?> • <rsl:rsl xmlns:rsl="http://www.globus.org/namespaces/2003/04/rsl" • xmlns:gram="http://www.globus.org/namespaces/2003/04/rsl/gram" • xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" • xsi:schemaLocation=" • http://www.globus.org/namespaces/2003/04/rsl • c:/ogsa-3.0/schema/base/gram/rsl.xsd • http://www.globus.org/namespaces/2003/04/rsl/gram • c:/ogsa-3.0/schema/base/gram/gram_rsl.xsd"> • <gram:job> • <gram:executable> <rsl:path> • <rsl:stringElement value="/bin/echo"/> </rsl:path> • </gram:executable> • <gram:directory> <rsl:path> • <rsl:stringElement value="/bin"/> </rsl:path> • </gram:directory> • <gram:arguments> • <rsl:string> <rsl:stringElement value="Hello World"/> </rsl:string> • </gram:arguments> • <gram:stdin> <rsl:path> • <rsl:stringElement value="/dev/null"/> </rsl:path> </gram:stdin> • <gram:stdout> • <rsl:pathArray> • <rsl:path> • <rsl:substitutionRef name="HOME"/> • <rsl:stringElement value="/stdout"/> • </rsl:path> • </rsl:pathArray> • </gram:stdout> • <gram:stderr> • <rsl:pathArray> • <rsl:path> • <rsl:substitutionRef name="HOME"/> • <rsl:stringElement value="/stderr"/> • </rsl:path> • </rsl:pathArray> • </gram:stderr> • <gram:count> <rsl:integer value="1"/> </gram:count> • <gram:jobType> • <gram:enumeration> • <gram:enumerationValue> <gram:multiple/> </gram:enumerationValue> • </gram:enumeration> • </gram:jobType> • <gram:gramMyJobType> • <gram:enumeration> • <gram:enumerationValue> <gram:collective/> </gram:enumerationValue> • </gram:enumeration> • </gram:gramMyJobType> • <gram:dryRun> <rsl:boolean value="false"/> </gram:dryRun> • <gram:saveState> <rsl:boolean value="true"/> </gram:saveState> • <gram:twoPhase> <rsl:integer value="600"/> </gram:twoPhase> • </gram:job> • </rsl:rsl> RSL and (GT 3.2) RSL-2 comparison for echo program &((executable=/bin/echo) (directory="/bin") (arguments="Hello World") (stdin=/dev/null) (stdout="stdout") (stderr="stderr") (count=1) )

  30. Running GT 3 Job • Command: managed-job-globusrun and arguments -- named master job factory service to process job and an xml file to specify job. • Command equivalent to GT 2 globusrun command.

  31. Globus 4.0

  32. GT 4 WS-GRAM

  33. In WS GRAM, jobs started by the ManagedExecutionJobService, which is a Java service implementation running within globus service container.

  34. Running GT 4 Job • Command: globusrun-ws and arguments to specify job. • Equivalent to GT 3.2 managed-job-globusrun command, and GT 2 globusrun command.

  35. GT4 job submission command globusrun-ws • Submit and monitor GRAM jobs • Replaces (java) managed-job-globusrun • Written in C, faster startup and execution • Supports multiple and single job submission • Handles credential management • Streaming of job stdout/err during execution

  36. Simple job submission • Step 1: Create proxy with grid-proxy-int command. • Step 2: globusrun-ws with parameters to specify job.

  37. Some globusrun-ws flags for job submission

  38. Job submission -submit Submits (or resubmits) a job to a job host in one of three output modes: batch, interactive, or interactive-streaming. This flag needed.

  39. Specifying where job is submitted(ManagedJobFactory) -F Specifies the “contact” for the job submission. Default https://localhost:8443/wsrf/services/ ManagedJobFactoryService In assignment 3, simply localhost and container port number used, i.e. -F localhost:8443

  40. Submitting a single job - c Causes globusrun-ws to generate a simple job description with the named program and arguments. This flag, if used, must be the last flag.

  41. Example Submit program echo with argument hello to default local host. % globusrun-ws –submit –c /bin/echo hello Submitting job...Done. Job ID: uuid:d23a7be0-f87c-11d9-a53b-0011115aae1f Termination time: 07/20/2005 17:44 GMT Current job state: Active Current job state: CleanUp Current job state: Done Destroying job...Done.

  42. A successful submission will create a new ManagedJob resource with its own unique EPR for messaging. • globusrun-ws will output this EPR to a file when requested, as the sole standard output when running in batch mode.

  43. Selecting a different host Example $ globusrun-ws –submit –F https://140.221.65.193:4444/wsrf/services/managedJobFactoryService –c /bin/echo hello

  44. Using an RSL file –f Similar to -c except job description held in a file. Example globusrun-ws –submit –f echo.xml where echo.xml is an RSL-2 file describing job.

  45. Contents of echo.xml <job> <executable>/bin/echo</executable> <argument>hello</argument> </job>

  46. Batch Submission -batchResults in ManagedJob EPR as the sole standard output (unless in quiet mode) and then exits. -o filename Created ManagedJob EPR written to file (if submission successful)

  47. Batch Job Submission $ globusrun-ws –submit –batch –o job_epr –s /bin/sleep 50 Submitting job…Done JoB ID: uuid:f9544174-60c5-11d9-97e3-0002a5ad41e5 Termination time: 01/08/2005 16:05 GMT

  48. Monitoring Batch Submission -monitor Attaches to an existing job in interactive or interactive-streaming output modes. -j filename EPR for ManagedJob read from file.

  49. Monitoring Batch Job globusrun-ws –monitor –j job_epr job state: Active Current job state: CleanUp Current job state: Done Requesting original job description...Done. Destroying job...Done

  50. Batch Submission -status Reports the current state of the job and exits -kill Requests immediate cancellation of job and exits.

More Related