Information and Monitoring Rob Byrom / WP3

advertisement

WP3

Information and

Monitoring

Rob Byrom / WP3

r.byrom@rl.ac.uk

Plan of Talk

• Final EDG Release

• Who’s using R-GMA?

• Migrating to EGEE

• Working within ARDA

• R-GMA Future Work

WP3

Information & Monitoring Robert Byrom/RAL - 05/1/2004 2

Final EDG Release

• Robust

– Significant improvements to robustness mid-December

– Information Catalogue ran continuously for 1 month without restart

• Performance

– Significant improvements to performance

– Carried out extensive testing using CMS simulation test framework

– Performance figures published daily & available via our web site

• http://hepunx.rl.ac.uk/edg/wp3/performance/index.html

• Authentication

– Implemented but currently disabled on app testbed

• Nagios

– Use alert system to identify problems on the various testbeds

• https://heplnx30.pp.rl.ac.uk/nagios/

• Ganglia Integration

– Ganglia data published via R-GMA (called Ranglia)

WP3

Information & Monitoring Robert Byrom/RAL - 05/1/2004 3

Who’s using R-GMA?

• BaBar/Southern Tier 2

– Ranglia

• CMS

– Boss

• David Colling, Peter Hobson et al

• D0

– Data challenge

• Jeff Templon - Nikhef

• LCG

– Publishing PBS accounting data for LCG

• Network Monitoring

– WP7

Information & Monitoring Robert Byrom/RAL - 05/1/2004

WP3

4

Migrating to EGEE

• WP3 to form the Information Collection ‘Cluster’

• JRA1 - Objectives

– Provide robust, supportable middleware components

• Select, re-engineer and integrate identified components

• Evolve towards Services Oriented Architecture

• Adopt emerging WS-RF standards

• Support multiple platforms

• Define defect handling processes and responsibilities

– Selection of Middleware based on requirements of:

• The applications (Bio & HEP)

• The Operations

WP3

Information & Monitoring Robert Byrom/RAL - 05/1/2004 5

EGEE Milestones and

Deliverables for the first year

Month

M03

M03

M03

M05

M05

M09

M10

M12

WP3

Deliverables

&

Milestones

MJRA1.1

DJRA1.1

MJRA1.2

MJRA1.3

DJRA1.2

MJRA1.4

MJRA1.5

DJRA1.3

Item

Lead

Partner

Tools for middleware engineering and integration deployed

(Document) Architecture and Planning (Release 1)

CERN

CERN

Software cluster development and testing infrastructure available

Integration and testing infrastructure in place including test plans (Release 1)

(Document) Design of grid services (Release 1)

CERN

CERN

CERN

Software for the Release Candidate 1

Integrated Release Candidate 1 enters testing and validation period (Release 1)

(Software) Software and associated documentation

(Release 1)

CERN

CERN

CERN

Information & Monitoring Robert Byrom/RAL - 05/1/2004 6

Working within ARDA

• Working in ARDA middleware ‘task force’

– Presented by Miron Livny at ARDA workshop

– Promoting an integrated approach to info services

– Hope document will be available soon

• First Web Service prototype for R-GMA

– Within two weeks

WP3

Information & Monitoring Robert Byrom/RAL - 05/1/2004 7

R-GMA Future Work

• Functionality

– Registry & Schema Replication

– Enhanced Mediator

– Authorization

– Multiple VO Support

• Re-engineering

– Packaging

– Web Services

Information & Monitoring Robert Byrom/RAL - 05/1/2004

WP3

8

Registry & Schema

Replication

• Registry Replication

– Each logical Registry has multiple physical

“copies”

– Transmit new and deleted records and a checksum

– Consumer & Producer use any Registry instance

• Fail over mechanism

– Prototype complete

WP3

• Schema Replication

– Mechanism to vote for a master schema

– Synchronise all Schemas with the master

Information & Monitoring Robert Byrom/RAL - 05/1/2004 9

Enhanced Mediator

• Mediator brokers the correct Producer/s for a

Consumer query

• Meta-data to be made available to

Consumers via RGMAWarnings

• Support for republisher hierarchies

• Ability to deal with more complex predicates

WP3

Information & Monitoring Robert Byrom/RAL - 05/1/2004 10

Authorization

• Reviewing requirements

• Considering the best way to express the required authorisation semantics

• Looking into GACL syntax

WP3

Information & Monitoring Robert Byrom/RAL - 05/1/2004 11

Multiple VOs

• Currently one virtual Registry & Schema per grid

– Used by all VOs

• Moving to one virtual Registry & Schema per

VO

– To be added as part of the EGEE re-engineering

– Information to multiple VOs need only be published once

WP3

12 Information & Monitoring Robert Byrom/RAL - 05/1/2004

Re-Engineering

• EGEE provides good opportunity to improve upon existing design

– Producing a documented design

• Agreed UML diagrams

– Will refactor to this design

– Looking into JMS

• To enable a pluggable streaming interface between

Producers and Consumers

WP3

Information & Monitoring Robert Byrom/RAL - 05/1/2004 13

Quality

• Emphasis on better ‘quality’ control from experiences with EDG

– System testing

• Currently using CMS testing framework with Nagios alerts

• Need to identify bugs/problems early on

– Improving approach to code design & implementation

• Using UML to document new ideas

• More unit testing to improve code coverage

WP3

Information & Monitoring Robert Byrom/RAL - 05/1/2004 14

Packaging

• Remove dependency on a particular platform

– Provide distribution in tar/zip format

WP3

Information & Monitoring Robert Byrom/RAL - 05/1/2004 15

Web Services

• Currently we have Schema, Registry and

Consumer Web Services

– Only Producers left to be converted (a few days work)

• Provide wrappers for compatibility with our current API’s

– Using eu-egee name space

• Continue to follow emerging standards

– WS-RF

– WS-Notification

WP3

Information & Monitoring Robert Byrom/RAL - 05/1/2004 16

Summary

WP3

• R-GMA robustness improved considerably

• R-GMA being used by BaBar/Southern Tier 2,

CMS, D0, LCG

• First release for EGEE due in December

2004

• Working within ARDA to provide Information &

Monitoring Service

• Web service prototype well under way

Information & Monitoring Robert Byrom/RAL - 05/1/2004 17

Download