DELETION SERVICE ISSUES ADC Development meeting 09.08.2011.

Slides:



Advertisements
Similar presentations
Case Study: Photo.net March 20, What is photo.net? An online learning community for amateur and professional photographers 90,000 registered users.
Advertisements

Debugging/Tuning Queries via iSeries Navigator Tom McKinley
Module 13: Performance Tuning. Overview Performance tuning methodologies Instance level Database level Application level Overview of tools and techniques.
Futures – Alpha Cloud Deployment and Application Management.
Bookshelf.EXE - BX A dynamic version of Bookshelf –Automatic submission of algorithm implementations, data and benchmarks into database Distributed computing.
GGF Toronto Spitfire A Relational DB Service for the Grid Peter Z. Kunszt European DataGrid Data Management CERN Database Group.
Chapter 6: Database Evolution Title: AutoAdmin “What-if” Index Analysis Utility Authors: Surajit Chaudhuri, Vivek Narasayya ACM SIGMOD 1998.
Data Warehouse success depends on metadata
Connect with life Praveen Srvatsa Director | AsthraSoft Consulting Microsoft Regional Director, Bangalore Microsoft MVP, ASP.NET.
19 February CASTOR Monitoring developments Theodoros Rekatsinas, Witek Pokorski, Dennis Waldron, Dirk Duellmann,
Copyright © 2007 Quest Software The Changing Role of SQL Server DBA’s Bryan Oliver SQL Server Domain Expert Quest Software.
LHC Experiment Dashboard Main areas covered by the Experiment Dashboard: Data processing monitoring (job monitoring) Data transfer monitoring Site/service.
Module 18 Monitoring SQL Server 2008 R2. Module Overview Monitoring Activity Capturing and Managing Performance Data Analyzing Collected Performance Data.
CERN - IT Department CH-1211 Genève 23 Switzerland t Monitoring the ATLAS Distributed Data Management System Ricardo Rocha (CERN) on behalf.
Implementation Yaodong Bi. Introduction to Implementation Purposes of Implementation – Plan the system integrations required in each iteration – Distribute.
Online Database Support Experiences Diana Bonham, Dennis Box, Anil Kumar, Julie Trumbo, Nelly Stanfield.
Monitoring Latency Sensitive Enterprise Applications on the Cloud Shankar Narayanan Ashiwan Sivakumar.
ATLAS DQ2 Deletion Service D.A. Oleynik, A.S. Petrosyan, V. Garonne, S. Campana (on behalf of the ATLAS Collaboration)
Eurotrace Hands-On The Eurotrace File System. 2 The Eurotrace file system Under MS ACCESS EUROTRACE generates several different files when you create.
Oracle 10g Database Administrator: Implementation and Administration Chapter 2 Tools and Architecture.
PanDA Summary Kaushik De Univ. of Texas at Arlington ADC Retreat, Naples Feb 4, 2011.
OSG Area Coordinator’s Report: Workload Management April 20 th, 2011 Maxim Potekhin BNL
Designing High Performance BIRT Reports Mica J. Block Director Actuate Corporate Engineers Actuate Corporation.
Database weekly reports Zbigniew Baranowski Carlos Fernando Gamboa.
EGEE-III INFSO-RI Enabling Grids for E-sciencE EGEE and gLite are registered trademarks GStat 2.0 Joanna Huang (ASGC) Laurence Field.
BNL DDM Status Report Hironori Ito Brookhaven National Laboratory.
1 Oracle Enterprise Manager Slides from Dominic Gélinas CIS
DDM Monitoring David Cameron Pedro Salgado Ricardo Rocha.
INTRODUCTION TO DBS Database: a collection of data describing the activities of one or more related organizations DBMS: software designed to assist in.
David Adams ATLAS DIAL/ADA JDL and catalogs David Adams BNL December 4, 2003 ATLAS software workshop Production session CERN.
Storage cleaner: deletes files on mass storage systems. It depends on the results of deletion, files can be set in states: deleted or to repeat deletion.
Author: Andrew C. Smith Abstract: LHCb's participation in LCG's Service Challenge 3 involves testing the bulk data transfer infrastructure developed to.
MultiJob pilot on Titan. ATLAS workloads on Titan Danila Oleynik (UTA), Sergey Panitkin (BNL) US ATLAS HPC. Technical meeting 18 September 2015.
INFSO-RI Enabling Grids for E-sciencE ARDA Experiment Dashboard Ricardo Rocha (ARDA – CERN) on behalf of the Dashboard Team.
XROOTD AND FEDERATED STORAGE MONITORING CURRENT STATUS AND ISSUES A.Petrosyan, D.Oleynik, J.Andreeva Creating federated data stores for the LHC CC-IN2P3,
Database authentication in CORAL and COOL Database authentication in CORAL and COOL Giacomo Govi Giacomo Govi CERN IT/PSS CERN IT/PSS On behalf of the.
The new FTS – proposal FTS status. EMI INFSO-RI /05/ FTS /05/ /05/ Bugs fixed – Support an SE publishing more than.
Distributed Logging Facility Castor External Operation Workshop, CERN, November 14th 2006 Dennis Waldron CERN / IT.
20 Copyright © 2008, Oracle. All rights reserved. Cache Management.
FTS monitoring work WLCG service reliability workshop November 2007 Alexander Uzhinskiy Andrey Nechaevskiy.
Summary of persistence discussions with LHCb and LCG/IT POOL team David Malon Argonne National Laboratory Joint ATLAS, LHCb, LCG/IT meeting.
1 A Scalable Distributed Data Management System for ATLAS David Cameron CERN CHEP 2006 Mumbai, India.
Site Services and Policies Summary Dirk Düllmann, CERN IT More details at
Status of tests in the LCG 3D database testbed Eva Dafonte Pérez LCG Database Deployment and Persistency Workshop.
DDM Central Catalogs and Central Database Pedro Salgado.
21 Copyright © 2008, Oracle. All rights reserved. Enabling Usage Tracking.
Distributed Data Management Miguel Branco 1 DQ2 status & plans BNL workshop October 3, 2007.
Replicazione e QoS nella gestione di database grid-oriented Barbara Martelli INFN - CNAF.
INFSO-RI Enabling Grids for E-sciencE File Transfer Software and Service SC3 Gavin McCance – JRA1 Data Management Cluster Service.
Troubleshooting Dennis Shasha and Philippe Bonnet, 2013.
VO Box discussion ATLAS NIKHEF January, 2006 Miguel Branco -
DB Questions and Answers open session (comments during session) WLCG Collaboration Workshop, CERN Geneva, 24 of April 2008.
1 Copyright © 2008, Oracle. All rights reserved. Repository Basics.
Design and Implementation of a High-Performance distributed web crawler Vladislav Shkapenyuk and Torsten Suel Proc. 18 th Data Engineering Conf., pp ,
SQL Server DBA Online TrainingSQL Server DBA Online Training.
Learn Structured Query Language to rule Database.
LCG Storage Management Workshop, CERN, 7th April 2005
Designing High Performance BIRT Reports
Performance Management
Monitoring and Fault Tolerance
Simone Campana CERN IT-ES
Elizabeth Gallas - Oxford ADC Weekly September 13, 2011
Accounting at the T1/T2 Sites of the Italian Grid
Data Federation with Xrootd Wei Yang US ATLAS Computing Facility meeting Southern Methodist University, Oct 11-12, 2011.
Conditions Data access using FroNTier Squid cache Server
100% Exam Passing Guarantee & Money Back Assurance
1Z0-320 Dumps
A Web-Based Data Grid Chip Watson, Ian Bird, Jie Chen,
5 Azure Services Every .NET Developer Needs to Know
Roadmap for Data Management and Caching
Presentation transcript:

DELETION SERVICE ISSUES ADC Development meeting

Productivity (overall) Productivity of service: A year ago: 500K of files per day (was enough) Now: 2M per day (2,5M – estimated value, 1,5M – lowest limit) (not always enough) Monitoring: Number of operation of DS increase at least x5 Speed of on-line aggregation decrease -> slow report generation Some reports stuck (overview for last week)

Productivity (in details) Not so high as wanted (less 10k of files for one site) Slow LFC clearing Problems: Low productivity of one threat (a lot of time spent by Deletion Agent for intercommunication with backend) One LFC cleaner per LFC server -> one cleaner per cloud (last year implementation)

New version of Deletion Service In testing with production data for two endpoints in BNL (BNL-OSG2-MCDISK, BNL-OSG-USERDISK) Deletion Agent deployed on atlddm17, Deletion server deployed on dedicated instance (atlddm16) – no impact from other services. New Features Changed algorithm of intercommunication of Storage cleaner for decreasing number of calls of CC (HTTP server) Number of catalog cleaners not depend from number of LFC servers Limitation of number of resolving files (give possibility to have proper size of huge operation table in DB)

First results. It’s preliminary results: more tests with loading needed Storage cleaner performance increase (22-28K of deleted files per hour). Decreased time for intercommunication with backend (bulk operations): No significant changes for LFC Cleaner for single thread, but number of threads can be increased Idea 1: number of operations with DB didn’t change –> CC performance? Idea 2: Oracle can update data with around 100Hz (for Deletion Service) Chunk sizeOld versionNew version sec1,5-5 sec.

Next steps (very nearest plans) Continue test usage with increasing of loading (next two weeks) -> will be good to have a plan. Implement same strategy for Catalog Cleaner as for Storage Cleaner (operations across datasets, ‘bulk’ operations for update) Special processing for jumbo datasets Current implementation of resolving are not coherent for jumbo datasets. It affect DB and CC, so different implementation needed.

Needed features. Some times of data should be removed for limited period, and will be good if biggest files will be deleted from storage early then smallest -> priority of deletion needed for this case. Directory deletion Local installation (implemented in current architecture – but validation needed)

“Plan B” Reviewing of architecture, in case implemented features will not give estimated effect Decentralization for most loading operation Significant changes in DB structure Etc. Plan B will means significant changes for Deletion Monitoring too…

Deletion monitoring Decrease of performance due to a lot of aggregation. Tunes of query have limits DB structure oriented for mostly for deletion process Not all metrics shown (LFC deletion missing) Some more metrics needed for defining problems/productivity of each component (Storage, LFC, Backend)

Deletion monitoring. Nearest plan Performance increase: extension of DB schema with summary tables, filled by triggers and jobs. Interface extensions Additional metrics: already collected by Deletion Agent and put to log... But can be send someware (asynchronously ideally) Should not be collected in production Oracle Can be collected in NoSQL -> most of operation will be insert, select with aggregation