February, 2015 Bill Loewe



Similar documents
NetApp High-Performance Computing Solution for Lustre: Solution Guide

An Alternative Storage Solution for MapReduce. Eric Lomascolo Director, Solutions Marketing

Seagate Lustre Update. Peter Bojanic

Architecting a High Performance Storage System

New Storage System Solutions

Commoditisation of the High-End Research Storage Market with the Dell MD3460 & Intel Enterprise Edition Lustre

High Performance Computing OpenStack Options. September 22, 2015

High-Availability and Scalable Cluster-in-a-Box HPC Storage Solution

Cray Lustre File System Monitoring

Lustre * Filesystem for Cloud and Hadoop *

Data management challenges in todays Healthcare and Life Sciences ecosystems

Panasas at the RCF. Fall 2005 Robert Petkus RHIC/USATLAS Computing Facility Brookhaven National Laboratory. Robert Petkus Panasas at the RCF

GPFS Storage Server. Concepts and Setup in Lemanicus BG/Q system" Christian Clémençon (EPFL-DIT)" " 4 April 2013"

Reference Design: Scalable Object Storage with Seagate Kinetic, Supermicro, and SwiftStack

Xyratex Update. Michael K. Connolly. Partner and Alliances Development

Lustre failover experience

UCS M-Series Modular Servers

Hyper-V over SMB Remote File Storage support in Windows Server 8 Hyper-V. Jose Barreto Principal Program Manager Microsoft Corporation

E4 UNIFIED STORAGE powered by Syneto

Application Performance for High Performance Computing Environments

IBM System x GPFS Storage Server

How To Create A Multi Disk Raid

HPC Advisory Council

DIABLO TECHNOLOGIES MEMORY CHANNEL STORAGE AND VMWARE VIRTUAL SAN : VDI ACCELERATION


THE SUN STORAGE AND ARCHIVE SOLUTION FOR HPC

How swift is your Swift? Ning Zhang, OpenStack Engineer at Zmanda Chander Kant, CEO at Zmanda

Current Status of FEFS for the K computer

Stovepipes to Clouds. Rick Reid Principal Engineer SGI Federal by SGI Federal. Published by The Aerospace Corporation with permission.

Introduction to NetApp Infinite Volume

Improving Lustre OST Performance with ClusterStor GridRAID. John Fragalla Principal Architect High Performance Computing

SMB Direct for SQL Server and Private Cloud

Lessons learned from parallel file system operation

NIVEO Network Attached Storage Series NNAS-D5 NNAS-R4. More information:

Performance Comparison of SQL based Big Data Analytics with Lustre and HDFS file systems

Lustre SMB Gateway. Integrating Lustre with Windows

The Use of Flash in Large-Scale Storage Systems.

Netapp HPC Solution for Lustre. Rich Fenton UK Solutions Architect

Evaluation of Dell PowerEdge VRTX Shared PERC8 in Failover Scenario

HPC Update: Engagement Model

IBM System x GPFS Storage Server

LS-DYNA Best-Practices: Networking, MPI and Parallel File System Effect on LS-DYNA Performance

Quantum StorNext. Product Brief: Distributed LAN Client

Intel RAID SSD Cache Controller RCS25ZB040

Investigation of storage options for scientific computing on Grid and Cloud facilities

Performance Comparison of Intel Enterprise Edition for Lustre* software and HDFS for MapReduce Applications

Hyper-V over SMB: Remote File Storage Support in Windows Server 2012 Hyper-V. Jose Barreto Principal Program Manager Microsoft Corporation

VMware Virtual SAN Backup Using VMware vsphere Data Protection Advanced SEPTEMBER 2014

Sun Storage Perspective & Lustre Architecture. Dr. Peter Braam VP Sun Microsystems

New and Improved Lustre Performance Monitoring Tool. Torben Kling Petersen, PhD Principal Engineer. Chris Bloxham Principal Architect

High Availability and Backup Strategies for the Lustre MDS Server

Big data management with IBM General Parallel File System

BlueArc unified network storage systems 7th TF-Storage Meeting. Scale Bigger, Store Smarter, Accelerate Everything

CERN Cloud Storage Evaluation Geoffray Adde, Dirk Duellmann, Maitane Zotes CERN IT

Achieving Real-Time Business Solutions Using Graph Database Technology and High Performance Networks

Increasing Storage Performance

POWER ALL GLOBAL FILE SYSTEM (PGFS)

Network Storage Appliance

Comparing SMB Direct 3.0 performance over RoCE, InfiniBand and Ethernet. September 2014

The Panasas Parallel Storage Cluster. Acknowledgement: Some of the material presented is under copyright by Panasas Inc.

Understanding Microsoft Storage Spaces

The last 18 months. AutoScale. IaaS. BizTalk Services Hyper-V Disaster Recovery Support. Multi-Factor Auth. Hyper-V Recovery.

Cluster Implementation and Management; Scheduling

Cisco Small Business NSS2000 Series Network Storage System

Four Reasons To Start Working With NFSv4.1 Now

Object storage in Cloud Computing and Embedded Processing

Direct NFS - Design considerations for next-gen NAS appliances optimized for database workloads Akshay Shah Gurmeet Goindi Oracle

IBM General Parallel File System (GPFS ) 3.5 File Placement Optimizer (FPO)

Appro Supercomputer Solutions Best Practices Appro 2012 Deployment Successes. Anthony Kenisky, VP of North America Sales

THE SUMMARY. ARKSERIES - pg. 3. ULTRASERIES - pg. 5. EXTREMESERIES - pg. 9

Windows Server 2012 授 權 說 明

modular Storage Solutions MSS Series

SMB Advanced Networking for Fault Tolerance and Performance. Jose Barreto Principal Program Managers Microsoft Corporation

What s New with VMware Virtual Infrastructure

Cray DVS: Data Virtualization Service

Cisco Small Business NSS3000 Series Network Storage System

Vess A2000 Series HA Surveillance with Milestone XProtect VMS Version 1.0

WOS OBJECT STORAGE PRODUCT BROCHURE DDN.COM Full Spectrum Object Storage

Building Storage Clouds for Online Applications A Case for Optimized Object Storage

EXAScaler. Product Release Notes. Version Revision A0

præsentation oktober 2011

QoS-Aware Storage Virtualization for Cloud File Systems. Christoph Kleineweber (Speaker) Alexander Reinefeld Thorsten Schütt. Zuse Institute Berlin

Desktop Consolidation. Stéphane Verdy, CTO Devon IT

EMC DATA DOMAIN OPERATING SYSTEM

Cloud Storage. Parallels. Performance Benchmark Results. White Paper.

Enabling High performance Big Data platform with RDMA

Installation Guide July 2009

Sonexion GridRAID Characteristics

Storage Architectures for Big Data in the Cloud

Eloquence Training What s new in Eloquence B.08.00

(Scale Out NAS System)

EMC AVAMAR. a reason for Cloud. Deduplication backup software Replication for Disaster Recovery

OSDsim - a Simulation and Design Platform of an Object-based Storage Device

Service Description CloudSure Public, Private & Hybrid Cloud

Parallels Cloud Server 6.0

June Blade.org 2009 ALL RIGHTS RESERVED

Transcription:

February, 2015 Bill Loewe

Agenda System Metadata, a growing issue Parallel System - Lustre Overview Metadata and Distributed Namespace Test setup and implementation for metadata testing Scaling Metadata Servers High Availability Seagate Confidential

Metadata Performance System Performance typically viewed in Bandwidth Bandwidth problem largely addressed, but metadata is a growing issue. We see this in workloads with high numbers of files to access and process. Genome processing CPU Chip manufacturing Video compositing/rendering Seagate Confidential

Lustre Parallel System Lustre is an open source, distributed parallel file system Object-based design provides extreme scalability Compute clients interact directly with storage servers Comprised of: Clients Metadata Servers and Targets Storage Servers and Targets

Lustre Distributed NamespacE (DNE) Distributed NamespacE (DNE) is a new feature available in Lustre 2.5 that allows multiple MDS / MDT components to participate in a single file system. DNE allows the namespace to be divided across multiple metadata servers. Enables the size of the namespace and metadata throughput to be scaled with the number of servers. The Lustre DNE project is comprised of 2 phases. Seagate Confidential

Phase 1, Lustre 2.5 Release Remote Directories -- Lustre sub-directories are distributed over multiple metadata targets (MDTs). Sub-directory distribution is defined by an administrator. Remote Directories Root dir b dir c dir d dir e dir a dir b2 dir c2 dir d2 dir e2 Seagate Confidential

Phase 2, Lustre 2.7 Striped Directories -- The contents of a given directory are distributed over multiple MDTs. Striped Directories Striped Directory dir c2 dir e2 Seagate Confidential

Engineered Storage Solutions for HPC, Big Data & Cloud High speed networking (IB/40GB/e) Parallel file system/object Data protection High availability Flash optimization system (Ext4) Linux OS BIOS/IPMI ClusterStor GEM diagnostics Custom X86 embedded server Seagate storage platforms Seagate Storage Devices Architected Integrated Optimized Qualified Supported Seagate Confidential

Lustre Components Clients Directory Operations, open/close, metadata, and concurrency I/O and locking MDS creation, file status, and recovery OSS OSS OSS Seagate Confidential

ClusterStor Management Unit (CMU): Management and Metadata (MDS/MDT) CSM Manager and MDS/MGS Nodes 2RU 4-node Sandy Bridge Servers Server 1: CSM Mgmt Server 2: Boot Server 3: MGS Server 4: MDS Fault Tolerance (active/passive) Serviceability 2U24 JBOD MDT SAS JBOD for MDS/MGS/Management Disk Configuration Qty 4 Lustre Management (MGS) Qty 4 ClusterStor Management and NFS Qty 2 Global Hot spares Qty 14 Drives for MDT

Scalable Storage Unit (SSU) SSU 5U84 Enclosure 2 Object Storage Servers s per SSU Two (2) trays of 42 HDD s each for Object Storage Targets H/A on each SSU Infiniband QDR/FDR and 40Gb Ethernet data network connectivity

ClusterStor & Lustre 2.5 DNE Hardware DNE is available in ClusterStor v2.0 MDT0 is master and default in DNE environment DNE Servers are configured in active / active pairs Seagate 2U24 with 2 MDS embedded server modules Scale Metadata Capacity / Performance with DNE Server pairs Root dir b dir c dir d dir e Base MDS dir a dir b2 dir c2 dir d2 dir e2

ClusterStor Hardware and the Lustre System Meta Data and Management Servers 2U x 4 Servers Meta Data Target Seagate 2U24 JOBD 1) Where is file? 2) is at. Client 3) Single (3,072Kb) 4) is broken into block stripe segments (1,024Kb) Object Storage Server Seagate Embedded Application Server Object Storage Target Seagate 5U84 Storage Bay Bridge Enclosure 5a) block stripe 1 of 3 (1,024Kb) 5b) block stripe 2 of 3 (1,024Kb) 5c) block stripe 3 of 3 (1,024Kb)

Op/s Scaling MDS and DNEs MDS + 4 DNE Servers (2 ADUs) mdtest create/stat/del Mean of 5 iterations 600,000 500,000 400,000 mdtest scaling MDS + 4 DNEs 300,000 200,000 100,000 Mean Create Mean Stat Mean Remove 0 Seagate Confidential

Op/s Metadata High Availability MDT failover will ensure that the Lustre filesystem remains available in the face of MDS node failure Based on existing OSS pair failover model Failover is graceful, quick, and non-disruptive Failback is automatic and nondisruptive 140,000 120,000 100,000 80,000 60,000 40,000 20,000 0 High Availability and Performance Before Failover Failed over After Failover Mean Create Mean Stat Mean Remove Seagate Confidential

Green Machine: Environmentally-Aware Cold Storage Solution Space Light weight Small foot print Cold storage optimized design Cooling Zero heat emission Ambient cooling/no fans High operating temp. tolerant HDDs Power Green Dynamic power management Low power servers Aggressive TCO goals Recyclable chassis Reduced metal Responsible disposal of old chassis Lowest Operating Cost Reduced Carbon footprint Best for the Planet 15

Typical Use cases Retrieve content, photographs etc. from deep archive while maintaining consistent user experience Online pictures/social media store use cases Pictures >45 days in cold storage Retrieve MRIs/X-rays of a patient Use cases leveraging Tape-based solutions 16

Thank you!