IBM Power Systems IBM i High Availability Strategies
IBM Power Systems - IBM i High Availability Strategies & Planning for 2012 Part 3 of 3 - HA and DR Technology Positioning Eric Hess IBM Power Systems, Advanced Technical Support - Americas February 2009 1 Eric Hess, IBM 2008 © 2009 IBM Corporation
Agenda - High Availability Strategies & Planning for 2012 § Part 1 of 3 - Introduction to High Availability and Disaster Recovery o What are your availability requirements ? o IBM Power Systems IBM i availability solution § Part 2 of 3 - Introduction to IBM HA and DR Technologies o IBM Power Systems IBM i multi- server availability solutions o High availability and Disaster recovery o What’s new in IBM Availability? § Part 3 of 3 - HA and DR Technology Positioning (this module): o HA and DR Technology Positioning o Availability with POWER 6 and Virtualization o Pulling it all together - next steps The latest client-ready version of this presentation can be found in Tech. Docs: § IBM Internal: w 3. ibm. com/support/techdocs § External: www-03. ibm. com/support/techdocs/atsmastr. nsf/Web/Tech. Docs (or ibm. com and search ‘techdocs’) § Partners: Partner. World In Techdocs, use the search word ‘Hess’ to find this presentation 2 Eric Hess, IBM 2008 © 2009 IBM Corporation
**** This page for IBM and Business Partners Only **** Additional Events 1. Introduction to HA (HA Strategies & Planning for 2012 part 1 of 3) - Techtalk 2. Introduction to IBM HA and DR Technologies (HA Strategies & Planning for 2012 part 2 of 3) – Recorded (in Techdocs) 3. IBM Power HA Strategy – Techtalk (Thursday Feb 19) 4. IBM CBU Enhancements – Techtalk 5. Technology Positioning (HA Strategies & Planning for 2012 part 3 of 3) - Techtalk 6. IASP Enablement - Techtalk 7. Sizing Power. HA - Techtalk 8. Geographical Mirroring for Native IBM i applications - Recorded 9. Geographical Mirroring for Hosted and VIOS environments - Recorded 10. DS 6000/DS 8000 Copy Services HA Sizing and Considerations - Techtalk 11. i. Cluster Enhancements - Techtalk 12. What’s New in IBM i Availability Speakers for these sessions include: Bob Gintowt, Dave Chorba, Selwyn Dickey, Steve Finnes, Fred Robinson, Chuck Stupca, Mike Warkentin, and other experts on availability solutions and planning § Look for email from “STG Sales Education & Events Newsletter” for Techtalk dates § Techtalks: Smart. Zone for Power System Technical Training: o. IBM: http: //lt. be. ibm. com/smartzone/powertech o. Partner: http: //www. ibm. com/services/weblectures/smartzone/powertech 3 Eric Hess, IBM 2008 © 2009 IBM Corporation
Recap of Part 1 and Part 2 § IBM offers a growing number of single system and multi-system availability solutions, including software, OS and storage based technologies § Assess your use of single server availability options § All sizes of businesses should assess their business resiliency requirements o § Use methodologies such as Business Impact Analysis (BIA) to better understand exposures and what should be done about them IBM i and Power Systems offer 6 fundament technologies for business resiliency (native attached storage) 1. Logical Replication software 2. Switched IASP - Independent Auxiliary Storage Pool 3. IBM i Geographical Mirroring via OS 4. IBM System Storage LUN Level Switching 5. IBM System Storage Copy Services – simple copy 6. IBM System Storage Copy Services with IBM i § Some HA and DR technologies are unique to the IBM i Virtual Environments o § Understand how application HA / DR requirements interact with design efforts for a virtualized solution environment IBM i 6. 1 in conjunction with Power. HA offers dramatic HA/DR enhancements 10 Eric Hess, IBM 2008 © 2009 IBM Corporation
Power Systems IBM i Availability Solutions IBM has a long time investment strategy to deliver advanced i 5/OS and IBM i availability in order to meet varying client needs 1999 1989 § Logical Replication (Partner) § Journaling § Disk Mirroring and RAID-5 § UPS § HABP Cluster Partnership § OS/400 Clusters with Logical Replication § Remote Journaling § Redundant power supplies § Select concurrent maintenance § High speed, parallel tape 2009 § Power. HA for IBM i § Independent Auxiliary Storage Pools § IBM i Geographical Mirroring § LUN Level Switching § IBM Copy Services Integration* § IBM i. Cluster (Logical Replication) § Remote Journaling enhancements § RAID 5/6, Hot Spare, Disk Mirroring § Dual power connections § Increased server resiliency § Concurrent node maintenance § i 6. 1 Quisece for backups § Virtual tape § IBM Lab Services for advanced availability consulting § Education/Services Vouchers * IBM Storage DS 6000 and DS 8000 11 Eric Hess, IBM 2008 © 2009 IBM Corporation
Technology Positioning Multi-Server Availability Solutions 12 Eric Hess, IBM 2008 © 2009 IBM Corporation
Common Planning Items and Costs § DB Journaling o Storage space and performance for any new journaling activity § HA Journal Performance, OS Option 42 (optional) o Can be required for any availability need if journaling has an impact on long running workloads like nightly batch § SMAPP – System Managed Access Path Protection o Its use directly effects recovery time of all IASP and Storage copy based solutions § Efficient tape backup combinations o Save While Active o High speed tape o Concurrent or parallel tape designs o i 6. 1 Quiesce o Flash. Copy § Second server or LPAR related hardware § Related vendors technology education and services 14 Eric Hess, IBM 2008 © 2009 IBM Corporation
Power. HA and IBM i Switched IASPs Switched IASP planning items: LPAR-1 *SYSBAS o IASP LPAR-2 *SYSBAS 5. 3, 5. 4 or - 6. 1 Power. HA for i o Basic Switched IASPs between LPARs o o o Redundant LPARs provide protection from LPAR outage and some planned and unplanned outage requirements Single copy of disk cannot address disk outage or backups Combine with replication solutions Primary use is for switching between LPARs o POWER 6 is last generation to support HSL, avoid Switchable HSL towers between servers Is not supported by VIOS environments Some Single Points of Failure exist o Server SPOFs – use all applicable single server availability options o Storage - single copy of disk so use better disk protections schemes 15 Eric Hess, IBM 2008 © 2009 IBM Corporation
Switched IASPs (1 -2) Description: o o o IASP based Basic, local High Availability – stand alone or foundation for other technologies Management and automation provided by IBM clusters or Power. HA for i Business resiliency use: o o o Local High Availability: Some RTO and RPO benefits when Journaling and SMAPP used Long distance Disaster Recovery: No, local LPAR solution only Both local HA and remote DR: No, however, can be combined with replication Planned outages: Some. Online backups: No, may use Save-While-Active Workload balancing: No, since this is a single disk copy and cannot be shared Typical client environments: o Small and medium SMB, and as a foundation in Large Enterprise Environments o Native IBM i 5. 4 & 6. 1 o IBM i 5. 4 & 6. 1 Availability and recovery characteristics: o o o No IPL required - vary on the IASP to an active server/LPAR RTO: 5 minutes to 1 hour depending on design and practice RPO: Real time Hosted partitions for AIX, Linux and IBM i 6. 1* o Windows Integration * (IBM i 6. 1 host and client only) 16 Eric Hess, IBM 2008 © 2009 IBM Corporation
Switched IASPs (2 -2) Key advantages: o o Storage agnostic – works with any supported internal, external or combination Strategic foundation, allows addition of other replication technologies as needed Simple architecture means robust switching and simple to operate Application change management simplicity Key considerations: o o o LPAR only - not strategic between servers since POWER 6 is last to support HSL/RIO Storage and I/O properly designed for uptime and performance (system pool and IASP) Single copy of data, does not offer concurrent access to data from other server or LPAR Typical planning and timelines: o o Identify application and data objects to be replicated Application(s) and data migration proof of concept Storage I/O hardware planning and implementation as req. 3 -6 months from planning to implementation Typical costs considerations: o o Additional storage or I/O hardware (disk, drawer, 12 x, etc. ) HA Switchable Resources – OS option 41 Power. HA for i - 5761 -HAS if 6. 1 IASP proof of concept ( if required) Vendors and Resources o Power team to lead o o o architecture ATS and Techline Lab Services (fee) Benchmark Center GTS (fee) IASP and Power. HA Education 17 Eric Hess, IBM 2008 © 2009 IBM Corporation
IBM i and System Storage LUN Level Switching Source Target DS 8000 Production *SYSBAS HA LUN Level Switching planning items: o *SYSBAS LUN Group (active) IASP LUN Group (none) o o Flash. Copy IBM i Switchable LUNs o 5. 4 and 6. 1 Toolkit o Switchable LUNs using native attached DS 6000/DS 8000 (LPAR or Server level) o o Server level redundancy offers protection from server or LPAR outage Add Flash. Copy to address backup windows Combine with continuous replication solution for advanced HA or DR Does not support VIOS o o o Good DS 6000 or DS 8000 skills Proper DS 6000 or DS 8000 design for performance Proper IASP and *SYSBAS design Space Efficient Flash. Copy assessment Multipath used for critical workloads Lab Services consulting engagement for design, any proof of concept requirements, and implementation IBM i 5. 4 or i 6. 1 SMART Fibre IOA helps reduce costs 18 Eric Hess, IBM 2008 © 2009 IBM Corporation
LUN Level Switching(1 -2) Description: o o Function provided by IBM clustering Local High Availability – stand alone or combined with Switched IASPs Automation provided by IBM i clusters Application change management simplicity Business resiliency use: o o o Local High Availability: Yes, RTO and RPO benefits when Journaling and SMAPP used Long distance Disaster Recovery: No, side by side or campus distance Both local HA and remote DR: No. Planned outages: Most. Online backups: Yes, with Flash. Copy or SWA Workload balancing: No, since this is a single disk copy and cannot be shared Typical client environments: o Small and medium SMB, and foundation in Large Enterprise Availability and recovery characteristics: o o o No IPL required - vary on (switch) IASP to an active server/LPAR RTO: 5 minutes to 1 hour depending on design and practice RPO: Real time Environments o Native IBM i 5. 4 & 6. 1 o IBM i 5. 4 & 6. 1 Hosted partitions for AIX, Linux and IBM i 6. 1 o Windows Integration o DS 6000 and DS 8000 19 Eric Hess, IBM 2008 © 2009 IBM Corporation
LUN Level Switching(2 -2) Key advantages: o o Cost effective, single image local HA Simple architecture means robust switching and simple to operate Application change management simplicity Combines easily with DS 6000 and DS 8000 Copy Services and Toolkit Key considerations: o o o Storage and I/O properly designed for uptime and performance (system pool and IASP) Requires additional copies, plus external storage costs versus internal Not yet supported by Power. HA for i Typical planning and timelines: o o o Application(s) and data migration proof of concept or test Server and DS Hardware planning and implementation as req. 3 -12 months from planning to implementation Typical costs considerations: o o o DS 6000 or DS 8000 costs over internal storage Additional storage or I/O hardware (disk, HSL, 12 x, etc. ) HA Switchable Resources – OS option 41 IBM i Copy Services toolkit (Lab Services) IASP proof of concept (if required) Vendors and Resources o DS storage team to own o o o DS pre-sales, Disk Magic, S. E. Flash. Copy, etc. Power team to lead architecture ATS and Techline Lab Services (fee) Benchmark Center (fee) GTS (fee) IASP and System Storage Education 20 Eric Hess, IBM 2008 © 2009 IBM Corporation
Power. HA and XSM Geographic Mirroring Source Target IBM i Geographic Mirroring planning items: o Provide LPAR-1 *SYSBAS IASP Tape Backup 5. 3, 5. 4, or 6. 1 Power. HA for i Geographic Mirroring Overview o IBM replication at virtual storage management layer for consistent switchover o Synchronous replication between servers o Consistent replication of all data in IASP o Administrative Domain (5. 4 and 6. 1) or Logical replication used to replication key system data o Target can be used for backups to tape (6. 1 preferred), if suspending replication is acceptable (RPO exposure consideration) Eric Hess, IBM 2008 the extra CPU and memory required to support Geographic Mirroring environments o Proper disk design o Design for Performance o Avoid disk pool I/O sharing o Two or more network connections for performance and reliability o Improper LAN or WAN planning will create end user performance problems o Separate network path for cluster management o Assess network related performance impacts when exceeding campus distance 21 © 2009 IBM Corporation
Geographic Mirroring (1 -2) Description: o o OS based solution, integrated into IBM i 5. 4 and i 6. 1 Local High Availability (side by side or campus) Use stand alone or combined with Switched IASP and Flash. Copy Automation provided by IBM clusters or Power. HA for i Business resiliency use: o o o Local High Availability: Yes, RTO and RPO benefits when Journaling and SMAPP used Long distance Disaster Recovery: No, intended for campus or short city distance Both local HA and remote DR: No Planned outages: Some. Online backups: Maybe, SWA on source, or suspend replication and use target copy Workload balancing: No, since this is a single disk copy and cannot be shared Typical client environments: o Small and medium SMB, and basic needs in Large Enterprise Availability and recovery characteristics: o o o No IPL required - vary on (switch) IASP to an active server/LPAR RTO: 5 minutes to 1 hour depending on design and practice RPO: real time, if journaling and SMAPP are used Environments o Native IBM i 5. 4 & 6. 1 o IBM i Hosted partitions for IBM i o Windows Integration o VIOS IBM i client o VIOS-Blades IBM i client 22 Eric Hess, IBM 2008 © 2009 IBM Corporation
Geographic Mirroring (2 -2) Key advantages: o o o Storage agnostic – works with any supported internal, external, or virtual storage Simple architecture means robust switching and simple to operate Application change management simplicity Key considerations: o o o Synchronous replication technology limits distance CPU, Memory and Storage I/O must be designed to support Geographical Mirroring Detaching target for backups suspends replication - no valid HA environment during backup to tape and during the subsequent partial resynchronization. Typical planning and timelines: o o Application(s) and data migration proof of concept or test WAN - LAN Network design (redundancy, latency, bandwidth) Hardware planning and implementation as req. 3 -12 months from planning to implementation Typical costs considerations: o o o Additional storage or I/O hardware (disk, HSL, 12 x, etc. ) HA Switchable Resources – OS option 41 Power. HA for i - 5761 -HAS if 6. 1 WAN or LAN hardware for capacity and performance IASP proof of concept (if required) Vendors and Resources o Power team to lead o o o architecture ATS and Techline Lab Services (fee) Network expertise Benchmark Center (fee) GTS (fee) o IASP and Power. HA Education 23 Eric Hess, IBM 2008 © 2009 IBM Corporation
Point In Time Copy IBM System Storage Copy Services Software * Snapshots for backups, OS and application upgrade role-backs, BI and queries DS 8000 Flash. Copy and Space Efficient Flash. Copy DS 6000 For local High Availability Peer To Peer Remote Copy (continuous copy) Metro Mirror (synchronous) Global Mirror (asynchronous) Combos Global Copy (asynchronous copy utility) Distance for HA or DR Consistency Group For data migration, or coupled with Flash. Copy for DR For point in time snapshots at DR site Incremental Flash. Copy with Global Copy (point in time) Eric Hess, IBM 2008 * Describes current native attached DS 6000 or DS 8000 functions. Not all System Storage copy services are available in an IBM i Virtualized environment 24 © 2009 IBM Corporation
Full System Copy Services + Global Mirror DS 8000 Production Full System Backups Flash. Copy DS 8000 LUN Level Switching planning items: DR Global Mirror Flash. Copy CG Local Site o 5. 4, or 6. 1, Toolkit Remote Site Full System Copy Services - GM o DS 6000 or DS 8000 (non-VIOS) o o o Asynchronous replication between storage servers Consistent replication of all data in System ASP (ASP 1) Not a good HA solution - relies on IPL (boot) to access copy Requires target system configuration changes before copy can be used Full system replication results in Crash Consistent copy o o o Good DS 6000 or DS 8000 skills Proper DS 6000 or DS 8000 design for performance Space Efficient Flash. Copy assessment Proper Network assessment and design Multipath used for critical workloads Lab Services consulting engagement for design implementation IBM i 5. 4 or i 6. 1 SMARTIOA fiber helps reduce costs Note, Full System Metro Mirror also an option, but not shown 25 Eric Hess, IBM 2008 © 2009 IBM Corporation
Full System Copy Services + GM (1 -2) Description: o o DS 6000 and DS 8000 Clone copy of system and ASPs (data, OS, program products, temp work areas, etc) Function provided by SAN subsystem software Automation can optionally be provide by Lab Services ‘Copy Services Toolkit’ offering Business resiliency use: o o o Local High Availability: No. RPO and RPO benefits when Journaling and SMAPP used Long distance Disaster Recovery: Yes Both local HA and remote DR: No, however, can be combined with replication Planned outages: No. Online backups: Yes, with Flash. Copy and additional LPAR Workload balancing: Some, via offline point in time snapshots copies Typical client environments: o Medium SMB, and as for basic DR in Large Enterprise Availability and recovery characteristics: o o o IPL required – abnormal IPL likely - power off source to avoid RTO: 30 minutes to hours RPO: seconds - minutes based on consistency group design and network (Metro Mirror RPO = 0) Environments o Native IBM i 5. 4 & 6. 1 o IBM i Hosted partitions o Windows Integration o Does not support VIOS 26 Eric Hess, IBM 2008 © 2009 IBM Corporation
Full System Copy Services + GM (2 -2) Key advantages: o o Can meet moderate DR requirements No object out of sync issues Incorporates into cross platform storage DR solutions Application change management simplicity Key considerations: o o o Unable to address planned outages (changes on target cannot be propagated to source) Replicates temp data so it uses the largest amount of network capacity of all solutions Requires additional copies, plus external storage costs versus internal Typical planning and timelines: o o o Server and DS Hardware planning and implementation as req. WAN Network design (latency, bandwidth) 3 -12 months from planning to implementation Vendors and Resources o Storage team to own presales for Disk Magic, S. E. Flash. Copy, network planning, etc. Power team to lead architecture ATS and Techline Lab Services (fee) Benchmark Center (fee) GTS (fee) o IBM DS 8000 Education o o Typical costs considerations: o o o DS 6000 or DS 8000 costs over internal storage Extra copies of disk required for backups and Global Mirror DS 6000 or DS 8000 Copy Services software Lab Services Copy Services toolkit WAN or LAN hardware for capacity and performance o o o 27 Eric Hess, IBM 2008 © 2009 IBM Corporation
IASP + Copy Services + Metro Mirror DS 8000 Production DR and Test *SYSBAS IASP Backup Metro Mirror *SYSBAS IASP o 5. 4, or 6. 1, Toolkit, Power. HA Local Site Combination DS 6000 -DS 8000 Copy Services cluster solution example Power. HA or Copy Services Toolkit - MM o DS 6000 or DS 8000 (non-VIOS) o o Synchronous replication between storage servers Consistent replication of all data in System ASP (ASP 1) Advanced HA and Backup solution – no IPL required o Combine with Switched LUNs, other Copy Services, 6. 1 Quiesce , etc. o o o Flash. Copy *SYSBAS o LUN Level Switching planning items: Consistent replication of all data in IASP Administrative Domain (5. 4 and 6. 1) or Logical replication used to replication key system data o o o o Good DS 6000 or DS 8000 skills Proper DS 6000 or DS 8000 design for performance Space Efficient Flash. Copy assessment Proper Network assessment and design Multipath used for critical workloads Lab Services consulting engagement for design implementation Power. HA for i or Toolkit IBM i 5. 4 or i 6. 1 SMARTIOA fiber helps reduce costs 28 Eric Hess, IBM 2008 © 2009 IBM Corporation
IASP + Copy Services + Global Mirror + Combo Power. HA or Copy Services Toolkit Combination o Global Mirror examples: o o o Production *SYSBAS o LUN Group (active) DS 6000 or DS 8000 (non-VIOS) Synchronous or asynchronous replication between storage servers Advanced HA, DR and Backup solution o Local Switched LUNs for HA o Local Flash. Copy with 6. 1 Quisce for nightly backups o Global Mirror for advanced DR across long distances o Optional ‘D’ copy at remote site allows uninterrupted HA or DR test Many other combinations possible DS 8000 Consistency Group HA *SYSBAS LUN Group (none) IASP Global Mirror Production *SYSBAS IASP Flash. Copy Local Site Optional Golden (D) Copy DR Site 5. 4, or 6. 1, Toolkit, Power. HA Note: *SYSBAS disks are logical representations only, and may be disks located within the DS 6000 or DS 8000 29 Eric Hess, IBM 2008 © 2009 IBM Corporation
External Storage Copy Services Network Requirements § IASP based SAN copy solutions require less network bandwidth than full system SAN copy o Only IASP data is replicated o Examples show significant ongoing network cost savings using IASPs with asynchronous SAN copy services o Previous experience suggests that a 300 Mbs network between sites for full system replication can be reduced to 90 Mbs by using IASPs o IASP based cluster solutions provide the best RTO and RPO, with lower network costs www. systeminetwork. com Business Continuity with External Storage. Article ID: 20782 5250 Interactive Application SAP Application 30 Eric Hess, IBM 2008 © 2009 IBM Corporation
IASP + Copy Services + MM or GM (1 of 2) Description: o o DS 6000 and DS 8000 Storage level replication function provided by SAN subsystem software IASP based for more granular replication Automation provided by Power. HA or Toolkit Business resiliency use: o o o Local High Availability: Yes, RTO and RPO benefits when Journaling and SMAPP used Long distance Disaster Recovery: Yes. Both local HA and remote DR: Yes, using combinations Planned outages: Yes. Online backups: Yes, with Flash. Copy or Save-While-Active Workload balancing: Some, via offline point in time snapshots copies Typical client environments: o Medium SMB, Large Enterprise (usually too costly for small SMB) Availability and recovery characteristics: o o o No IPL required - vary on (switch) IASP to an active server/LPAR RTO: 5 minutes to hour depending on design and practice RPO: Real time, if journaling and SMAPP are used Environments o Native IBM i 5. 4 & 6. 1 o IBM i Hosted partitions o Windows Integration o Does not support VIOS 32 Eric Hess, IBM 2008 © 2009 IBM Corporation
IASP + Copy Services + MM or GM (2 of 2) Key advantages: o o o Can meet advanced HA, DR and backup requirements Fully integrated into IBM i via Power. HA or Toolkit – unique to IBM storage solutions No object out of sync issues, robust switching and simple to operate Incorporates into cross platform storage DR solutions Application change management simplicity Large number of installation and extensive IBM experience and skills Key considerations: o o o Costs – GM requires additional copies, plus external storage costs versus internal issues DS 6000 or DS 8000 skills as required Move to IASPs Typical planning and timelines: o o Application(s) and data migration proof of concept or test Identify application and data objects to be replicated 6 -18 months from planning to implementation Storage I/O hardware planning and implementation as req. Typical costs considerations: o o o DS 6000 or DS 8000 costs over internal storage Extra copies of disk required for backups and Global Mirror Power. HA for i and/or Copy Services software Lab Services Copy Services toolkit WAN or LAN hardware for capacity and performance Vendors and Resources o Storage team to own pre- o o o sales for Disk Magic, S. E. Flash. Copy, network planning, etc. Power team to lead architecture ATS and Techline Lab Services (fee) Benchmark Center (fee) GTS (fee) IASP, Power. HA and Storage Education 33 Eric Hess, IBM 2008 © 2009 IBM Corporation
Logical Replication planning items: o Are skills are in place to plan, implement, operate and maintain Production Backup o Do not use ‘replicate everything’ Data Copy design planning approach Tape Backup IBM i Remote Journaling o Poor application change management will cause out of Apply sync issues Management and Control Process 3 rd party software DB and Non-DB objects o For critical HA, limit use of target read-only copy 5. 4 or 6. 1 Typical Logical Replication Infrastructure o High I/O queries can delay the apply process Logical Replication Overview o Add 3 rd copy if needed o IBM Remote Journaling is used for replication of DB and other object types o Understand backup requirements o Vendor unique logic provides replication for other o A target copy backup with SWA objects types or other DB needs by itself does not replace the need to quiesce source server o Vendor provides management and control applications or users o Target copy is available for read-only workload o Higher tier products can support broadcast replication to 2 or more copies Source Target 34 Eric Hess, IBM 2008 © 2009 IBM Corporation
Solution Comparisons: Logical Replication (1 -2) Description: o o Add-on vendor software – does not require any optional OS features Journal (logging) based replication and management (journaling required) Replication by object Automation provided by ISV. Use of IBM base cluster in complex designs suggested Business resiliency use: o o o Local High Availability: Yes, with RPO benefits from synchronous Remote Journaling Long distance Disaster Recovery: Yes, via journaling or asynch Remote Journaling Both local HA and remote DR: Yes, multiple copies via cascade or broadcast options Planned outages: Most. Online backups: Yes, replication may continue, while apply process on B/U is halted Workload balancing: Some, via concurrent access and proper planning and controls Typical client environments: o Small SMB, medium SMB, and Large Enterprise Availability and recovery characteristics: o o o Environments o All environments, including VIOS No IPL required. Wait for apply process on B/U to complete. RTO: 15 minutes to small hours depending on design, practice RPO: near real time to minutes based on date type, activity, distance 35 Eric Hess, IBM 2008 © 2009 IBM Corporation
Solution Comparisons: Logical Replication (2 -2) Key advantages: o o Storage agnostic – works with any supported internal, external or combination Only solution to offer concurrent access to data on target Able to support simultaneous replication to local HA and remote DR copies Smaller network bandwidth required than other solutions Key considerations: o o o Tends to require monitoring, operations and development team resources Target RPO can fall behind during write intensive batch jobs or during backups on target Some applications can be problematic in keeping data synchronized Typical planning and timelines: o o Journal and vendor product planning Identify application and data objects to be replicated 3 -12 months from planning and implementation Must be fully tested on a regular basis (e. g. 3 -6 months) Typical costs considerations: o o o Vendor software and options Operations resource to monitor daily Proper network Education and services for proper design and implementation HA Journal Performance Option 42 might be needed Vendors and Resources IBM i. Cluster o www. ibm. com/systems/p ower/software/availability Vision Solutions o www. visionsolutions. com Maximum Availability no. MAX o www. maxava. com Trader’s Quick-Edd o www. quick-softwareline. com 36 Eric Hess, IBM 2008 © 2009 IBM Corporation
IBM i 6. 1 Power. HA § Provides choices in design, implementation and management of IBM i 6. 1 and SAN Storage IASP Cluster based replication technologies for HA and DR Power. HA for i (5761 -HAS) a. k. a HASM Cross Site Mirroring (XSM) “HA Switchable Resources” - option 41 – IBM i Geographic Mirroring • IBM i storage mgmt page level replication • Synchronous • Any internal or external storage Metro Mirror (SAN)* • SAN hardware replication • Synchronous • IBM SAN storage only Global Mirror (SAN)* • SAN hardware replication • Asynchronous • IBM SAN storage only Flash. Copy (SAN)* • SAN hardware replication • Point in time • IBM SAN storage only * This new support and capability added to IBM i and XSM with Power. HA and 6. 1, IBM SAN Only See www. redbooks. ibm. com: Implementing Power. HA for IBM I - SG 24 -7405 -00 (Nov 2008) 37 Eric Hess, IBM 2008 © 2009 IBM Corporation
Copy Services toolkit with IBM i 6. 1 Power. HA for i Functions Power. HA for i Toolkit ( stand alone environment ) IBM i operating system product (Power. HA LPP ) Yes No Requires Power. HA LPP N/A Yes- via tool kit release which supports Power. Ha Cluster switch/failover Yes Start/Stop Flash. Copy Yes Health check provided Yes Resets multipath for disk registration on switchover Yes Automates cluster node starting for switchover/Flash. Copy No Yes Supports IBM i 5. 4 No Yes Automate queise function No Yes Provides DS scripts for replication Yes Full System or non IBM i replication management No Yes Provides CL programming interface for IBM storage No Yes Eric Hess, IBM 2008 39 © 2009 IBM Corporation
IBM i Availability Technologies Virtualized Environments 40 Eric Hess, IBM 2008 © 2009 IBM Corporation
HA and DR Technologies for IBM i Virtual Environments § VIOS (hosting IBM i partition) - Power servers and IBM Power blades o o o IBM and ISV Logical Replication IBM i XSM Geographical Mirroring (Power. HA) with DS 4000 or SVC IBM Storage Copy Services (full copy, manual process only) o DS 4000 Point in time copy: Flash. Copy or Volume Copy o DS 4000 Continuous copy: Synchronous Enhanced Remote Mirroring (Metro Mirror) o SVC Copy Services (SAN Volume Controller) § Notes o o o Power. HA for i does not support copy services of DS 4000 or SVC Lab Services ‘Copy Services Toolkit’ does not support copy services of DS 4000 or SVC Full-system Flash. Copy and Volume. Copy requires that the production IBM i logical partition (LPAR) be powered off or in a restricted state New 6. 1 ‘Quiesce’ function not supported with DS 4000 or SVC copy services See planning documentation for other planning items and restrictions IBM i and Midrange External Storage • www. redbooks. ibm. com/abstracts/sg 247668. html? Open • IBM i Virtualization and DS 4000 Read-me First (October 20 th, 2008) • www-03. ibm. com/systems/resources/systems_i_os_i_virtualization_and_ds 4000_readme. pdf 41 Eric Hess, IBM 2008 © 2009 IBM Corporation
Choosing a Solution Summary 42 Eric Hess, IBM 2008 © 2009 IBM Corporation
Selection Criteria § Generalizations (e. g. categorizing into major problem sets) can distort the answer § Should use as a guide to potentially rule out certain solution types § In practice, consider all major and supporting decision factors such as: o Up time requirements o Recovery time objective o Recovery point objective o Resilience requirements o Concurrent access requirements o Geographic dispersion requirements o Tolerance for end user disruption o Outage type coverage o Experience, skills and resources o Cost and budget o Service and support i 5/OS High Availability Clusters: Data Resilience Solutions Document Version 1. 0 Version date: October 22, 2004 Authors: Steven J Finnes Robert T Gintowt Michael J Snyder © Copyright IBM Corporation 2004. All rights reserved. See Redpaper redp 0888: www. redbooks. ibm. com 43 Eric Hess, IBM 2008 © 2009 IBM Corporation
**** This page for IBM and Business Partners Only **** Additional Events 1. Introduction to HA (HA Strategies & Planning for 2012 part 1 of 3) - Techtalk 2. Introduction to IBM HA and DR Technologies (HA Strategies & Planning for 2012 part 2 of 3) – Recorded (in Techdocs) 3. IBM Power HA Strategy – Techtalk (Thursday Feb 19) 4. IBM CBU Enhancements – Techtalk 5. Technology Positioning (HA Strategies & Planning for 2012 part 3 of 3) - Techtalk 6. IASP Enablement - Techtalk 7. Sizing Power. HA - Techtalk 8. Geographical Mirroring for Native IBM i applications - Recorded 9. Geographical Mirroring for Hosted and VIOS environments - Recorded 10. DS 6000/DS 8000 Copy Services HA Sizing and Considerations - Techtalk 11. i. Cluster Enhancements - Techtalk 12. What’s New in IBM i Availability Speakers for these sessions include: Bob Gintowt, Dave Chorba, Selwyn Dickey, Steve Finnes, Fred Robinson, Chuck Stupca, Mike Warkentin, and other experts on availability solutions and planning § Look for email from “STG Sales Education & Events Newsletter” for Techtalk dates § Techtalks: Smart. Zone for Power System Technical Training: o. IBM: http: //lt. be. ibm. com/smartzone/powertech o. Partner: http: //www. ibm. com/services/weblectures/smartzone/powertech 44 Eric Hess, IBM 2008 © 2009 IBM Corporation
Thank You IBM i High Availability Strategies & Planning for 2012 Part 3 of 3 Eric Hess, texas@us. ibm. com IBM 2009 Technical Sales Support Americas 45 Eric Hess, IBM 2008 © 2009 IBM Corporation
Trademarks and Disclaimers Copyright IBM Corporation 1994 -2009. All rights reserved. References in this document to IBM products or services do not imply that IBM intends to make them available in every country. Trademarks of International Business Machines Corporation in the United States, other countries, or both can be found on the World Wide Web at http: //www. ibm. com/legal/copytrade. shtml. Intel, Intel logo, Intel Inside logo, Intel Centrino logo, Celeron, Intel Xeon, Intel Speed. Step, Itanium, and Pentium are trademarks or registered trademarks of Intel Corporation or its subsidiaries in the United States and other countries. Linux is a registered trademark of Linus Torvalds in the United States, other countries, or both. Microsoft, Windows NT, and the Windows logo are trademarks of Microsoft Corporation in the United States, other countries, or both. UNIX is a registered trademark of The Open Group in the United States and other countries. Java and all Java-based trademarks are trademarks of Sun Microsystems, Inc. in the United States, other countries, or both. Other company, product, or service names may be trademarks or service marks of others. Information is provided "AS IS" without warranty of any kind. The customer examples described are presented as illustrations of how those customers have used IBM products and the results they may have achieved. Actual environmental costs and performance characteristics may vary by customer. Information concerning non-IBM products was obtained from a supplier of these products, published announcement material, or other publicly available sources and does not constitute an endorsement of such products by IBM. Sources for non-IBM list prices and performance numbers are taken from publicly available information, including vendor announcements and vendor worldwide homepages. IBM has not tested these products and cannot confirm the accuracy of performance, capability, or any other claims related to non-IBM products. Questions on the capability of non-IBM products should be addressed to the supplier of those products. All statements regarding IBM future direction and intent are subject to change or withdrawal without notice, and represent goals and objectives only. Some information addresses anticipated future capabilities. Such information is not intended as a definitive statement of a commitment to specific levels of performance, function or delivery schedules with respect to any future products. Such commitments are only made in IBM product announcements. The information is presented here to communicate IBM's current investment and development activities as a good faith effort to help with our customers' future planning. Performance is based on measurements and projections using standard IBM benchmarks in a controlled environment. The actual throughput or performance that any user will experience will vary depending upon considerations such as the amount of multiprogramming in the user's job stream, the I/O configuration, the storage configuration, and the workload processed. Therefore, no assurance can be given that an individual user will achieve throughput or performance improvements equivalent to the ratios stated here. Photographs shown may be engineering prototypes. Changes may be incorporated in production models. 46 Eric Hess, IBM 2008 © 2009 IBM Corporation
- Slides: 37