Sql server 2012_sp1_01_of_12_required 9s_and_protection_level300_dark_deck

965 views
862 views

Published on

Published in: Technology
0 Comments
1 Like
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
965
On SlideShare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
33
Comments
0
Likes
1
Embeds 0
No embeds

No notes for slide

Sql server 2012_sp1_01_of_12_required 9s_and_protection_level300_dark_deck

  1. 1. REQUIRED 9S & PROTECTIONMission Critical Confidence
  2. 2. THE FANTASTIC 12 OF 2012 1 2 3 4 5 6 7 8 9 10 11 12
  3. 3. SQL SERVER DELIVERS REQUIRED 9SMAXIMIZE UPTIME FOR MISSION CRITICAL APPLICATIONS
  4. 4. ALWAYSON AVAILABILITY GROUPS
  5. 5. EXAMPLE HA/DR TOPOLOGY
  6. 6. BENEFITS*Database backup is copy-only type, Log backup is regular. See http://msdn.microsoft.com/en-us/library/hh245119.aspx
  7. 7. Seamless connectivity to primary through ListenerVirtual Network Name 1 Virtual IP per subnetCluster resource Part of AG resource group Fails over with AG (binds to network interface)
  8. 8. AVAILABILITY GROUP REPLACES DB MIRRORING
  9. 9. NEW TOPOLOGY BENEFITS
  10. 10. CONSIDERATIONS FOR AVAILABILITY GROUPS
  11. 11. CLUSTER CONSIDERATIONSCluster Members must be in same Windows domain or trusted domainsCluster needs quorum to avoid split brainThe number of voting members determines the cluster tolerance to failureConfiguring cluster quorum:1. Select cluster members to vote Primary Auto failover target Other nodes in local data center (not necessarily hosting SQL Server instances)2. Select quorum type: Odd number of votes, use “Node Majority” Even number of votes a) Add an additional node and use “Node Majority” b) Add a file share and use “Node and File Share Majority”
  12. 12. EXAMPLE TOPOLOGY – CLUSTER QUORUM CONFIGURATION
  13. 13. ALWAYS ON ACTIVE SECONDARIES
  14. 14. ACTIVE SECONDARY – READABLE SECONDARY
  15. 15. OFFLOADING READ WORKLOADS
  16. 16. OFFLOADING READ WORKLOADS
  17. 17. OFFLOADING READ WORKLOADS
  18. 18. CONFIGURING SECONDARY AS READABLE
  19. 19. CLIENT CONNECTIVITYRead / Write WorkloadConnecting using AG ListenerConnection using FAILOVER_PARTNER (if connectionstring of existing applications can’t be changed)Read Only WorkloadConnection using VNN andApplicationIntent=ReadOnlyConnection to the secondary instance directlyReadOnly RoutingMulti subnet failover scenario:New client libraries => MultiSubnetFailover=TrueOld client libraries configure appropriate clientconnection timeout
  20. 20. CONFIGURING ROUTING OF READ-ONLY CONNECTIONS
  21. 21. CONFIGURING ROUTING OF READ-ONLY CONNECTIONS
  22. 22. READABLE SECONDARY – DATA LATENCY
  23. 23. LOG SYNCHRONIZATION TO SECONDARY
  24. 24. BACKUP* CAPABILITIES*Database backup is copy-only type, Log backup is regular. See http://msdn.microsoft.com/en-us/library/hh245119.aspx
  25. 25. CREATING BACKUPS
  26. 26. SQL SERVER FAILOVER CLUSTERING COMPONENTS
  27. 27. ALWAYSON MULTI-SITE FAILOVER CLUSTER INSTANCEHA & DR SOLUTIONProvide High Availability at the Instance LevelUnit of failover = SQL server instanceMaintain same virtual network name after failover. Clients re-connect to same nameInstance restart requires database to go through recoveryProvide Disaster Recovery at the Instance LevelProvide Disaster Recovery protection from site failure: be it network, power, infrastructure or other sitedisasters.Require storage based replication technology and networking considerationsMulti-subnet support
  28. 28. ALWAYSON FAILOVER CLUSTER INSTANCES
  29. 29. MULTI-SUBNET CLUSTERING
  30. 30. ALWAYSON FAILOVER CLUSTER INSTANCEFailover Condition Policy
  31. 31. CONFIGURE NEW IP ADDRESS - ADDNODE
  32. 32. MULTI-SUBNET CLUSTERING REQUIREMENTSSQL Server SKU: Enterprise EditionOS version: Windows Server 2008 R2 +SAN replication for cross-site DR (e.g. EMC SRDF)Single AD domain for all nodes
  33. 33. BEST PRACTICES – 1. QUORUM MODEEven number of nodes
  34. 34. BEST PRACTICES – 2. HEARTBEAT SETTINGSDefault ValueFrequency is once per 1 secondIf miss 5 heartbeat then initial failoverTune the setting for cross subnet heartbeatCrossSubnetDelay can be up to 4 secondsCrossSubnetThreshold can be up to 10http://technet.microsoft.com/en-us/library/dd197562(WS.10).aspx
  35. 35. BEST PRACTICES – 3. SAN REPLICATIONUse block replicationFile and Application replication is not supportedPreserve block size and write-order to prevent data corruptionChoose Replication Mode according to the network latencySync if Network latency < 10ms (regional DR)Async if Network latency > 10ms (Geo DR)http://technet.microsoft.com/en-us/library/dd197562(WS.10).aspx
  36. 36. BEST PRACTICES – 4. DNS SETTINGSParallel Client Connection (new client connection stringparameters)sqlcmd.exe -M switchADO .Net, native clients MultiSubnetFailover=trueOLE DB, legacy clients Connection Timeout = 21* (# of IPs)Reduce DNS replication delayDNS/AD inter-site replication schedule, by default is 180 minutesSet replication frequency to 15 minutes (minimum value)
  37. 37. TEMPDB ON LOCAL DISKWhy We Enable This?Tempdb access occupies large % of SAN I/OFast local HDD/SSD becomes standard ServerconfigurationBenefitsBetter overall performanceCost savingImportant Note!Ensure that tempdb local paths are available toSQL Service on all the nodes
  38. 38. PREDICTABLE RECOVERY TIME –INDIRECT CHECKPOINTSPreviously checkpoints mode (automatic vs. fixed intervals)Variance in failover timeVariance in IO loadNew in SQL Server 2012:Background checkpointing over timeSmoother IO loadMore predicable failover timesConfigurable per databaseOff-by-default for backcompat
  39. 39. INDIRECT CHECKPOINT: DIRTY PAGES
  40. 40. FLEXIBLE FAILOVER POLICY –Failure Condition Levels
  41. 41. WHY WINDOWS SERVER CORE FOR SQL SERVER 2012?
  42. 42. WINDOWS SERVER 2012 INSTALLATION
  43. 43. WINDOWS TIPS FOR SQL SERVER DEPLOYMENTS http://tinyurl.com/cn7ttsm http://tinyurl.com/28h883x http://tinyurl.com/729mr58
  44. 44. .NET FRAMEWORK & SERVER CORE http://tinyurl.com/d2uapdp) o How To: http://tinyurl.com/6nmueob (DISM) or http://tinyurl.com/882un6y (PS) .NET 3.51 (incl. How to Install .NET 4.0 How to Install 2.0)Windows Server In Band Server Manager Out of Band SQL media,2008 R2 (GUI), DISM, download PowerShell Install 4.0.2 or later for AlwaysOn enhancementsWindows Server In Band Server Manager In Band Already enabled2012 (GUI), DISM, (.NET 4.5) PowerShell
  45. 45. .NET FRAMEWORK & SERVER CORE DISM (Full) DISM (Core) PS (Full) PS (Core)Windows Server NetFx3 • NetFx2-ServerCore Net-Framework- Use DISM (will be2008 R2 • NetFx2-ServerCore- Core enabled with WOW64 remote • NetFx3-ServerCore management) • NetFx3-ServerCore- WOW64Windows Server Net- Net-Framework-Core Net-Framework- Net-Framework-2012 Framework- Core Core Core
  46. 46. SQL SERVER 2012 INSTALL BEHAVIOR ON FULL GUI OS• Have to manually install .NET Framework 3.51• Will install .NET 4.0 (if necessary)
  47. 47. SQL SERVER 2012 INSTALL BEHAVIOR ON SERVER CORE• Will enable/install .NET Framework (3.51 and 4.0) if not already done
  48. 48. WSFC & SERVER CORE
  49. 49. WSFC & SERVER CORE DISM (Full) DISM (Core) PS (Full) PS (Core)Windows Server FailoverCluster- • FailoverClust Failover- • Failover-2008 R2 FullServer er-Core Clustering Clustering • FailoverClust • WoW64- er-Core- FailoverClust WOW64 erWindows Server FailoverCluster- FailoverCluster- Failover- Failover-2012 FullServer FullServer Clustering Clustering
  50. 50. WSFC & SERVER CORE DISM PS PowerShell cmdlets • FailoverCluster- RSAT-Clustering- AdminPak PowerShell • FailoverCluster- PowerShell Old Automation FailoverCluster- RSAT-Clustering- Server AutomationServer AutomationServer (deprecated) Old cluster.exe FailoverCluster- RSAT-Clustering- (deprecated) CmdInterface CmdInterface
  51. 51. ACTIVATION WITHOUT INTERNET CONNECTIVITY
  52. 52. INSTALLING SQL SERVER 2012 ON SERVER CORE o http://tinyurl.com/bmozrgw, http://tinyurl.com/cnft2fm
  53. 53. INSTALLING SQL SERVER 2012 ON SERVER CORE http://tinyurl.com/7jovob2
  54. 54. INSTALLING SQL SERVER 2012 ON SERVER CORE
  55. 55. ADDITIONAL INFORMATION The Microsoft SQL Server support policy for Microsoft Clustering http://support.microsoft.com/kb/327518• KB943984 – The Microsoft Support Policy for Windows Server 2008 or Windows Server 2008 R2 Failover Clusters http://support.microsoft.com/kb/943984• KB2545685 – Recommended hotfixes and updates for Windows Server 2008 R2 SP1 Failover Clusters http://tinyurl.com/7r8flpj• SQL Server 2012 Bare Metal Workshop Series http://tinyurl.com/7l66mks• Server Core team blog http://tinyurl.com/7e4rq8o
  56. 56. NEXT STEPSSQL Server 2012 Case Studies:http://www.microsoft.com/casestudies/Case_Study_Advanced_Search.aspx (Search onSQL Technologies)SQL Server 2012 Hands On Labs:http://www.microsoft.com/sqlserver/en/us/learning-center/virtual-labs.aspxSQL Server 2012 Certification:http://www.microsoft.com/learning/en/us/certification/cert-sql-server.aspxSQL Server 2012 Best Practices:http://technet.microsoft.com/en-us/sqlserver/bb671430
  57. 57. BETTER AVAILABILITYWINDOWS SERVER 2012 Significantly decrease downtime Enhanced Live Migration Cluster-Aware Updating (CAU) Windows Server Core Dynamic Quorum in Cluster Hyper-V Replica
  58. 58. BETTER AVAILABILITY WITH WINDOWS SERVERTECHNICAL BENEFITS
  59. 59. KEY LEARNINGS FROM EARLY CUSTOMER DEPLOYMENTS
  60. 60. MULTI-SITE FAILOVER CLUSTER INSTANCE FOR HA AND DR
  61. 61. MULTI-SITE FAILOVER CLUSTER INSTANCE FOR HA AND DRFileshare Witness in 3rd Data Center
  62. 62. MULTI-SITE FAILOVER CLUSTER INSTANCEKey ElementsA single SQL Server failover cluster instance (FCI) providing HA as well as DRSpanning across multiple sites (usually multiple subnets as well)Key components:Storage Storage level replication Storage failover mechanism Provided by the storage vendor Work with your storage vendor to get the appropriate software and best practicesNetwork Multi-subnet support in SQL Server configuration and engine Key improvement in SQL Server 2012 IP address OR dependency set within SQL Server setup SQL Engine skips binding to any IP’s which are not online at start-up RegisterAllProvidersIP for Network Name improves application failover time
  63. 63. MULTI-SITE FAILOVER CLUSTER INSTANCEDeployment ConsiderationsStorage ValidationStorage Validation Check Requirement is relaxed due to make-up of multi-site storageinfrastructure (but still get the pop-up!)Multi-site FCI Solution does not require passing the storage validation tests, to besupported. http://support.microsoft.com/kb/943984Appropriate Quorum ModelValidation suggests “Node and Disk Majority” which can be ignoredConsider “Node and File Share Majority” or “Node Majority” based on number of nodes
  64. 64. MULTI-SITE FAILOVER CLUSTER INSTANCEDeployment ConsiderationsTEMPDB on Local DiskNot specific to “multi-site” FCIs, but has some great positive side effects for “multi-site”scenarios Enables use of local storage for TEMPDB Can use solid state storage to improve performance of TEMPDB-heavy workloads Saves money on storage replication cost Reduces cross-data center storage replication traffic
  65. 65. AVAILABILITY GROUPS FOR HA AND DR
  66. 66. AVAILABILITY GROUPS FOR HA AND DRDeployment ConsiderationsPre-requisites:BOL Link: Prerequisites, Restrictions, and Recommendations for AlwaysOn Availability Groups(SQL Server)Unit of FailoverGroup of databases – not the instanceConsider Contained Database for containing logins for failoverFor jobs and other objects outside the database, simple customization neededConsiderations for Replacing Log ShippingNo delayed apply on the secondaryRemoving log shipping means the regular log backup job is removed Need to re-establish periodic log backup (essential for truncating the log)
  67. 67. CUSTOMER EXAMPLE: MSIT SAP ERPBusinessCritical to the Microsoft businessUtilized for finance, payment, and procurement among other functions, for Microsoft.Utilized across 89 countries , supporting 92,000 employeesApplication~100K internal web users plus external web usersUp to 1500+ concurrent users2 million dialog steps per business day240K+ batch job executions per month6.6 terabytes of compressed data0.8 seconds user response time99.995% availabilityhttp://www.microsoft.com/casestudies/Microsoft-SQL-Server-2012-Enterprise/Microsoft-IT/Microsoft-Ensures-Smooth-Operation-of-ERP-System-and-Cuts-Disaster-Recovery-Time/710000000493
  68. 68. CUSTOMER EXAMPLE: MSIT SAP ERP
  69. 69. SPECIAL CASE: AUTOMATIC FAILOVER FOR DRUse of 3rd Data Center
  70. 70. CUSTOMER EXAMPLE: BWIN.PARTYBusinessOnline gaming and gambling.Real money handling system for bwin.party.Authoritative system for Responsible Gaming Limitations.Application RequirementsMultiple databases, and multiple availability groups in the topology.Plan for the worst case scenario: Loss of a complete datacenter.Must still be able to do maintenance during the worst case.RPO: Zero data loss.RTO: 10 seconds or less.>99.99% availability in the last number of years.>99.99% availability even with maintenance.http://www.microsoft.com/casestudies/Microsoft-SQL-Server-2012/bwin.party/Company-Cuts-Reporting-Time-by-up-to-99-Percent-to-3-Seconds-and-Boosts-Scalability/710000000087
  71. 71. CUSTOMER EXAMPLE: BWIN.PARTY
  72. 72. FAILOVER CLUSTER INSTANCE (HA), AVAILABILITY GROUP (DR)
  73. 73. FAILOVER CLUSTER INSTANCE (HA), AVAILABILITY GROUP (DR)Deployment ConsiderationsPre-requisites:Windows Server Service packs / QFEs: Asymmetric Storage Windows Server 2008 with http://support.microsoft.com/kb/976097 OR, Windows Server 2008 R2 SP1 Node Votes: http://support.microsoft.com/kb/2494036 Validate disk test QFE: http://support.microsoft.com/kb/2531907Prerequisites, Restrictions, and Recommendations for AlwaysOn Availability Groups (SQL Server)Different units of failover for HA and DRInstance-level failover for local HA (FCI)Group of databases (AG) for DRAG Failover ModeIn FCI+AG configuration, FCI provides automatic failover, and AG provides manual failover
  74. 74. FAILOVER CLUSTER INSTANCE (HA), AVAILABILITY GROUP (DR)Deployment ConsiderationsAsymmetric StorageKey concept behind this architectureNew Windows Server Failover Clustering capability introduced in: Windows Server 2008 R2 SP1 Windows Server 2008 with QFESymmetric storage = a cluster disk that is shared between all the WSFC nodesAsymmetric storage = a cluster disk that is shared between a subset of nodesInstance NamingEach FCI within the WSFC needs to have a different instance nameDatabase File Paths(recommended) use identical drive letters for the disks for each FCI(recommended) use identical file paths for data and log files for each FCI
  75. 75. CUSTOMER EXAMPLE: SERVICEU CORP(PART OF ACTIVE NETWORK)BusinessServiceU provides web-based online scheduling, event management, payment processing, and other services tocustomers in 15 countries and 50 united states.Application Requirements99.99% uptime (maximum allowable downtime of 52 minutes per year including scheduled maintenance).Security – Level 1 PCI Service Provider.Performance.http://www.microsoft.com/casestudies/Microsoft-SQL-Server-2012/ServiceU/Online-Company-Reduces-Downtime-and-Helps-Its-Customers-to-Improve-Service/4000011506
  76. 76. CUSTOMER EXAMPLE: SERVICEU CORP(PART OF ACTIVE NETWORK)
  77. 77. QUORUM CONSIDERATIONS
  78. 78. QUORUM MODEL AND NODE VOTESAvailability Groups for HA and DR
  79. 79. QUORUM MODEL AND NODE VOTESFailover Cluster Instance (HA), Availability Group (DR)
  80. 80. QUORUM MODEL AND NODE VOTESHow to set / view
  81. 81. MIGRATION PLANNING AND KEY CONSIDERATIONSHardware: new hardware, reuse existing hardware?Windows Server: Upgrading the OS as well?Windows Clustering: involve the Windows System Administration team and the networkingteamQuorum considerations across multiple data centersCluster network communication across multiple data centersStages: migrate the whole configuration at once, or migrate the DR afterwards?Application connection string changeIf the secondary is a Failover Cluster Instance: needs re-seedingUninstall existing SQL FCI on the DR site nodesDestroy existing WSFC at the DR siteJoin DR site nodes to Windows cluster created in the primary data centerRe-install SQL FCI on DR nodesBackup database(s) from primary, and Restore on the secondary
  82. 82. SQL SERVER 2012 ALWAYSON CUSTOMER EXAMPLES
  83. 83. BETONSOFTBusinessInternational gaming software provider.Software development of online casino gaming platform.Shared platform for 5 operators (15 brands).Over 110 games, including: Slots, Roulette, Video Poker, Blackjack, Craps.ApplicationGame play data is generated rapidly and used infrequently for “Playback”.Game state and game history.Realtime operator metrics (less than 2 seconds).http://www.microsoft.com/casestudies/Microsoft-SQL-Server-2012-Enterprise/BetOnSoft/Online-Gaming-Firm-Implements-Real-Time-Analytics-and-Scales-for-Planned-Growth/710000000130
  84. 84. BETONSOFT - ARCHITECTUREInfrastructure1 data center. A separate DR data center is being planned.HardwareDell R810 servers with 4 X 8-core, 256 GB RAM.Fusion IO cards for storage: Two cards on each server (1.2 TB, 640 GB).AlwaysOn HA / DR Solution
  85. 85. BETONSOFT – HA ARCHITECTURE
  86. 86. SUMMARY: BETONSOFTGains and LearningsModified application to remove distributed transactions Distributed transactions were broken into two separate transactions and the application handles consistencyAG synchronous replica has very minimal impact on performance and throughput of theapplication
  87. 87. SOFTBANK (AGOOP) - WEBGISBusinessSoftbank One of the largest mobile phone carrier in Japan Also runs broadband and international phone lineAGOOP Subsidiary of Softbank Specialized to GIS (Geographic Information Systems) Provide services for Softbank and external customersApplicationWebGIS is the GIS for internal Softbank mobile for analysis to help make plans for their future businessHA/DR requirements: RPO=0 RTO=1 hour (unplanned), Up to 1/2 day for maintenance (planned)http://www.microsoft.com/casestudies/Microsoft-SQL-Server-2012/SoftBank-Mobile-Corporation/Japanese-Cellular-Provider-Gains-High-Availability-and-Improves-Performance/710000000498
  88. 88. SOFTBANK (AGOOP) - ARCHITECTUREInfrastructureSingle data centerHardwareDell PowerEdge R710 servers, 2-Socket (16 logical cores total)Storage: Powervault MD3000/MD1000AlwaysOn HA / DR Solution
  89. 89. SOFTBANK (AGOOP) – HA ARCHITECTURE
  90. 90. SUMMARY: SOFTBANK (AGOOP)Gains and LearningsAlwaysOn Availability groups provide a mechanism for High Availability and also for read-scale.Easy to scale out reads using AlwaysOn availability group and readable secondary Previously had multiple data tiers with transactional replication, with AlwaysOn integrated solution. With AlwaysOn maintenance overhead in managing different databases in the replication architecture is gone. Allowed analysts to create reports on data without impact to the primary workload.
  91. 91. DELL – TRANSIENTCARTBusinessDell.com is one of the busiest websites in the world. Over 10,000 concurrent users during peaks.Most sales and support processes are executed online through Dell.comDell did $61 billion US in revenue for 2011.ApplicationTransientCart is a shopping cart for Dell.com Maintains a user’s configuration for 45 minutes Allows for users to quickly retrieve and store items without re-configuring every time. Very heavy OLTP write application. Multiple “carts” to handle load, consolidated into 1 for SQL Server 2012Business requires 4 9’s uptime (including planned and unplanned downtime)http://www.microsoft.com/casestudies/Microsoft-SQL-Server-2012-Enterprise/Dell/Dell-Speeds-Recovery-Time-by-99-Percent-Boosts-Service-and-Efficiency-and-Cuts-Costs/710000000488
  92. 92. DELL - ARCHITECTUREInfrastructure2 data centers, distance of 1.5 miles apartDifferent subnets in each data centerHardwareDell PowerEdge R710 servers, 2-socket/8-core and 24Gb of RAMCompellent SAN storageAlwaysOn HA / DR Solution
  93. 93. DELL (TRANSIENTCART) - HA/DR ARCHITECTURE
  94. 94. SUMMARY: DELL TRANSIENTCARTGains and LearningsAble to achieve failovers in less than 1 minute: RTO < 1 min.Synchronous AG’s allow for RPO = 0.Meeting SLA of 52 min downtime (99.99%), so far zero perceived downtime. Upgrade from RC to RTM had no application downtime impactListener (VNN) helped avoid extended manual process they previous had to go through ina DR scenario.Have a coordinated plan to recover from disaster. Worked with different groups to buildthe system and maintain scripts and process to recover in case of failure.In Process: Backup from secondary (waiting on support from 3rd party) and readablesecondary capabilities.
  95. 95. MICROSOFT IT SAP ERP DEPLOYMENTBusinessCritical to the Microsoft businessUtilized for finance, payment, and procurement among other functions, for Microsoft.Utilized across 89 countries , supporting 92,000 employeesApplication~100K internal web users plus external web usersUp to 1500+ concurrent users2 million dialog steps per business day240K+ batch job executions per month6.6 terabytes of compressed data0.8 seconds user response time99.995% availabilityhttp://www.microsoft.com/casestudies/Microsoft-SQL-Server-2012-Enterprise/Microsoft-IT/Microsoft-Ensures-Smooth-Operation-of-ERP-System-and-Cuts-Disaster-Recovery-Time/710000000493
  96. 96. MSIT SAP ERP - ARCHITECTUREInfrastructure2 data centers ~160 miles apartHardwareDatabase Servers: 4 X 8 cores, 256 GB of memoryEMC SANAlwaysOn HA / DR Solution
  97. 97. MSIT SAP ERP – HA/DR ARCHITECTURE
  98. 98. SUMMARY: MSIT SAP ERPGains and LearningsReplaced “combination of multiple technologies for HA + DR”, with a single technologyRequired a lot of work across different IT teamsDO NOT MISS !DB316 How does Microsoft run their SAP landscape on windows and SQL Server Elke BreglerFriday July 27, 13:00 - 14:15
  99. 99. MSIT – SERVICES INTEGRATION LAYER (SIL)BusinessThe Service Integration Layer (SIL) is a SOA based middleware messaging platform thatwill expose common capabilities offered by the Microsoft Service & Support business.This abstraction and services layer facilitates seamless integration of multiple serviceconsumers and service providers.ApplicationMSSOLVE, Metropolis, and other Incident Management systems would be leveraging SILfor data access to downstream customers (EDW, DETEGO).
  100. 100. MSIT SIL - ARCHITECTUREInfrastructure2 data centers ~160 miles apartHardware16 processors, 72 GB RAMDirect attached storageAlwaysOn HA / DR Solution
  101. 101. MSIT SIL – HA/DR ARCHITECTURE
  102. 102. SUMMARY: MSIT SIL
  103. 103. BWIN.PARTYhttp://www.microsoft.com/casestudies/Microsoft-SQL-Server-2012/bwin.party/Company-Cuts-Reporting-Time-by-up-to-99-Percent-to-3-Seconds-and-Boosts-Scalability/710000000087
  104. 104. BWIN.PARTY - ARCHITECTUREInfrastructureMultiple data centers (3 used for the scenario we are focusing on here).Data centers hosting primary and secondary replicas are 11 km apartHardwareFujitsu RX-600 S5 - 4 socket Nehalem-EX CPUs, 32 processor cores, 256 GB RAMExternal SAS Disks for Data Files - Attached with 4x 8GB Fiber Channel, FusionIO SSD’s for TlogAlwaysOn HA / DR Solution
  105. 105. BWIN.PARTY – HA/DR ARCHITECTURE
  106. 106. SUMMARY: BWIN.PARTYEliminated need to maintain mirroring connection strings (failover_partner) in allapplications. This was painful, and in some cases (some 3rd party applications) not even supported.Eliminated Data Warehouse load restrictions due to Log Shipping (restore and queryconflict).Migration involves other teams, not just the DBA team.Troubleshooting now either needs way smarter DBAs or other teams involved.Migration: Need to change connection string if more than 1 secondary replica.Different machines used different OS versions before. This is no longer possible.
  107. 107. CAREGROUP HEALTHCAREFour Hospitals located in Boston, MAAmong Top 5 Largest Healthcare Systems in the USA146 Mission Critical Clinical Applications2+ Million Patient Medical RecordsRanked #1: Most Innovative Healthcare IT nationwide (InformationWeek)80+ databases rated “AAA”: RPO 0 & RTO 0300+ databases rated “AA”: RPO =<1 hour & RTO 1 hourRest of the databases rated “A”: RPO & RTO 1 dayA “AA” class application is discussed in the next few slides.http://www.microsoft.com/casestudies/Microsoft-SQL-Server-2012/Beth-Israel-Deaconess-Medical-Center/Hospital-Improves-Availability-and-Speeds-Performance-to-Deliver-High-Quality-Care/5000000011
  108. 108. CAREGROUP – “AA” ARCHITECTUREInfrastructure2 data centers a few miles apart, collocated with the hospitalsHardwareHardware for “AA” class of applications are virtualized using Hyper-VEMC Clarion SAN used for storageAlwaysOn HA / DR Solution
  109. 109. CAREGROUP “AA” – HA/DR ARCHITECTURE
  110. 110. SUMMARY: CAREGROUP HEALTHCARETest and validation Performed parallel run for a week in pre-production Performance of AG at par with DBM performanceNeeded close working between Windows admin team and DBA teamNeeded to rethink quorum model
  111. 111. MS DYNAMICS CRM ONLINEOnline with SQL Server 2012 ~4,000 customers. 2000+ Servers , 30000 + customers provisioned online, 200+ trials added each dayMultiple versions of CRM online supported at any given timeDB size ranging from 100 MB to 60+ GB, each organization gets their own databaseEnable failovers at multiple levels and keep it simpleRedundancy and fault recovery at all levelsBusiness continuity and SLA commitments 99.99% (4 9’s availability)
  112. 112. MS DYNAMICS CRM ONLINE- ARCHITECTUREInfrastructure6 data centers worldwide, 2 in each geography (primary and DR)Multiple subnetsHardwareHP ProLiant, 2 socket Intel Quad-Core Xeon L5630 2.26GHz, 48GB Memory, 1TB DASDTwo embedded Gb Ethernet NICs, P410 Smart Array controller with 256MB BBWC; N+1 RedundantAlwaysOn HA / DR Solution
  113. 113. DYNAMICS CRM ONLINE – HA/DR ARCHITECTURE
  114. 114. SUMMARY: MS DYNAMICS CRM ONLINEPerformed successful Availability Group failovers in production.Performed successful data center failover in production (Hongkong failing over to Singapore).Simple and Unified: New architecture that natively supports multiple data centers and meets all HA/DR needs with SQL Server 2012Important Gaps Filled: Parallel performance testing with Database Mirroring technologies Successfully pushed AlwaysOn technologies past recommended boundaries Identified multiple product development bugs and issues Compliance with Microsoft Business Continuity Standards Ready for ISO Certifications Removed a big road block for enterprise adoptionConsistent, Solid Foundation: Big step towards a future of further scale and robustness for CRM Online service
  115. 115. SUNDIO GROUPLeading tour operator in Europe.Sells trips to holiday destinations throughout Europe.10 brands, 9 countriesOver 700K passengers in 2011Online travel reservation systemhttp://www.microsoft.com/casestudies/Microsoft-SQL-Server-2012/Sundio-Group/Online-Travel-Firm-Boosts-Availability-and-Performance-to-Handle-Peak-Sales-Volumes-with-Microsoft-Solution/4000011822
  116. 116. SUNDIO GROUP - ARCHITECTUREInfrastructure1 data centerHardwareHP BL680 G7 blade serversEMC VMAX SANAlwaysOn HA / DR Solution
  117. 117. SUNDIO GROUP – HA/DR ARCHITECTUREFCI for local HA, AG for Offloading read workload
  118. 118. SUMMARY: SUNDIO GROUPFailover Cluster Instance and Availability Groups (FCI + AG) solution FCI used for HA AG used for offloading read workloadUse possible owners appropriately to ensure they don’t overlap for AG primary and AGsecondaryTransactional replication for read scale out For the workload that requires only a subset of the data and different indexes
  119. 119. BALTIKA BREWERIESThe leader of the Russian beer market with 39.7% market share.The largest FMCG (Fast Moving Consumer Goods) company in Russia.The Baltika brand is the №1 beer in Europe in terms of sales (Euromonitor) Annual Revenue: 79,307 million RBL or 2,643 million USD - (2010 year)CRM Application through ISV-Monolit.The core of Baltika’s selling process.Provides a workspace for sales reps, merchandisers, service agents: Taking and processing orders from distributors Data for sales and marketing analytics1500 back office/5000 mobile employees, 350 distributors.100 000 orders a day, 50 000 distributor’s data exchange sessions a day.Tier 1 application: 24x7 nonstop services as accessed from different regions and time-zoneshttp://www.microsoft.com/casestudies/Microsoft-SQL-Server-2012-Enterprise/Baltika-Brewing-Company/Brewer-Increases-Efficiency-and-Insight-Boosts-Performance-by-up-to-78-Percent/710000000527
  120. 120. BALTIKA - ARCHITECTUREInfrastructure1 data centerHardwareHP ProLiant BL685c G7 Four 12-Core AMD Opteron, 2200 MHz proc., 256 Gb RAMHP EVAAlwaysOn HA / DR Solution
  121. 121. BALTIKA – HA/DR ARCHITECTURE
  122. 122. SUMMARY: BALTIKANo impact/overhead on Primary replica when using Asynchronous secondaryRead-scale was able to be implemented via Availability Groups readable secondary replicasGraphic below shows # of commands (1k-9M) and time (seconds) to deliver transactions 100k commands the time for replication goes up significantly
  123. 123. SERVICEU CORP (PART OF ACTIVE NETWORK)ServiceU provides web-based online scheduling, event management, payment processing, and other services tocustomers in 15 countries and 50 united states.99.99% uptime (maximum allowable downtime of 52 minutes per year including scheduled maintenance).Security – Level 1 PCI Service Provider.Performance.Try to eliminate manual intervention as much as possible.Eliminate single points of failure as much as possible.Highly trained personnel, extensive monitoring, good documentation, standardization across the enterprise.http://www.microsoft.com/casestudies/Microsoft-SQL-Server-2012/ServiceU/Online-Company-Reduces-Downtime-and-Helps-Its-Customers-to-Improve-Service/4000011506
  124. 124. SERVICEU - ARCHITECTUREInfrastructure2 data centers ~500 miles apart. Each data center is identical in capacity and infrastructureHardwareDell R815 (2 X 12 cores), 128 GB RAMEMC SANAlwaysOn HA / DR Solution
  125. 125. SERVICEU - HA / DR ARCHITECTURE
  126. 126. SUMMARY: SERVICEUUpgraded to SQL Server 2012 with 90 seconds of downtime (half of which was actual upgradedowntime, and the other half for final testing during the upgrade).Asymmetric disk as quorum resource.DO NOT use Windows Failover Cluster Manager to perform Availability Group Failover.Concurrent ADD NODE operation yields unexpected results in a SQL Server Failover Cluster Instance.Ensure correct possible owners have been set for each resource within each FCI.In a FCI + AG setup, the SQL Instance names must be unique within the Windows Cluster.
  127. 127. EDGENETLeader in Data Services, Guided Selling and Marketing SolutionsOrganize product information to increase sales: Help retailers sell configurable products Help consumers compare and purchase the right product for them.Collect, certify and distribute product data Google/Bing Search & Shopping, and for Retailers One of Four Active US GDSN-certified poolsProvides high availability and disaster recovery for data within the pool of applications above Near real-time data replication with MSDTC support Need for data to be available for Exports & BI WorkloadsSLA: 99.99% Annual uptimehttp://www.microsoft.com/casestudies/Microsoft-SQL-Server-2012/Edgenet/Data-Provider-Supports-Growth-and-Gains-Competitive-Advantage-with-Microsoft/4000011528
  128. 128. EDGENET - ARCHITECTUREInfrastructure2 data centers: Milwaukee, WI and Atlanta, GA (850 miles in between)Multiple subnetsHardwareNEC Express 5800/A1080a-D GXEMC Clariion CX4-80, EMC Recover Point CE – Disk Based ReplicationAlwaysOn HA / DR Solution
  129. 129. EDGENET - HA / DR ARCHITECTURE
  130. 130. SUMMARY: EDGENETAllowed for true multi-site failover without having to create a VLAN.Advantage of TempDB on Local Disk Saves money on storage replication cost. Reduces cross-data center storage replication traffic. Enables use of local solid state storage to improve performance.Provided ability to offload BI workload and exports to read-only secondary database.
  131. 131. RESOURCES
  132. 132. © 2011 Microsoft Corporation. All rights reserved. Microsoft, Windows, Windows Vista and other product names are or may be registered trademarks and/or trademarks in the U.S. and/or other countries.The information herein is for informational purposes only and represents the current view of Microsoft Corporation as of the date of this presentation. Because Microsoft must respond to changing market conditions, it should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of any information provided after the date of this presentation. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.

×