vmworld 2015: advanced sql server on vsphere
TRANSCRIPT
Advanced SQL Server on vSphereScott Salyer, VMware, Inc
Wanda He, EMC
VAPP5598
#VAPP5598
CONFIDENTIAL 2
• This presentation may contain product features that are currently under development.
• This overview of new technology represents no commitment from VMware to deliver these features in any generally available product.
• Features are subject to change, and must not be included in contracts, purchase orders, or sales agreements of any kind.
• Technical feasibility and market demand will affect final delivery.
• Pricing and packaging for any new technologies or features discussed or presented have not been determined.
Disclaimer
The Percentage of Applications in Virtualized InfrastructureHas Increased Dramatically over the Last Few Years (VMware Core Metrics Survey July 2015)
CONFIDENTIAL 3
Microsoft SQL is the most common application running in on-premise virtual infrastructure
NA EU dAP BRIC SMB COMM ENT
57% 73% 70% 74% 68% 71% 64%
47% 51% 39% 56% 43% 51% 54%
41% 43% 46% 61% 36% 46% 57%
45% 54% 37% 41% 43% 49% 46%
34% 38% 59% 51% 37% 39% 48%
26% 27% 32% 37% 24% 34% 33%
25% 30% 23% 35% 16% 30% 39%
29% 16% 31% 27% 22% 22% 30%
15% 23% 30% 28% 19% 24% 25%
15% 22% 22% 30% 17% 21% 25%
71% 62% 62% 64% 65% 64% 68%
48% 54% 49% 55% 50% 51% 53%
51% 45% 49% 49% 44% 49% 53%
36% 35% 39% 46% 37% 40% 37%
20% 15% 20% 26% 15% 17% 25%600 450 230 323 653 346 604
Region Company Size
Microsoft SQL
Microsoft SharePoint
SAP
Microsoft Exchange
Oracle Databases
Oracle Applications
High Performance Computing
Custom BCA/ industry-specific
Oracle Middleware
IBM Middleware
Business critical
Important
Development
Test
Staging
67%
49%
46%
45%
42%
29%
28%
25%
22%
21%
66%
51%
49%
38%
19%
Applications in Virtualized Infrastructure
> Total< Total
N = 1603
Level of Criticality of Applications in Virtualized Infrastructure
(Select all that apply)
(Select all that apply)
Virtualizing Applications Sessions and Offerings
• 30 Breakout Sessions with 5 Panels & 4 Quick Talks• 10 Group Discussions• One-on-One Meet the Experts Sessions• Checkout the Hands on Labs
Sign up for the Independent Oracle User Group (IOUG) VMware Special Interest Group (SIG)www.ioug.org/vmware
CONFIDENTIAL 5
RDBMS Books from VMware Press
Book signing @ 1PM Tuesday Sept 1
vmwarepress.com
http://www.pearsonitcertification.com/store/virtualizing-oracle-databases-on-vsphere-9780133570182http://www.pearsonitcertification.com/store/virtualizing-sql-server-with-vmware-doing-it-right-9780321927750
CONFIDENTIAL 6
Agenda
1 Introduction – Why Virtualize SQL Server?
2 Designing for Performance• Storage Design• Memory Optimization• NUMA Considerations• ExtremeIO Enhancements• Monitoring and Troubleshooting
3 Consolidating Multiple SQL Server Workloads
4 SQL Server Availability
IntroductionWhy virtualize SQL Server?
CONFIDENTIAL 8
Quick Facts• SQL Server database servers account for ~10% of all x86 workloads and are typically
underutilized (6-20% CPU utilization)
• Many Database Administrators (DBAs) are hesitant to virtualize database servers due to perceived issues with performance, availability, and licensing
• Running SQL Server workloads on vSphere can help to reduce physical server and licensing costs while increasing availability without sacrificing performance
• The VMware SDDC platform offers management benefits that extend to both the infrastructure administrator and the DBA– In-depth application monitoring and trend analysis– Automation and provisioning of database components for developers (self-service)– Application and site-resiliency
CONFIDENTIAL 9
Reduce hardware costs by > 50%• Consolidate servers by 4X – 20X
Provision databases on demand• Minutes to provision in production and in the lab
Reduce licensing costs• Potentially increase utilization of SQL Server licenses
(depending on degree of consolidation)
Increase application Quality of Service• Scale dynamically• Built-in high availability and simple disaster recovery
DB On Demand
Quality of Service
DB Consolidation
Why Deploy SQL Server on VMware SDDC?
Licensing
Complete isolation between systems on the same host• Protects databases and applications against network-based threatsSecurity
Designing for PerformanceTier-1 Production Workloads
Storage Design
CONFIDENTIAL 12
Factors Affecting Storage Performance
vSCSI adapterApplication
VMKernel
FC/iSCSI/NAS
VMKernel admittance ( Disk.SchedNumReqOutstanding)
Per path queue depthAdapter queue depth
Storage network (link speed,zoning, subnetting)
HBA target queuesArray SPs# of disks
Disk queue depthAdapter queue depth
Adapter type
CONFIDENTIAL 13
Optimize for Performance – Queue Depth• vSCSI Adapter
– Be aware of per device/adapter queue depth maximums (KB 1267)– Use multiple PVSCSI adapters, follow KB 2053145 for large-scale IO intensive SQL deployment (not applicable for
typical deployment)
• VMKernel Admittance– VMKernel admittance policy affecting shared datastore (KB 1268), use dedicated data store for mission critical SQL
Server– VMKernel admittance changes dynamically when SIOC is enabled (recommended for lower-tier SQL Server
deployments)
• Physical HBAs– Follow vendor recommendation on max queue depth per LUN (http://kb.vmware.com/kb/1267)– Follow vendor recommendation on HBA execution throttle– Be aware settings are global if host is connected to multiple storage arrays– Pick the right multipathing policy based on vendor storage array design
CONFIDENTIAL 14
Optimize for Performance – Storage Network• Link Type/Speed
– FC vs. iSCSI vs. NAS– Latency suffers when bandwidth is saturated
• Zoning and Subnetting– Place hosts and storage on the same switch, minimize ISL– Use 1:1 initiator to target zoning or follow vendor recommendation– Enable jumbo frame for IP based storage (MTU needs to be set on all connected physical and
virtual devices)– Make sure different iSCSI IP subnets cannot transmit traffics between them
CONFIDENTIAL 15
VMDK Lazy Zeroing*• Default VMDK allocation policy lazy zeroes 1M VMFS
blocks on first write
• Write penalty on an untouched VMDK
• SQL Server operations could be affected by lazy zeroing– Write operations– Read operations that use tempdb extensively– Bulk load/index maintenance
• For best performance, format VMDK as eagerzeroedthick *
* Zero offload capability in VAAI improves zeroing in supported arrays
1 host 2 hosts
4 hosts
8 hosts
16 hosts
0
20
40
60
80
100
120
140
160
180
200
Effect of Zeroing on Storage Performance
"Post-zeroing" "Zeroing"
Thro
ughp
ut (M
Bps
)
Choose Storage which supports VMware vStorage APIs for Array Integration (VAAI)
CONFIDENTIAL 16
SQL Server Data Files
• For optimal performance, dedicate data stores / LUNs to Data, TempDB, and Log files– Defaults place data and logs on the same volume…change with ALTER DATABASE -> MODIFY FILE statement– Data and TempDB files can share the same data store / LUN but dedicate a LUN to Log files
– Stripe data across as many physical spindles as possible
• Use multiple TempDB files – start with 1 TempDB file per CPU core (up to 8)– Files should be of equal size – SQL Server favors allocations in files with more free space (hot spots)
• Number of Data files will depend on the size of the database and backup considerations– Or alternately .25 Data files per CPU core
• If using storage tiering use the following file placement priority – fastest to slowest drive:– TempDB Data Files > Transaction Log Files > Data Files
• Use RAID 10 for log, use RAID 10 or RAID 5 for data, tempdb (not applicable to XtremIO)
Follow Microsoft and Storage Vendor Recommendations!
CONFIDENTIAL 17
Strict Best Practices SQL Server VM Disk Layout ExampleCharacteristics:• OS on shared DataStore/LUN• 1 database; 4 equally-sized data
files across 4 LUNs• 1 TempDB; 4 (1/vCPU) equally-sized
tempdb files across 4 LUNs• Data, TempDB, and Log files spread
across 3 PVSCSI adapters– Data and TempDB files share PVSCSI
adapters
• VMDK/DataStores could be RDMsAdvantages:• Optimal performance; each Data,
TempDB, and Log file has a dedicated VMDK/Data Store/LUN
• I/O spread evenly across PVSCSI adapters
• Log traffic does not contend with random Data/TempDB traffic
NTFS Partition: 64K cluster size
C:\ D:\ H:\ E:\ I:\ L:\ T:\
DataFile1.mdf
DataFile5.ndf
LogFile1.ldf
TmpLog1.ldf
OS
ESX Host
LUN1
Data Store 1
VMDK1
LUN2
VMDK2
LUN3
VMDK3
LUN4
VMDK4
SQL ServerOS
Can be placed on a DataStore/LUN
with other OS VMDKs
Can be Mount Points under a drive as well.
OS VMDK
Can also be a shared LUN since TempDB is usually in Simple
Recovery Mode
PVSCSI1LSI1
F:\ J:\ G:\ K:\
TmpFile1.mdf
TmpFile2.ndf
TmpFile3.ndf
TmpFile4.ndf
Data Store 2 Data Store 3 Data Store 4
LUN5
VMDK5
LUN6
VMDK6
Data Store 5 Data Store 6
LUN5
VMDK5
LUN6
VMDK6
PVSCSI2
Data Store 5 Data Store 6
LUN5
VMDK5
LUN6
VMDK6
PVSCSI3
Data Store 5 Data Store 6
DataFile3.ndf
DataFile7.ndf
Disadvantages:• You can quickly hit the ESX LUN Limit!• More complicated storage management
CONFIDENTIAL 18
Realistic SQL Server VM Disk Layout ExampleCharacteristics:• OS on shared DataStore/LUN• 1 database; 8 Equally-sized data files
across 4 LUNs• 1 TempDB; 4 files (1/vCPU) evenly
distributed and mixed with data files to avoid “hot spots”
• Data, TempDB, and Log files spread across 3 PVSCSI adapters
• VMDK/DataStores could be RDMs
Advantages:• Fewer LUNs used• I/O spread evenly/TempDB
hot spots avoided• Log traffic does not contend with
random Data/TempDB traffic
NTFS Partition: 64K cluster size
C:\ D:\ E:\ F:\ G:\ L:\ T:\
DataFile1.mdf
DataFile2.ndf
TmpFile1.mdf
DataFile4.ndf
DataFile3.ndf
TmpFile2.ndf
DataFile5.ndf
DataFile6.ndf
TmpFile3.ndf
DataFile7.ndf
DataFile8.ndf
TmpFile4.ndf
LogFile.ldf TmpLog.ldfOS
ESX Host
LUN1
Data Store 1
VMDK1
LUN2
Data Store 2
VMDK2
LUN3
Data Store 3
VMDK3
LUN4
Data Store 4
VMDK4
LUN5
Data Store 5
VMDK5
LUN6
Data Store 6
VMDK6
SQL ServerOS
Can be placed on a DataStore/LUN with other
OS VMDKs
Can be Mount Points under a drive as well.
OS VMDK
Can also be a shared LUN since TempDB is
usually in Simple Recovery Mode
PVSCSI1LSI1 PVSCSI2 PVSCSI3
CONFIDENTIAL 19
Block Alignment• Configure storage presented to vSphere hosts using vCenter to
ensure VMFS block alignment
• Misalignment can result in up to 50% performance hit
• Windows 2008+ should align sectors by default…double-check!– http://msdn.microsoft.com/en-us/library/dd758814.aspx – http://blogs.msdn.com/b/jimmymay/archive/2014/03/14/disk-partition-alignme
nt-for-windows-server-2012-sql-server-2012-and-sql-server-2014.aspx (Jimmy May - MSDN Blogs)
• For example, OEM setups with recovery partitions could result undesirable starting offsets for the Windows partition
Unaligned partitions result in additional I/O
Aligned partitions reduce I/O
stripe unit size value should be an integer
Memory Optimization
CONFIDENTIAL 21
Large Pages in SQL Server Configuration Manager (Guest)• Use Large Pages in the guest
– ON by default in 2012/2014 if Lock Pages in Memory User Right is granted to SQL Server Service Account (sqlservr.exe) and if the VM has 8+ GB of memory - http://msdn.microsoft.com/en-us/library/ms178067.aspx
– For older versions, start SQL Server with trace flag -T834
NUMA Considerations
CONFIDENTIAL 23
Non-Uniform Memory Access (NUMA)• Designed to avoid the performance hit when several
processors attempt to address the same memory by providing separate memory for each NUMA Node.
• Speeds up Processing
• NUMA Nodes Specific to Each Processor Model
• Enable Non-Uniform Memory Access (NUMA) in BIOS– ESXi is NUMA-aware and will schedule vCPUs onto a
single NUMA node (if it fits)– SQL Server is NUMA-aware
CONFIDENTIAL 24
NUMA Best Practices• http://www.vmware.com/files/pdf/techpaper/VMware-vSphere-CPU-Sched-Perf.pdf
• Avoid Remote NUMA access– Size # of vCPUs to be <= the # of cores on a NUMA node (processor socket)
• Where possible, align VMs with physical NUMA boundaries
– For wide VMs use a multiple along with virtual NUMA
• Hyperthreading– Initial conservative sizing: set vCPUs equal to # of cores– HT benefit around 20-25%, < for CPU intensive batch jobs (based on OLTP workload tests )
• # of virtual sockets and # of cores / virtual socket: where possible keep default 1 core / socket
• ESXTOP to monitor NUMA performance in vSphere– Coreinfo.exe to see NUMA topology in Windows Guest
• If vMotioning, move between hosts with the same NUMA architecture to avoid performance hit (until reboot)
CONFIDENTIAL 25
Non-Wide VM Sizing Example (VM fits within NUMA Node) • 1 vCPU per core with hyperthreading OFF
– Must license each core for SQL Server
• 1 vCPU per thread with hyperthreading ON– 10%-25% gain in processing power– Must license each thread for SQL Server “numa.vcpu.preferHT” to true to force 24-way VM
to be scheduled within NUMA node
SQL Server VM: 24 vCPUs
NUMA Node 0: 128 GB Memory
0 1 2 3 4 5 6 7 8 9 10 11
SQL Server VM: 12 vCPUs
NUMA Node 0: 128 GB Memory
0 1 2 3 4 5 6 7 8 9 10 11
Hyperthreading OFF Hyperthreading ON
CONFIDENTIAL 26
SQL Server VM: 24 vCPUs
NUMA Node 0: 128 GB Memory
0 1 2 3 4 5 6 7 8 9 10 11
NUMA Node 1: 128 GB Memory
0 1 2 3 4 5 6 7 8 9 10 11
Virtual NUMA Node 1Virtual NUMA Node 0
Hyperthreading OFF
Wide VM Sizing Example (VM crosses NUMA Node) • Extends NUMA awareness to the guest OS
• Enabled through multicore UI– On by default for 8+ vCPU multicore VM– Existing VMs are not affected through upgrade– For smaller VMs, enable by setting numa.vcpu.min=4
• Do NOT turn on CPU Hot-Add
• For wide virtual machines, confirm feature is on for best performance
XtremIO Enhancements
CONFIDENTIAL 28
Common Challenges• Consolidation creates random performance issues
• Copying data is common – VMs, databases, analytics
• Workflow complexity – downstream data services, lifecycle management
MASSIVE OPPORTUNITYFOR IMPROVEMENT
CONFIDENTIAL 29
XtremIO All-Flash Storage Array with Rich Data Services
Advantages for SQL Server: Scaling: Linear, Dynamic for Capacity & IOPs Performance: Consistent & Predictable at Scale Workflow: Zero Impact Copy Services Data-Reduction: In-line and Unstoppable Enterprise-Ready: Data Protection & Availability Simplicity: Zero Capacity Planning & Tuning
SCALE-OUTCONSISTENT & PREDICTABLE
SUB-MS LATENCIESDATA SERVICES
INLINE ALL THE TIMEMANAGEMENT SIMPLICITY
& AUTOMATION
CONFIDENTIAL 30
Performance – Consistent Low Latency• Consolidate 1, 2, 4, and 8 SQL
Server VMs• DB size: ~1TB each• XtremIO dual X-Brick• OLTP like workload, 90% read,
10% write, majority 8K random
• Near linear scalability of IOPS• Consistent sub-millisecond avg.
disk latency
CONFIDENTIAL 31
Performance – Increase Consolidation Efficiency
• Zero tuning performance enhancement
• Reduce of batch run time from 7hrs to 1hr
• Opportunity to increase consolidation density with more efficient resource utilization
• Maximize return of investment on SQL Server licenses
Sales D
ata 1
Sales D
ata 2
Sales D
ata 3
Inven
tory D
ata 1
Inven
tory D
ata 2
Inven
tory D
ata 3
Inven
tory D
ata 4
Inven
tory D
ata 5
Receiv
ables
Data
1
Receiv
ables
Data
1
Receiv
ables
Data
10
1
2
3
4
5
6
7
8
Traditional Storage XtremIO
Bat
ch J
ob R
un T
ime
(hrs
)
Before and After Customer Study
CONFIDENTIAL 32
Storage Efficiency• “Zero” optimized
– Instant eagerzeroedthick – Zero space for pre-allocated space
• Space efficient VM clones, database copies
• Automatic compression, full compatibility w/ SQLnative compression
PHYSICAL
Database Size on VM = 1TB
Logical Volume = 610GB
Provisioned VMDK = 2TB
PHYSICALPhysical = 383GB
Thin Provision
Zero Optimized
Dedupe, Compressed
5.3:1
Overall Efficiency
CONFIDENTIAL 33
Simplicity – Capacity Planning and Operational Management
DONE!
XTREMIO
Volume Size
RAID group/Number of spindles
Storage pool/tiering
Number of LUNs
Separate data vs. log
Tempdb placement
Separate OLTP vs. analytics workloads
HISTORIC COMPLEXITIES
Right sizing: hot vs. active vs. cold
DESIGN / DEVELOPUPGRADE
TUNE
MAINTAIN
REPORTING
Backup
Test/Dev Copies
Offload Query Tuning
dbcc checkdb
App-Consistent Snapshot Backup
Reporting Copies
Copy for Failback
Agility – Enhanced Lifecycle Management w/ Copy Services
Demo: Rapid Deployment of AlwaysOn Secondary Replica
Monitoring and Troubleshooting
CONFIDENTIAL 37
Performance Needs Monitoring at Every Level
ApplicationGuest OS
ESXi Stack
Physical Server
Connectivity
Peripherals
Application Level App Specific Perf tools/statsGuest OS CPU Utilization, Memory Utilization, I/O Latency
Virtualization Level vCenter Performance Metrics /Charts
Limits, Shares, Virtualization Contention
Physical Server Level CPU and Memory Saturation, Power Saving
Connectivity Level Network/FC Switches and data paths Packet loss, Bandwidth Utilization
Peripherals Level SAN or NAS Devices Utilization, Latency, Throughput
STARTHERE
38
VirtualMachine
Storage LUN
PhysicalDisks
Guest OS disk
VMwareData store(VMFSVolume)
.vmdk file
Storage Array
Logical Storage Layers: from Physical Disks to vmdks
CONFIDENTIAL
KAVG• Tracks latency of I/O passing thru
the Kernel• Investigation Threshold: 1ms
DAVG• Tracks latency at the device driver;
includes round-trip time between HBA and storage
• Investigation Threshold: 15 - 20ms, lower is better, some spikes okay
Aborts (ABRT/s)• # commands aborted / sec • Investigation Threshold: 1
GAVG• Tracks latency of I/O in the guest
VM• Investigation Threshold: 15-20ms
KB Article Link:http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1008205
CONFIDENTIAL 39
In-guest SQL Server Performance Monitoring• Perfmon
– SQL Server specific counters: \SQLServer:*\*– VMware Tools includes a Perfmon DLL that provides visibility into host CPU, memory, and disk usage
• SQL Server Extended Events: https://technet.microsoft.com/en-us/library/bb630282(v=sql.105).aspx
• SQL Server Profiler– Monitor SQL Server performance at T-SQL statement level
• SQL Server Dynamic Management Views (DMVs)– Monitor the internal health of SQL Server– Sys.dm_*
• SQL server Replay https://msdn.microsoft.com/en-us/library/ff878183.aspx
Consolidating Multiple SQL Server Workloads
CONFIDENTIAL 41
Consolidation Options• Scale-up approach
– Multiple instances per VM– Fewer virtual machines– More complex workload management
• Use Min Server and Max Server memory OR Resource Governor
– Potential reduction in SQL Server licensing cost
• Scale-out approach– Single instance per VM– Potential increase in mgmt. overhead– Better isolation/performance – Easier security and change mgmt.– DRS more effective with smaller VMs– Faster migration (vMotion)
41
CONFIDENTIAL 42
Running with Mixed SQL Server Workloads• Consider workload characteristics, and manage pCPU overcommitment as a
function of typical utilization– OLTP workloads can be stacked up to a sustained utilization level– OLTP workloads that are high usage during daytime and batch workloads that run during off-peak hours
mix well together– Batch/ETL workloads with different peak periods are mixed well together
• Consider operational history, such as month-end and quarter-end– CPU and memory Hot Add can be used to scale up VMs– Additional VMs can be added to handle peak periods if scale out is a possibility– Reduce virtual machine density, or add more hosts to the cluster
CONFIDENTIAL 43
OLTP vs. Batch Workloads
What this says:• Average 15% Utilization
• Moderate sustained activity (around 28% during working hours 8am-6pm)
• Minimum activities during none working hours
• Peak utilization of 58%
What this says:• Average 15% Utilization
• Very quiet during the working day(less than 8% utilization)
• Heavy activity during 1am-4am, withavg. 73%, and peak 95%Batch Workload (avg. 15%)
OLTP Workload (avg. 15%)
SQL Server Availability
CONFIDENTIAL 45
vSphere Availability Features• vSphere vMotion
– Can reduce virtual machine planned downtime– Relocate SQL Server VMs without end-user interruption– Perform host maintenance any time of the day
• vSphere DRS– Monitors state of virtual machine resource usage– Can automatically and intelligently locate virtual machine – Can create a dynamically balanced SQL deployment
• VMware vSphere High Availability (HA)– Does not require Microsoft Cluster Server– Uses VMware host clusters– Automatically restarts failed SQL virtual machine in minutes– Heartbeat detects hung virtual machines– Application HA can provide availability at the SQL Server service level!
CONFIDENTIAL 46
MicrosoftClustering onVMware
vSpheresupport
VMwareHAsupport
vMotionDRSsupport
StoragevMotionsupport
MSCSNodeLimits
Storage Protocols support Shared Disk
FC
In-Guest
OS iSCSI
NativeiSCSI
In-Guest
OS SMB
FCoE RDM VMFS
SharedDisk
MSCS withShared Disk
Yes Yes1 Yes No 5 2 (pre-5.1 only)
Yes Yes No Yes5 Yes4 Yes2 Yes3
Exchange SingleCopy Cluster
Yes Yes1 Yes No25 (5.1 only)
Yes Yes No Yes5 Yes4 Yes2 Yes3
SQL Clustering Yes Yes1 Yes No25 (5.1 only)
Yes Yes No Yes5 Yes4 Yes2 Yes3
SQL AlwaysOnFailover ClusterInstance
Yes Yes1 Yes No25 (5.1 only)
Yes Yes No Yes5 Yes4 Yes2 Yes3
NonsharedDisk
Network LoadBalance
Yes Yes1 Yes YesSame asOS/app
Yes Yes Yes N/A Yes N/A N/A
Exchange CCR Yes Yes1 Yes YesSame asOS/app Yes Yes Yes N/A Yes N/A N/A
Exchange DAG Yes Yes1 Yes Yes Same asOS/app
Yes Yes Yes N/A Yes N/A N/A
SQL AlwaysOnAvailabilityGroup
Yes Yes1 Yes YesSame asOS/app
Yes Yes Yes N/A Yes N/A N/A
vMotion with Shared Disk Configurations:Requires Hardware 11 Compatibility (vSphere 6.0)Physical Mode Required for vSCSI Controller
Non-Shared Disk Configurations:No VMware Restrictions
* Use affinity/anti-affinity rules when using vSphere HA
VMware Knowledge Base Article: http://kb.vmware.com/kb/1037959
Support for Microsoft Clustering on vSphere
CONFIDENTIAL 47
vSphere HA with Shared Disk Clustering• Supports up to five-node cluster in vSphere 5.1 and
above
• Host attach (FC) , FCoE* or in-guest (iSCSI)
• Supports RDM only– RDMs can be vMotioned in vSphere 6.0!!!– Use DRS affinity/anti-affinity rules to avoid running cluster
virtual machines on the same host
• vSphere HA + failover clustering– Seamless integration, virtual machines rejoin clustering
session after vSphere HA recovery– Can shorten time that database is in unprotected state
Failover clustering supported with vSphere HA since vSphere 4.1
http://kb.vmware.com/kb/1037959
vSphere Cluster 1 – HA-enabled
SQL Server Databases
Normal OperationPhysical disks do not move.
vSphere Cluster 1 – HA-enabled
SQL Server Databases
Cluster Failover
DB FailoverPhysical disks do not move.
vSphere Cluster 1 – HA-enabled
SQL Server Databases
Recovery
HA Reboots
Physical disks do not move.
CONFIDENTIAL 48
vSphere HA with AlwaysOn Availability Groups• Seamless integration
• Protect against hardware/software failure
• DRS anti-affinity rule avoids running virtual machines on the same host
• Support multiple secondary and readable secondary
• Provide local and remote availability
• Full feature compatibility with availability group
• vSphere HA + failover clustering– Seamless integration, virtual machines rejoin clustering
session after vSphere HA recovery– Can shorten time that database is in unprotected state Witness
Direction of data flow
vSphere Cluster 1 – HA-enabled
SQL Server Databases
Normal Operation
Full Quorum
vSphere Cluster 1 – HA-enabled
SQL Server Databases
Cluster Failover
DB Failover
Witness-to-PartnerQuorum
vSphere Cluster 1 – HA-enabled
SQL Server Databases
HA Recovery
HA Reboots
Full Quorum
CONFIDENTIAL 49
Resources • Visit us on the web to learn more on specific apps
– http://www.vmware.com/solutions/business-critical-apps/– Specific page for each major app– Includes Best Practices and Design/Sizing information
• Visit our Business Critical Application blog – http://blogs.vmware.com/apps/
CONFIDENTIAL 50
Questions?
Advanced SQL Server on vSphereScott Salyer, VMware, Inc
Wanda He, EMC
VAPP5598
#VAPP5598