Hyper-V is Microsoft's hypervisor-based virtualization system for x86-64 systems that supports isolation in terms of partitions. These partitions can each run copies of operating systems and allow applications to run in isolation. As companies move to a more virtualized environment, many will choose to use different virtualization technologies. Each of these technologies, Hyper-V included, give companies opportunities to better manage capacity in data centers. Some people have argued that virtualization and the management tools that are built in have eliminated the need for a Capacity Management process and/or a Capacity Management staff. This couldn't be farther from the truth, as a poorly managed virtualized environment can cause performance problems for all the services that run in the environment. Therefore, proper Capacity Management processes are even more important in a virtualized environment. Once a company recognizes that Capacity Management is vital, the next step is to put a process and a set of tools in place that will help the Capacity Manager understand and make appropriate recommendations to management. This webinar will look at the following: •A brief overview of Hyper-V •A look at the data and information that's available to the Capacity Manager •Some unique challenges that Hyper-V brings to the Capacity Manager •How a properly managed Hyper-V environment can help maximize the use of the deployed hardware

1. Capacity Management
For Hyper-V

2. Agenda
• A brief overview of Hyper-V
• A look at the data and information that's available to the
Capacity Manager
• Some unique challenges that Hyper-V brings to the Capacity
Manager

3. Overview of Hyper-V

4. What is Hyper-V?
• A software virtual machine monitor for x64 systems that shares the
same design as Xen
• Type 1 Hypervisor
• First production release was on 26 June 2008
• Key elements are:
• The hypervisor (around 100k in size)
• Parent or root partition (the first and controlling guest)
• Child partitions
• Two versions

5. What is Hyper-V?
• Windows 2008 R2
• Hyper-V role
• Windows + virtualization
• Live Migration
• Clustering capability
• Hyper-V Server 2008 R2
• Light weight version
• Purely virtualization

6. Architecture Diagram from MSDN

7. Hyper-V Guest Versions

8. Hyper-V Core – Dynamic Memory
• Available with SP1
• Adjust memory based on workload
• Memory management
• Startup RAM
• Max RAM
• Memory buffer & pressure
• Memory priority

9. Hyper-V Core – Dynamic Memory
• Dynamic memory buffer and pressure
• Pressure = ratio of memory needs/has
• Buffer = Percentage of committed memory
• Dynamic Memory Priority
• Set at the VM level

10. Hyper-V Core - Live Migration
• Source and destination host must be part of same failover cluster
• VM must be on shared storage
• Host processors must be the same
• Manufacturer and processor
• You need SCVMM R2
• Underlying OS must be Windows Server 2008 R2

11. Hyper-V vs vmware
• Cost savings
• Licenses very cheap
• New vmware cost memory
• Potentially better performance with other MS applications
• Access to internal MS teams
• Less functionality (although starting to catch-up)

12. Monitoring Hyper-V

13. Performance Monitoring
• Capturing the data
• SCOM/SCVMM
• Raw performance counters
• Interpreting the data

14. System Center Operations Manager
• Provides central source of monitoring for Hyper-V
• Management packs
• Minimal metrics
• No focus on Capacity Management
• Inbuilt aggregation
• Provides multiple monitoring levels
• Host
• Guest
• Application

15. System Center Virtual Machine Manager
• Multiple host management
• Multiple hypervisor management
• Template and library management
• Integrated P2V
• VM performance monitoring
• Live Migration
• Manage vmware estate as well (via vCenter)

16. Capturing Performance Data
• Main sources of information are the Hyper-V performance counters as
seen from the root partition
• 21 functioning counters that provide around 600 metrics in total
• Vendor products should interrogate these remotely via WMI
• Perfmon metrics within each guest partition may not be reliable
• For CPU etc.
• However certain other metrics can be used
• Monitoring via SCVMM

17. Performance counters
• CPU
• Hyper-V Hypervisor Logical Processor
• Hyper-V Hypervisor Root Virtual Processor
• Hyper-V Hypervisor Virtual Processor
• Processor
• Memory
• Hyper-V Hypervisor Partition
• Hyper-V Hypervisor Root Partition
• Hyper-V VM Vid Partition
• Hyper-V Dynamic Memory Balancer
• Hyper-V Dynamic Memory VM
• Memory

18. Performance counters
• Network
• Network Interface
• Hyper-V Virtual Switch
• Hyper-V Legacy Network Adapter
• Hyper-V Virtual Network Adapter
• Storage
• Physical Disk
• Hyper-V Virtual IDE Controller
• Hyper-V Virtual Storage Device

19. Hyper-V RVP CPU Utilization
WS2008ENTHyper-V Hypervisor Root Virtual Processor(_Total)% Total Run Time
40
35
30
25
WS2008ENTHyper-V Hypervisor
20 Root Virtual Processor(_Total)%
Total Run Time
15
10
5
0
12:30
12:38
12:46
12:54
13:02
13:10
13:18
13:26
13:34
13:42
13:50
13:58
14:06
14:14
14:22
14:30
14:38

20. CPU viewed within Root Partition
RVP Internal CPU Total Util. Reported (%)
45
40
35
30
25
CPU Total Util. Reported (%)
20
15
10
5
0
0
0
0
0
0
0
0
0
0
0
0
0
0
:3
:4
:5
:0
:1
:2
:3
:4
:5
:0
:1
:2
:3
12
12
12
13
13
13
13
13
13
14
14
14
14

21. Hyper-V VP CPU % for WS2003STD
WS2003STD Average CPU% VP Time
70.00
60.00
50.00
40.00
WS2003STD Average CPU%
VP Time
30.00
20.00
10.00
0.00
12 1
9
12 7
13 5
13 3
1
13 9
13 7
13 5
3
13 1
14 9
14 7
5
14 3
1
:3
:3
:4
:5
:0
:1
:1
:2
:3
:4
:5
:5
:0
:1
:2
:3
12
12
13
13
14

22. CPU viewed from within WS2003STD guest
CPU Total Util. Reported (%)
70
60
50
40
CPU Total Util. Reported (%)
30
20
10
0
0
8
6
4
2
0
8
6
4
2
0
8
6
4
2
0
8
:3
:3
:4
:5
:0
:1
:1
:2
:3
:4
:5
:5
:0
:1
:2
:3
:3
12
12
12
12
13
13
13
13
13
13
13
13
14
14
14
14
14

23. Hyper-V VP CPU % for Fedora9
WS2008ENTHyper-V Hypervisor Virtual Processor(Fedora9x86_64:Hv VP 0)% Total Run
Time
60
50
40
WS2008ENTHyper-V
Hypervisor Virtual
30
Processor(Fedora9x86_64:Hv
VP 0)% Total Run Time
20
10
0
12:31
12:39
12:47
12:55
13:03
13:11
13:19
13:27
13:35
13:43
13:51
13:59
14:07
14:15
14:23
14:31

24. CPU viewed from within Fedora9 guest
CPU Utilization Total Reported (%)
60
50
40
CPU Utilization Total Reported
30
(%)
20
10
0
12:30
12:38
12:46
12:54
13:02
13:10
13:18
13:26
13:34
13:42
13:50
13:58
14:06
14:14
14:22
14:30
14:38

25. Capacity Challenges

26. Challenges – Getting the data
• WMI access directly to the host
• Provides a view on physical and partition usage
• Misses the wider cluster view
• Lack application/process information
• Via SCOM/SCVMM
• Provides wider view of performance
• Default metrics light on performance/capacity
• Multiple platforms
• Windows and Linux information

27. Challenges – The levels
• Cluster
• Individual application clusters
• The wider Hyper-V estate
• Host
• How is the host performing
• How much capacity is available
• Guest
• Check dynamic memory settings
• Application performance

28. Simple performance guidelines
• CPU performance
• Logical processors
• Virtual processors
• MSDN troubleshooting guide
• Memory performance
• Memory available and paging
• Disk I/O performance
• Logical disk latency metrics
• .VHD usage, care with static/dynamic
• Network performance
• Bytes/sec and output queue length

29. Any Questions?

30. Capacity Management
For Hyper-V