Dell | PowerVault MD3000i | User`s guide | Dell PowerVault MD3000i User`s guide

Dell PowerVault MD3000i User`s guide
Planning for
Hyper-V
PowerVault MD3000i
and MD3000 Storage
Arrays
www.dell.com/MD3000
www.dell.com/MD3000i
PLANNING FOR HYPER-V
Table of Contents
Planning for Hyper-V on the Dell PowerVault MD3000/MD3000i Storage Arrays ........................4
Disclaimer .................................................................................................................................4
Supported Hardware and Software...........................................................................................4
Planning for Hyper-V Virtualized Configurations...........................................................................5
Pre-installation Requirements ...................................................................................................5
Planning the SAS Direct Attach Installation with the Dell PowerVault MD3000 Storage Array 7
Planning the IP SAN Installation with the Dell PowerVault MD3000i........................................8
IP SAN Considerations .........................................................................................................8
IP SAN Network Infrastructure for Performance .................................................................10
IP SAN Optimization for Throughput and I/Os ....................................................................12
Considerations for Storage and Network Adapters (Dell PowerVault MD3000/MD3000i
Storage Arrays) .......................................................................................................................13
Connecting the Dell PowerVault MD3000 Storage Array.................................................13
Connecting the Dell PowerVault MD3000i Storage Array ................................................14
Additional Considerations for Virtual Storage .........................................................................15
Considerations When Using Pass-through Physical Disks or Virtual Hard Drives..................17
Directly-mapped iSCSI Virtual Disks to the Virtual Machine (Dell PowerVault MD3000i
Storage Array).....................................................................................................................18
Considerations for the Virtual Machine Child Partition........................................................18
Clustering with Hyper-V ......................................................................................................18
Contact Information.................................................................................................................22
Appendix A: System Troubleshooting .....................................................................................22
Appendix B: References .........................................................................................................22
Appendix C: Terminology........................................................................................................24
November 09 Page 2 PLANNING FOR HYPER-V
Table of Figures
Figure 1 Dell PowerVault MD3000 Storage Array Configuration ..................................................8
Figure 2 Dell PowerVault MD3000i Storage Array (IP SAN) ........................................................9
Figure 3 Dell PowerVault MD3000 Storage System (SAS) Cabling Configuration .....................14
Figure 4 Dell PowerVault MD3000i Storage Array Network Configuration .................................14
Figure 5 Virtual Storage Configuration........................................................................................16
Figure 6 Dell PowerVault MD3000 SAS Storage Array Configuration in a Clustered Environment
....................................................................................................................................................20
Figure 7 Dell PowerVault MD3000i iSCSI Storage Array Configuration in a Clustered
Environment................................................................................................................................21
November 09 Page 3
PLANNING FOR HYPER-V
Planning for Hyper-V on the Dell PowerVault
MD3000/MD3000i Storage Arrays
This document provides information about planning to implement Microsoft Hyper-V
technology with the Dell PowerVault MD3000/MD3000i storage arrays.
Hyper-V provides the software infrastructure and management tools in the Windows
Server 2008 operating system to create and manage a virtualized server computing
environment. The configuration in this document uses Windows Server 2008 Release 2
with the following enhancements to Hyper-V:
•
Live migration – A transparent move of virtual machines from one cluster node to
another with no loss of virtual machine connectivity between nodes. Live
migration uses clustered shared volumes (CSVs).
•
Dynamic virtual machine storage – Support for hot plug-in and hot removal of
storage.
•
Enhanced processor support – Up to 32 physical cores.
•
Enhanced networking support – Support for jumbo frames on virtual machines.
For information about how to deploy Hyper-V, refer to Implementing Hyper-V on the Dell
PowerVault MD3000/MD3000i. For information about backup and recovery options, refer
to Data Protection on the Dell PowerVault MD3000/MD3000i in a Hyper-V Environment.
This document provides useful information for system administrators and IT planners or
designers who are in the process of considering or planning to use Hyper-V in their
environment.
Disclaimer
The steps in this document were performed on a test setup. No live data was involved.
This document implies no warranties. These instructions are not guaranteed to work in
every situation. For the latest information about Dell IP SAN storage arrays, refer to the
Dell storage array installation guides. Go to:
http://www.dell.com
Supported Hardware and Software
Microsoft Hyper-V requires specific hardware. Refer to the Microsoft Windows Server
catalog for Hyper-V to find out if your systems qualify.
November 09 Page 4 PLANNING FOR HYPER-V
http://go.microsoft.com/fwlink/?LinkId=111228
You also can check the Dell-supported hardware for Microsoft Hyper-V contained in the
Dell Solutions Overview Guide for Microsoft Hyper-V.
http://support.dell.com/support/edocs/software/HyperV/en
If you need additional information, go to the following website.
http://www.dell.com/hyperv
Both the Dell PowerVault MD3000/MD3000i storage arrays are supported with Microsoft
Windows Server 2008 Release 2 running the Hyper-V role.
Planning for Hyper-V Virtualized Configurations
Planning for Hyper-V virtualized configurations includes these tasks:
ƒ
Making sure that pre-installation requirements are met
ƒ
Configuring the network adapter
ƒ
Deciding how to assign physical drives to virtual machines
ƒ
Deciding when to use pass-through physical disks or virtual hard drives (VHDs)
ƒ
Planning child partitions
ƒ
Deciding how to cluster
Pre-installation Requirements
To install and use the Hyper-V role, make sure you have the following processing
features on your system:
•
An x64-based processor.
Hyper-V is not available for 32-bit systems or Itanium-based systems.
You can download the AMD utility for Hyper-V compatibility from the following
website:
http://support.amd.com/us/Processor_TechDownloads/AMD-V_HyperV_Compatibility_Check_Utility.zip
•
Hardware-assisted virtualization.
Processors with Intel Virtualization Technology (Intel VT)
November 09 Page 5
PLANNING FOR HYPER-V
AMD Virtualization Technology (AMD-V)
•
Hardware-enforced data execution prevention (DEP) must be available and
enabled.
The Intel XD bit must be enabled (execution disable bit).
The AMD NX bit must be enabled (no execute bit).
Install the updates that you need. To receive a current list of all of the updates, go to:
http://technet.microsoft.com/en-us/library/dd430893.aspx
For more information about the Dell PowerVault MD3000i storage array, download one
of the sources listed on the following webpage.
http://www.dell.com/content/products/productdetails.aspx/pvaul_md3000i?c=us&cs=555
&l=en&s=biz&dgc=IR&cid=14054&lid=407200
For more information about how to plan for an IP SAN network with the Dell PowerVault
MD3000i storage array, refer to the IP-SAN-Best-Practices.
http://www.dell.com/downloads/global/solutions/public/white_papers/IP-SANBestPractice-WP.pdf
For more information about the Dell PowerVault MD3000 storage array, go to:
http://www.dell.com/content/products/productdetails.aspx/pvaul_md3000?c=us&cs=555
&l=en&s=biz&dgc=IR&cid=14054&lid=407200
November 09 Page 6 PLANNING FOR HYPER-V
Planning the SAS Direct Attach Installation with the
Dell PowerVault MD3000 Storage Array
The Dell PowerVault MD3000 storage array, with SAS host interface connectivity, can
connect with up to four servers. The configuration that you choose depends on the
number of servers you have, whether you are connecting to servers or a cluster node,
and the level of data redundancy that your configuration requires.
•
Redundancy
Redundancy is established by installing separate data paths between the server and
the storage array, in which each path connect to the different RAID controller
modules. Redundancy protects the server from losing access to data in the event of
path failure because both RAID controllers can access all of the disks in the storage
array. Figure 1 shows redundant, high-availability cabling configuration for two hosts.
This configuration is a best practice for Microsoft Hyper-V installations.
Note
November 09 Use two dual-ported host bus adapters (HBAs) and connect to one port from
each of these HBAs to improve availability. With such a configuration, a cable or
a SAS HBA could fail, and the access to the data is still guaranteed.
Page 7
PLANNING FOR HYPER-V
Figure 1 Dell PowerVault MD3000 Storage Array Configuration
•
RAID level
An appropriate RAID level should be chosen based on your applications. RAID 1 or
higher provides some level of redundancy that is useful in the case of failed physical
disks. Each RAID level works best with certain applications. Consider the RAID level
when you configure the Dell PowerVault MD3000 storage array.
•
Power
Connect the power supplies for the Dell PowerVault MD3000 storage array to
separate power sources. Doing this ensures that even if one component fails due to
a power issue, the alternative path continues to work.
Planning the IP SAN Installation with the Dell
PowerVault MD3000i
You can implement an IP SAN in various ways based on business requirements,
available resources, and intended application. For instance, one important but easily
overlooked item that can improve the manageability of your IP SAN implementation is to
assign a consistent and representative naming scheme to the storage arrays. This
assignment is especially useful if the SAN has more than one storage array attached.
You can use the “blink array” feature of the Modular Disk Storage Manager to correctly
identify each storage array physically.
This section contains some general implementation guidelines; however, these are
guidelines and might not benefit all applications.
IP SAN Considerations
Redundancy Redundancy requires having a second set of hardware and communication paths so
that if one piece of hardware on one path breaks down, a second path can be used.
In an IP SAN, a second controller in the storage array can perform redundancy and
redundancy is available when you use two different switches for the iSCSI network.
Figure 2 shows such a configuration.
November 09 Page 8 PLANNING FOR HYPER-V
Figure 2 Dell PowerVault MD3000i Storage Array (IP SAN)
Physical Network Infrastructure A fully redundant IP SAN is characterized by multiple physical independent iSCSI
data paths between the servers and the storage array. Make sure that each data
path is on a separate subnet.
iSCSI Sessions In the case of an iSCSI target like the Dell PowerVault MD3000i storage array,
establish multiple sessions to the storage subsystem from each server. Set up one
session per port from each of the network cards to each RAID controller module.
This method allows one session to restart if a link goes down and does not affect any
of the other links.
RAID Level Choose an appropriate RAID level based on your applications. RAID 1 or higher
RAID levels provide some level of redundancy that is useful in the case of failed
physical disks. Each RAID level works best with certain applications, so consider the
RAID level when you configure the Dell PowerVault MD3000i storage array.
November 09 Page 9
PLANNING FOR HYPER-V
Power Connect the MD3000i power supplies to separate power sources. This connection
ensures that if one component fails due to a power issue, the alternative path
continues to work.
Security To secure data across an IP SAN, implement a physically independent network for
the iSCSI data traffic, and isolate it from other Ethernet traffic. Having an isolated
network dedicated to storage traffic means that the storage array avoids network
congestion with other non-storage traffic using the network.
-
CHAP
To have secure access between your host and storage array, enable target and
mutual CHAP authentication on the hosts and storage arrays. Follow standard
CHAP password guidelines for best security. Set a strong password that meets
standard ID guidelines on all of the devices in your IP SAN.
Virtual LAN (VLAN) If you want to physically isolate your IP SAN data traffic from general network traffic,
use VLANs. Turn on VLAN tagging on the Dell PowerVault MD3000i storage array. A
port can either transmit all tagged IP packets or all non-tagged IP packets.
Note: You must enable a VLAN throughout the entire iSCSI SAN from the network
interface cards (NICs) and switches to iSCSI ports; otherwise, behavior might be
inconsistent. To simplify and avoid troubleshooting at deployment, make sure
that the NICs, switches, and storage array are fully operational before you enable
the VLAN feature for the entire solution.
IP SAN Network Infrastructure for Performance
This section describes some general implementation guidelines for performance. These
guidelines are general and might not apply to all applications. Review your application
requirements prior to implementing these guidelines.
General Network Practices Make sure the category rating for the cables used are gigabit-Ethernet-compliant
(CAT5e or CAT6). Design your network to have the least amount of hops between
the storage arrays and the servers. This design greatly reduces the failure points,
and it simplifies the manageability and reduces the latency and complexity of your
network architecture (particularly in the area of redundancy).
Use managed switches because they provide advance features to help you optimize
and maintain your network for your application. Use auto-negotiation only because
November 09 Page 10 PLANNING FOR HYPER-V
gigabit Ethernet networks are designed to always have auto negotiation enabled. If a
particular application requires a specific speed or duplex mode, you can accomplish
this by changing the advertisement options of the switch.
Spanning Tree Protocol Disable the spanning-tree protocol (STP) on the switch ports that connect end nodes
(iSCSI initiators and storage array network interfaces). If you still decide to enable
STP on the switch ports, turn on the STP FastPort feature on the ports to allow
immediate transition of the ports into forwarding state.
Note: FastPort immediately transitions the port into the STP forwarding mode upon
linkup. Because the port still participates in STP and it is to be a part of the
loop, it eventually transitions into the STP blocking mode.
Note: PowerConnect switches default to Rapid Spanning Tree Protocol (RSTP),
which is an evolution in STP that provides for faster spanning tree
convergence and is preferable to STP when available.
Note: Use STP for a single cable connection between switches or for trunking for
multiple cable connections between switches.
TCP Congestion Avoidance TCP congestion avoidance is an end-to-end flow control protocol that limits the
amount of data sent between a TCP sender and a TCP transmitter. This protocol
uses a sliding window to size the data being sent to the TCP receiver. This protocol
starts with a small segment size and keeps increasing with each acknowledged
segment sent until a segment is dropped. After a segment is dropped, TCP starts
this avoidance over again.
Ether Flow Control It is best practice to enable flow control on the switch ports that handle iSCSI traffic.
In addition, if a server is using a software iSCSI initiator-and-NIC combination to
handle iSCSI traffic, enable flow control on the NICs to also obtain a performance
benefit. In SAN configurations in which many servers (initiators) are communicating
with storage devices, if senders transmit data simultaneously, the data transmission
might exceed the throughput capacity of the receiver. When this event occurs, the
receiver might drop packets, which forces the senders to retransmit the data after a
delay. Although this event does not result in any loss of data, latency increases
because of the re-transmissions, and I/O performance degrades. By enabling flow
control, you can avoid the imbalance in the network traffic between the devices that
send network traffic and the devices that receive the traffic.
November 09 Page 11
PLANNING FOR HYPER-V
Note: PowerConnect switches default to flow control being off. The Dell PowerVault
MD3000i storage array auto-configures to the switch when the flow control is
turned on.
Unicast Storm Control A traffic “storm” occurs when a large outpouring of IP packets creates excessive
network traffic that degrades network performance. Many switches have traffic storm
control features that prevent ports from being disrupted by broadcast, multicast, or
unicast traffic storms on physical interfaces. These features typically work by
discarding network packets when the traffic on an interface reaches a percentage of
the overall load (usually 80 percent, by default). Because iSCSI traffic is unicast
traffic and can typically use the entire link, disable unicast storm control on switches
that handle iSCSI traffic. However, the use of broadcast and multicast storm control
is encouraged. Refer to your switch documentation for information on disabling
unicast storm control
Jumbo Frames Most enterprise gigabit-Ethernet equipment provides some support for jumbo frames.
Enabling jumbo frames accelerates iSCSI performance by about five percent and
reduces server CPU utilization by two percent to three percent with standard or
smarter NICs. You should enable jumbo frames on the switch ports that handle
iSCSI traffic. In addition, if a server is using a software iSCSI initiator-and-NIC
combination to handle iSCSI traffic, you must also enable jumbo frames on the NICs
to obtain the performance benefit and ensure consistent behavior.
Note: You must enable jumbo frames throughout the entire IP SAN from the NICs,
switches, and array ports; otherwise, behavior might be inconsistent. To
simplify troubleshooting initial deployments, make sure that the NICs,
switches, and storage arrays are fully operational before you enable jumbo
frames.
IP SAN Optimization for Throughput and I/Os
When designing your IP SAN, review the various factors in your network and the
applications that you are using. Consider these general rules when you design your
IP SAN.
To maximize the data throughput of your storage array, all data ports need to be
used. If your application is I/O intensive, use an iSCSI offload NIC. Consider
manually balancing your virtual disk ownership so that no single controller is
processing an excessive amount of I/O relative to the other controller. The Dell
PowerVault MD3000i storage system supports active/active controllers, with each
controller being able to simultaneous process I/O. The asymmetric design of the
controllers means that a virtual disk (logical unit number or LUN) is owned by a
controller, and all I/O access to the virtual disk is only possible through the owning
controller. To take advantage of both the controllers for I/O access, distribute virtual
November 09 Page 12 PLANNING FOR HYPER-V
disks among the controllers and modify them to balance I/O access so as to balance
controller utilization.
Layer 3 Optimization Differentiated Services (DiffServ) provides a good method for managing your
network traffic. Some switches have a proprietary implementation of this that is
called Quality of Service (QoS). DiffServ uses the Differentiated Services Code Point
(DSCP) to distinguish between service levels of each IP connection. These service
level agreements are on a Per-Hop Basis (PHB). As such, within the internal
corporate network traffic flows can be predictable, but after a WAN link leaves the
company, the service agreements are no longer valid. Four levels are typically used
with DiffServ.
1. Default PHB – Typically this level is best-effort traffic.
2. Expedited Forwarding (EF) PHB – Good for low-loss, low-latency traffic.
3. Assured Forwarding (AF) – Behavior group.
4. Class Selector PHBs – Maintain backward compatibility with the IP
precedence field.
To choose what service level to use, examine the needs of the applications
connected to the storage array. For instance, if you have your servers set up to
iSCSI boot, or are using virtualization to “hide” the storage array (and the guest OS is
booting off a C: drive that is actually on the storage array), select EF because the
data must get to the server and, if there is latency, the server locks up. On the other
hand, you might want all of your traffic coming in from the Worldwide Web set to the
lowest possible class of AF so the service level does not affect your critical data.
Considerations for Storage and Network Adapters
(Dell PowerVault MD3000/MD3000i Storage Arrays)
When planning for the network configuration, make sure that you have at least two
network adapters assigned for your virtual machine traffic. If you are planning to add
MS-Exchange, MS-SQL, MS-SharePoint, or a similar configuration to your virtual
machines, use two dedicated network connections for each of those applications. For
example, if you have a file server, MS-Exchange, and a SQL server running on virtual
machines, you should have six network connections (two for MS-Exchange, two for the
SQL server, one for the file server, and one for managing the Hyper-V server). For
clustering, have two additional network connections for the cluster heartbeat.
Connecting the Dell PowerVault MD3000 Storage Array
The connection to the Dell PowerVault MD3000 storage array is SAS based. Figure 3
shows how the SAS storage cabling is configured.
November 09 Page 13
PLANNING FOR HYPER-V
Figure 3 Dell PowerVault MD3000 Storage System (SAS) Cabling Configuration
Connecting the Dell PowerVault MD3000i Storage Array
For the iSCSI-based Dell PowerVault MD3000i storage array, you must have two NICs
for I/O to the storage array for each server. Set up the management ports (out of band)
for managing the MD3000i storage array because they are required for the initial
configuration. For best performance when planning the network connection to the Dell
PowerVault MD3000i storage array, make sure that the cabling looks similar to Figure 4.
Figure 4 Dell PowerVault MD3000i Storage Array Network Configuration
November 09 Page 14 PLANNING FOR HYPER-V
The Module Disk Storage Manager comes with the device-specific module (DSM) for
Microsoft multipath I/O (MPIO). MPIO provides load balancing between the two ports of
one RAID controller module. For best performance, make sure that load balancing is
enabled and active. Load balancing lets you use up to 2 Gb/s for each virtual disk on the
Dell PowerVault MD3000i storage array.
MD3000i Load‐Balance Policies You can change the load-balance policy using the Windows 2008 Disk Management
option. From the server desktop, right-click the My Computer icon, and select
<Manage> to open the Computer Management dialog. Click <Disk Management> to
show the list of virtual disks that are attached to the server. Right-click the virtual disk on
which you would like to set the load balance policy, and select <Properties>. From the
MPIO tab, select the load-balance policy that you want to set for this virtual disk
You can choose one of these load-balance policies to optimize I/O performance.
Round Robin with Subset
The round-robin with subset I/O load balance policy routes I/O requests, in rotation, to
each available data path to the RAID controller module that owns the virtual disks. This
policy treats all paths to the RAID controller module that owns the virtual disk equally for
I/O activity. Paths to the secondary RAID controller module are ignored until the
ownership changes. The basic assumption for the round-robin policy is that the data
paths are equal.
Least Queue Depth with Subset
The least queue depth with subset policy is also known as the least I/Os, or least
requests, policy. This policy routes the next I/O request to a data path that has the least
outstanding I/O requests queued. For this policy, an I/O request is simply a command in
the queue. The type of command or the number of blocks that are associated with the
command are not considered. The least queue depth with subset policy treats large
block requests and small block requests equally. The data path selected is the path
group of the RAID controller module that owns the virtual disk.
Least Path Weight with Subset
The least path weight with subset policy assigns a weight factor to each data path to a
virtual disk. An I/O request is routed to the path with the lowest weight value to the RAID
controller module that owns the virtual disk. If more than one data path to the virtual disk
has the same weight value, the round-robin with subset path selection policy is used to
route I/O requests between paths with the same weight value.
Additional Considerations for Virtual Storage
Multiple ways exist to assign physical storage to a virtual machine. For the Dell
PowerVault MD3000/MD3000i storage array, you have two options:
November 09 Page 15
PLANNING FOR HYPER-V
•
Assign the virtual disks to the parent server, and create a file system and a virtual
hard drive (VHD) for the virtual machine.
•
Assign an unformatted physical disk directly to the virtual machine (pass-through
physical disk).
With the Dell PowerVault MD3000i storage array, in addition to the previously mentioned
two options, you have a third option: You can assign a physical disk directly to the virtual
machine by installing the iSCSI software initiator on the virtual machine as you would on
a physical server.
LUN1 – VHD Method
LUN2 – Pass-through Method
LUN3 – Storage Direct Method
Figure 5 Virtual Storage Configuration1
VHDs have several different options:
•
Dynamic expansion – The VHD file grows as data is stored to the physical disk
with a maximum size not exceeding the size specified during creation. When
using the new virtual machine wizard in Hyper-V Manager, select <Dynamically
Expanding> when creating a new virtual machine.
•
Fixed size – The VHD file uses the full amount of space specified during creation.
•
Differencing – The VHD file exists as a child disk in a parent/child relationship
with an existing VHD. The parent VHD is left intact, and changes that occur are
stored on the child (differencing) VHD.
1
2009, Jose Barreto’s Blog, http://blogs.technet.com/josebda/archive/2008/03/06/more-onstorage-options-for-windows-server-2008-s-hyper-v.aspx.
November 09 Page 16 PLANNING FOR HYPER-V
Note: For the best performance in production environments, use fixed-size VHDs
because they eliminate the overhead created by expanding the VHD in a normal
production environment.
Size each LUN to support the target VHD. To avoid warning messages generated by the
parent partition about the physical disk running out of capacity, allocate an additional
200 MB to the LUN, and leave this additional 200 MB as empty space on the virtual disk.
Considerations When Using Pass-through Physical
Disks or Virtual Hard Drives
Keep the following considerations in mind when you use pass-through physical disks or
VHDs.
•
VHDs support the ability to export the drive to another parent server.
•
VHDs support VM checkpoints.
•
Pass-through physical disks offer slightly higher performance than VHD devices
and support virtual disk sizes greater than 2 TB. However, pass-through physical
disks do not support VSS-based backup or checkpoints. VSS-based backup with
hardware providers works as long as the backup is initiated from the parent.
Inside the virtual machine, only software-based VSS backup is supported.
•
You can use both VHDs and pass-through physical disks in clustered
environments.
•
If you are booting from a pass-through physical disk, you must store the VM files
in a different location.
•
For every virtual machine, you must create a separate pass-through physical
disk. You cannot share a pass-through physical disk for two virtual machines.
•
The physical layout of the pass-through physical disk is the same as for any
physical server.
Note: For the highest performance, use pass-through physical disks but only as data
devices. If you need the best flexibility, assign fixed-size VHDs.
For more information about how to maximize performance for different applications on
the Dell PowerVault MD3000/MD3000i storage arrays, go to:
http://www.dell.com/downloads/global/products/pvaul/en/pvault_md3000_whitepaper.pdf
Note: The configuration of your physical disk layout for performance on the Dell
PowerVault MD3000/MD3000i storage array depends on the application
requirements of your virtual machine (child partition), not on the Hyper-V parent
server.
November 09 Page 17
PLANNING FOR HYPER-V
Directly-mapped iSCSI Virtual Disks to the Virtual Machine (Dell
PowerVault MD3000i Storage Array)
You can easily transfer the virtual disk from a physical server to the virtual machine.
However, transferring the virtual disk reduces the number of servers connected to the
Dell PowerVault MD3000i storage array. Every virtual machine uses exactly one server
partition. VSS-based backup from the parent does not include the directly mapped iSCSI
LUNs of the child partition. Booting also is not possible unless you use a third-party
software. The Dell PowerVault MD3000i storage array only requires a software initiator.
The iSCSI HBA is not supported with the Dell PowerVault MD3000i storage array. Only
software initiators are supported with the Dell PowerVault MD3000i storage array.
Note: Direct mapping can be used for data devices that were originally used on a
physical server or for data devices that are being moved back to a physical
server.
Considerations for the Virtual Machine Child Partition
Usually, when you create virtual hard drives, you can store your virtual machine
configuration files in the same location as the virtual hard drives. If you plan to use
multiple virtual machines, you can create one virtual disk on your Dell PowerVault
MD3000/3000i storage array for storing your configuration files and assign the data
physical disks separately. Depending on the performance required for your different
virtual machines, you might use one MD3000/3000i virtual disk for multiple virtual
machine data disks. When performing capacity calculations, add approximately 200 MB
to the virtual machine configuration data for each virtual machine. When configuring the
virtual machine, booting is supported only on virtual IDE adapters, not on the virtual
SCSI adapter. However, you can assign an MD3000/3000i virtual disk as an IDE device
to the virtual machine, and performance is not affected.
If you plan to use pass-through physical disks as boot disks and data disks, you must
store your virtual machine configuration files in a different location. This location can be
the local physical disk of your server, or an MD3000/3000i virtual disk, which is available
to the parent and formatted with NTFS. You can assign a pass-through device only after
creating the virtual machine. To assign a pass-through device, select <Assign Hard
Drive Later> during the virtual machine creation process. After the virtual machine is
created, change the virtual machine options to <Assign the Pass-Through Device>.
Note: The pass-through physical disk must have a signature and be in an offline state
on the parent so that you can assign the physical disk to a virtual machine.
Clustering with Hyper-V
If you want to use clustering for Hyper-V, carefully observe the following requirements.
•
November 09 Use identical server hardware to eliminate issues regarding failover.
Page 18 PLANNING FOR HYPER-V
•
Avoid having a single point of failure in your communications network. Connect
cluster-nodes through a network that uses redundant switches or similar
hardware.
Note: You can connect cluster nodes with a single network. However, it is best
to use a redundant communications network.
•
All cluster disks must be basic disks. Dynamic disks are not supported.
MD3000-specific cluster requirements:
•
Only use two dual-ported SAS adapters for each server, and connect one port of
each adapter to the Dell PowerVault MD3000 storage array in clustered
environments to achieve the highest level of redundancy.
MD3000i-specific cluster requirements:
•
The storage configuration must be redundant. Use only two or more network
connections for iSCSI traffic to the Dell PowerVault MD3000i storage array.
•
You must have dedicated iSCSI network adapters for iSCSI environments. You
cannot share communication and iSCSI traffic on the same adapter in clustered
configurations.
All cluster-nodes must be able to communicate with the same virtual disks on the Dell
PowerVault MD3000/MD3000i storage array. You can use Storage Partitioning to create
a server group with the cluster nodes and assign the necessary virtual disks to them.
Make sure that only the cluster-nodes have access to those virtual disks and that other
servers do not have access to those virtual disks.
When planning, create an additional virtual disk for the quorum disk. Make this virtual
disk at least 1 GB in size. This disk contains the quorum information and is used as the
witness disk. Format the quorum virtual disk with NTFS.
If you are planning to use pass-through devices on your virtual machines, you must
make sure that the virtual machine configuration files are also stored on the shared
storage and that the shared storage is available to all cluster nodes. You can create an
additional virtual disk on the Dell PowerVault MD3000/MD3000i storage array or
increase the size of the quorum virtual disk to store the configuration files there. You
also might create an additional virtual disk with approximately 300 MB to 500 MB
allocated for each virtual machine. For example, 10 virtual machines equal a 5-GB disk
size.
November 09 Page 19
PLANNING FOR HYPER-V
Figure 6 Dell PowerVault MD3000 SAS Storage Array Configuration in a Clustered
Environment
November 09 Page 20 PLANNING FOR HYPER-V
Figure 7 Dell PowerVault MD3000i iSCSI Storage Array Configuration in a Clustered
Environment
November 09 Page 21
PLANNING FOR HYPER-V
Contact Information
Go to:
http://www.dell.com/
Or call 1-800-WWW-DELL.
Appendix A: System Troubleshooting
For information about troubleshooting MD3000 storage arrays, refer to the
“Troubleshooting Problems” chapter of the Dell™ PowerVault™ Modular Disk Storage
Manager User’s Guide. Go to:
http://support.dell.com/support/edocs/systems/md3000/en/index.htm
Appendix B: References
Cisco, undated. Flooding Control
http://www.cisco.com/univercd/cc/td/doc/product/lan/c2900xl/29_35xu/olhelp/floodctl.htm
Dell, undated. Dell™ PowerVault™ Modular Disk Storage Manager CLI Guide
http://support.dell.com/support/edocs/systems/md3000/en/index.htm.
Dell, 2007. Manuals: Dell™PowerVault™MD3000
http://support.dell.com/support/edocs/systems/md3000/en/index.htm
Dell, 2007. PowerVault MD3000i SAN Array for Storage Consolidation
http://www.dell.com/content/products/productdetails.aspx/pvaul_md3000i?c=us&1=en&s
=bsd&cs=04
Dell, 2007. Using iSCSI: Dell™ PowerVault™ Modular Disk Storage Manager User’s
Guide
http://support.dell.com/support/edocs/systems/md3000/en/UG/HTML/iscsi.htm
Microsoft, 2007. How to Align Exchange I/O with Storage Track Boundaries
http://technet.microsoft.com/en-us/library/aa998219.aspx
Microsoft, 2007. Planning Storage Configurations for Exchange 2007
November 09 Page 22 PLANNING FOR HYPER-V
http://technet.microsoft.com/en-us/library/bb124518.aspx
Microsoft, 2008. Microsoft Storage Technologies – iSCSI
http://www.microsoft.com/windowsserver2003/technologies/storage/iscsi/default.mspx
(FAQs and more information on Microsoft iSCSI)
Microsoft Download Center, 2007. Microsoft iSNS Server 3.0
http://www.microsoft.com/downloads/details.aspx?familyid=0dbc4af5-9410-4080-a545f90b45650e20&displaylang=en
Microsoft Exchange Team Blog, 2007. Exchange 2007 Mailbox Server Role Storage
Requirements Calculator
http://msexchangeteam.com/archive/2007/01/15/432207.aspx
November 09 Page 23
PLANNING FOR HYPER-V
Appendix C: Terminology
Bandwidth – The amount of data that the storage array can process over time.
Bandwidth is measured in megabytes per second (MB/s).
Child partition – A partition that is associated with the parent partition. The child
partition is subordinate to the parent partition.
Initiator – The system component that originates an input/output (I/O) command over an
I/O bus or network. I/O adapters, network interface cards, and intelligent RAID controller
module device I/O bus control application specific integrated circuits (ASICs) are typical
initiators. A peripheral device is a target. (The Dictionary of Storage Networking
Terminology)
Latency – The interval of time between submitting a request and receiving a response.
MPIO (Multipath I/O) – In computer storage, an arrangement whereby more than one
logical path exists between the central processing unit (CPU) in a computer system and
its storage devices. This path is routed through the buses and bridge devices that
connect the CPU and its storage devices. If one RAID controller module, port, or switch
fails, the operating system can route I/O through the remaining RAID controller module
so that work can continue.
NIC (network interface card) – An adapter that connects an intelligent device to a
network. The common terminology is a network interface card or an Ethernet NIC
network.
Node – An addressable entity connected to an input/output (I/O) bus or network and
used primarily to refer to computers, storage devices, and storage arrays. The
component of a node that connects to the bus or network is a port. (The Dictionary of
Storage Networking Terminology).
Parent Partition – The primary partition. The parent partition creates links to child
partitions, and has access to those child partitions.
Pass-through Physical Disk – An unformatted physical disk directly assigned to the
virtual machine.
Portal – A service that links initiators and targets when their IP addresses are not
known.
Quorum Virtual Disk – A quorum virtual disk is a virtual disk on the shared storage
array that stores the quorum information.
Response time – The interval of time between submitting a request and receiving a
response.
Session – A group of iSCSI connections. As each connection is established, the session
can negotiate its own unique parameters.
November 09 Page 24 PLANNING FOR HYPER-V
Signature – An embedded code that identifies the device. The signature is used when
passing information between devices.
Target – The storage destination. In this document, the target is the MD3000i storage
array.
TCP/IP (Transmission Control Protocol/ Internet Protocol – A suite of protocols that
includes Transmission Control Protocol (TCP), Internet Protocol (IP), User Datagram
Protocol (UDP), and Internet Control Message Protocol (ICMP). These protocols form
the basic set of communication protocols used on the Internet. (The Dictionary of
Storage Networking Terminology)
Throughput – The number of individual I/Os the storage array can process over time.
Throughput is measured in I/Os per second (IOPS).
November 09 Page 25
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertising