Sunteți pe pagina 1din 17

Clustering Planning

Microsoft IT Camps - Virtualization

Hyper-V High Availability


Load Balancing
Windows Server Network Load Balancing (NLB)
Simple L2 software-based load-balancer Load-balance within a host and across hosts (assuming switches support NLB)

Hardware Load Balancer


Full-featured, but often expensive

Replication
Copy VMs to another host

Backup / Restore
Save running VM Restore if it crashes on the same host or in a different datacenter

Hyper-V Primarily uses Failover Clustering

Failover Clustering & Hyper-V


VM High-Availability
Automatic failover & recovery Hardware health detection Host OS health detection VM health detection Application/service health detection

VM Mobility
Live migration Quick migration Failover

Deployment Flexibility
VM not bound to a single host

Failover Clustering SKUs


Windows Server 2008 R2
Enterprise Datacenter Itanium

Microsoft Hyper-V Server 2008 R2


FREE Host OS Download it today! Built on Server Core for even higher availability from fewer updates Failover Clustering, Hyper-V, CSV, Live Migration More information: http://www.microsoft.com/hvs
SKU Differences Failover Cluster Nodes VM Guest Licenses Enterprise 16 4 Datacenter 16 Unlimited Itanium 8 Unlimited* Hyper-V Server 16 0
* Itanium requires a 3rd party hypervisor

Failover Clustering Basics


1. 2. 3. 4. 5. 6. 7. 8. 9. 10. Deploy redundant hardware everywhere Shared storage accessible by all nodes Clustered application writes data to shared storage so all nodes can assess it A registry hive containing current cluster configuration information is replicated to each node Nodes monitor the health of other nodes If a node fails, the health check will fail, and a failover will happen Another node knows what the failed node was hosting through the registry and start that application The application reads its saved data from the shared storage Client connects to app on new node Clients may experience a slight interruption in service during failover but none during a live migration (of a VM)

Simulated Failover
Public
VMs & Workloads

Shared Storage

Cluster Storage
Fibre Channel (FC)
Fibre Channel over Ethernet (FCoE)

Serial-Attached SCSI (SAS)

iSCSI

Requires a Storage Area Network (SAN) SCSI-3 Commands Basic GPT and MBR disks supported Multipath IO (MPIO) recommended Persistent Reservations (PRs) required

Cluster Networking
Integrated with Windows Server TCP/IP Stack Multiple networks for added redundancy Separate networks by function
Public Network Internal Cluster Network (health checks, cluster communication) Hyper-V Live Migration & Management Storage (iSCSI, FCoE) Backup

IPv4, IPv6 & IPv6 Tunneling DHCP or Static IP Addresses

Quorum Overview

Cluster membership model to avoid partitions and split-brain Majority is greater than 50% Always have an odd number of votes and at least 3 voters Possible voters: Nodes (1 each) + 1 Witness (Disk or File Share) 4 Quorum Types
Node Majority Node and Disk Majority
Vote Vote

Node and File Share Majority Disk Only (not recommended)


Vote

Vote

Vote

Cluster Disk Disk Witness

Remote File Share File Share Witness

Supported Roles and Features


Common
Hyper-V SQL Exchange File Server DFS-Namespace DFS-Replication DHCP DTC iSNS MSMQ NFS Print Remote Desktop Connection Broker WINS

Custom Scripts
Active Directory CS Active Directory RMS HPC IIS WWW & FTP iSCSI Target Generic Application Generic Script Generic Service Other Server Custom LOB Applications Etc

Also Ship In Box


Generic Containers

3rd Party

Cluster Groups
VMs
Virtual Machine Virtual Machine Configuration File
Information about the VMs features Cluster disk visible by all nodes

Cluster Shared Volumes

Other Workloads (File Server, DHCP, etc.)


Network Name (NetName)
Easy name for clients to connect

IP Address

IP address associated with NetName Uses DHCP if enabled in environment


Highly-available service or app Shared storage so all nodes can access data

Service / Application Physical Disk Resource

Host vs. Guest Clustering


Host Clustering
Cluster service runs inside (physical) host and manages VMs VMs move between cluster nodes

Guest Clustering
Cluster service runs inside a VM Apps and services inside the VM are managed by the cluster Apps move between clustered VMs

Cluster Cluster SAN iSCSI

What Host Clustering Delivers


Avoids a single point of failure when consolidating Survive Host Crashes
VMs restarted on another node

Restart VM Crashes
VM OS restarted on same node

Recover VM Hangs
VM OS restarted on same node

Cluster

Zero Downtime Maintenance & Patching


Live migrate VMs to other hosts

SAN

Mobility & Load Distribution


Live migrate VMs to different servers to load balance

What Guest Clustering Delivers


Application Health Monitoring
App or service within VM crashes or hangs and moves to another VM

Application Mobility
Apps or services moves to another VM for maintenance or patching of guest OS

Cluster

iSCSI

Combining Host & Guest Clustering


Best of both worlds for flexibility and protection
VM high-availability & mobility between physical nodes Application & service high-availability & mobility between VMs

Increases complexity

Guest Cluster

CLUSTER SAN

iSCSI

CLUSTER SAN

Mixing Physical and Virtual


Mixing physical & virtual nodes is supported
Must still pass Validate

Requires iSCSI storage Scenarios


Spare node is a VM in a Private Cloud Shared host running multiple spare nodes as VMs Test environment with limited hardware

iSCSI

Takeaways
Failover Clustering is the primary highavailability technology for Hyper-V Clustering has extra considerations around shared storage and networking Almost any datacenter workload can run on a Failover Cluster

S-ar putea să vă placă și