Module 11: High Availability and Fault Tolerance
vCenter Server Availability:
– Cluster DB
– Multiple redundant servers
– HA, vCenter server heartbeat
A cluster enabled for vSphere HA can have:
– up to 32 hosts per cluster
– up to 512 VM per host (regardless of the number of host/cluster)
– up to 3,000 VM per cluster
HA Failure scenario: Host, Guest and Application
Admission control policy: Host failures cluster tolerates/ Percentage of cluster resources reserved as failover spare capacity/ Specify failover hosts
VMware recommends redundant heartbeat networking for vSphere HA cluster
Before changing the networking configuration on the ESXi host (adding port group, removing vSwitches):
– Deselect Enable Host Monitoring
– Place host in maintenance mode
vSphere HA Architecture
FDM (Fault Domain Manager) service
Master/Slave host
HA Failure Scenarios:
– failed slave host (Network heartbeats/ Datastore heartbeats)
– failed master host (election process)
– Isolated host
– Network partition
Fault Tolerance is used for mission critical server that can not tolerate downtime
– no more than 4 FT VM on a single host
– store ISO files on shared storage
– disable BIOS-based power management
Module 12: Scalability
DRS (Distributed Resource Scheduler)
– Initial placement
– Load balancing
– Power management
EVC is a cluster feature that prevents vMotion migrations from failing because of incompatible CPUs
For best performance, VMware recommends that you store the swap file in the same directory as the virtual machines.
Affinity/ Anti-affinity rules: Preferential vs. Required
Storage DRS requires that storage I/O control be enabled on all the datastores (I/O load history is checked once every 8 hours)
Maintenance mode/ Standby mode
VM performance tuning
Module 13: Patch Management
Update Manager
5 types of baselines:
– Host patch
– Host extension
– Host upgrade
– Virtual machine patch
– Virtual appliance upgrade
Module 14: Installing VMware Components
Installing ESXi
Install vCenter server
Jackie Chen's IT Workshop 