Section 9 - Configure and Administer vSphere
Availability Solutions
Objective 9.1 -
Configure Advanced vSphere HA Features
Enable/Disable
Advanced vSphere HA Settings:
vSphere HA >
Edit > Edit Cluster Settings > Advanced Options > Add
das.isolationAddress[...]
- the address to ping to determine if a host is isolated from the network
das.useDefaultIsolationAddress
{true/false} - by default HA uses the default gateway of the console network as
an isolation address
das.isolationShutdownTimeout
- period of time the system waits for a VM to shut down before powering it
off (only applies if host’s isolation response is ‘Shut Down’). Default value
300 seconds.
das.slotMemInMb -
Defines the maximum bound on the memory slot size
das.slotCPUinMhz -
Defines the maximum bound on the CPU slot size
das.vmMemoryMinMb -
Defines the default memory resource value assigned to a VM if its memory
reservation is 0 (default 0 MB)
das.vmCPUminMhz -
Defines a default CPU resource value assigned to a VM if its CPU reservation is
0 (default 32 MHz)
das.ioStatsInterval
- Change the default I/O stats interval for VM Monitoring sensitivity (default
120 seconds)
das.ignoreInsufficientHbDatastore
- Disables config issues created if host does not have sufficient heartbeat
datastores
das.heartbeatDsPerHost
- Change number of heartbeat datastores required (default 2)
fdm.isolationPolicyDelaySec
- Wait time in seconds before executing the isolation policy
das.respectVmVmAntiAffinityRules
- Determines if HA enforces VM-VM anti-affinity rules
das.maxResets -
Max number of reset attempts made by VMCP
das.maxTerminates
- Max number of retries made by VMCP for VM termination
das.terminateRetryIntervalSec
- If VMCP fails to terminate a VM, wait time in seconds before retry
das.config.fdm.reportFailoverFailEvent
- When set to 1, enables generation of a detailed per-VM event when an attempt
by HA to restart a VM is unsuccessful (default 0)
vpxd.das.completeMetaDataUpdateIntervalSec
- seconds after a VM-Host affinity rule is set during which HA can restart
a VM (default 300 seconds)
das.config.fdm.memReservationMb
- by default vSphere HA agents run with memory limit of 250 MB - use this
option to lower the limit
vSphere HA
utilizes the concept of master and slave hosts to build out a HA cluster:
- hosts communicate with each other using network
heartbeats every second
- master host is responsible for detecting failure of
slave hosts in the cluster
- if master stops receiving heartbeats from a slave, it
checks slave’s datastore heartbeats, tries ICMP ping of slave’s management IP;
if all checks fail, slave is considered failed and its VMs are restarted
Identify ‘VM
Overrides’:
Each VM in a
vSphere HA cluster is assigned cluster default settings for:
- VM Restart Priority
- Host Isolation Response
- VM Component Protection
- VM Monitoring
VM Component
Protection (VMCP) Settings: Protects against datastore accessibility
failures
- PDL: Permanent
Device Loss - storage device reports datastore is no longer accessible
- APD: All Paths
Down - transient or unknown accessibility loss, or unidentified delay in
I/O processing
Configure: Tick ‘Protect against Storage Connectivity
Loss’
- PDL Failures:
VM is automatically failed over to a new host unless VMCP is configured to only
‘Issue Events’
- ADP Events:
‘Delay for VM failover’ then Conservative or Aggressive restart
Objective 9.2 -
Configure Advanced vSphere DRS Features
Identify
Distributed Resource Scheduler (DRS) Affinity Rules:
- VM-Host Affinity (or Anti-Affinity)
- VM-VM Affinity (or Anti-Affinity)
Affinity = DRS
tries to keep (groups) together
Anti-Affinity = DRS
tries to keep (groups) apart
Note: Types of DRS
Group = ‘Host Group’ or ‘VM Group’
VM-Host rule
specifications:
- Must run on hosts in group
- Should run on hosts in group
- Must not run on hosts in group
- Should not run on hosts in group
VM-VM rule types:
- Separate Virtual Machines
- Keep Virtual Machines Together
Note: Create a VM to VM anti-affinity rule -
to make sure that in the event of hardware failure of a host, VMs are not on
the same host at the same time.
Configure
Distributed Resource Scheduler (DRS) Automation Levels:
Manual: Initial
placement - Recommended host is displayed
Manual: Migration
- Recommendation is displayed
Partially Automated: Initial
placement - Automatic
Partially Automated: Migration
- Recommendation is displayed
Fully Automated: Initial
Placement - Automatic
Fully Automated: Migration
- Recommendation is executed automatically
Comments
Post a Comment