AZURE202x Microsoft Azure Virtual Machines – Module 3

January 12, 2017

AZURE202x Microsoft Azure Virtual Machines – Module 3 에서 제공하는 무료 강좌인 Microsoft Azure Virtual Machines 의 모듈 3 부분에 대한 정리입니다.

Public IP Address

  • For public-facing services
  • VMs, Internet facing load balancers, VPN gateways and application gateways
  • Can configure on the public IP resource
  • dynamic address
    • IP is not allocated on its creation, but assigned on starting and released on stopping
  • static address
    • IP is assigned on its creation, and released on deletion or on changing to dynamic

Private IP Address

  • Within Azure VNet
  • To extend on-prem with VPN Gateway or ExpressRoute
  • VMs, internal load balancers and application gateways
  • Can configure on the network interface resource
  • We can’t choose a specific IP address
  • dynamic address
    • default method assigned by resource’s subnet via DHCP. Changes when start/stop the resource.
  • static address
    • can set static with a valid IP within the resource’s subnet.

Availability Set

  • Logical grouping of 2+ VMs to prepare planned/unplanned failures
  • Availability Set Principles
    • Multiple VMs in one Availability Set for redundancy
    • Each application tier into separate Avaliability Sets
    • Load Balancer with Availability Sets combination
  • Each VM in one Availability Set is placed in one update domain and two fault domains

Update Domains

  • To perform incremental or rolling updates during the deployment
  • Set of VMs with physical hardware that can be updated/rebooted at the same time
  • Only one update domain is rebooted at a time.
  • 5 UD by default, but up to 20 UDs

Fault Domains

  • To share a common set of hardware, switches for single point of failures
  • VMs in AS are placed in at least two FDs to avoid affects from hardware failures, network outages, power interruptions or software updates

Scale Set

  • Azure Compute resource to deploy/manage a set of identical VMs
  • Support true auto-scale without provisioning to build large-scale targeting big compute, big data, centralised workloads
  • Often integrated with Azure Insight to measure scale up/down, Load Balancer, and NAT rules to spread workload over the available VMs
  • Guidelines
    • Automatically created with load balancer NAT rules to enable SSH or RDP
    • Scale Set varies between 0 and 100 VMs and easily change the number of VMs
    • Set Max/Min number of VMs and define triggers based on resource consumption
    • On scaling out, VMs are balanced across update/fault domains for maximum availability
    • On scaling in, VMs are removed with maximum availability