Hello everyone, thank you for visiting my blog again.
Hope you all doing great.
Here in this blog, we will try to understand about VMware cloud foundation by looking at simple Q&A.
I recently got a chance to attend VCF bootcamp based on VCF 4.3 version, although it is not the latest VCF version, but that bootcamp helped me to understand VCF more better. I tried to put my takeaway points from this bootcamp in the form of Q&A.
1. What are two architecture types in VCF?
We have two types: consolidated and standard
Consolidated architecture:
Run management VMs and customer workloads on one domain which is management domain.
The resources on this cluster will be shared across these different workloads by using resource pools.
Standard architecture:
Management domain is dedicated to run management appliances
And customer workloads will be running on VI workload domains.
2. What are the deployment options available in VCF?
By default single site deployment:
with in this site, if we need high availability we can stretch the cluster, then it is called stretched deployment.
Note:
management domain need to be stretched first before stretching VI domains
workload domain with vLCM base line option only can be stretched.
stretched cluster forms two availability zones (AZ1 AZ2) are with in the same region, it is only high availability solution but not considered as Disaster recovery solution.
if customer have some edge locations or remote office centers but want to connect to the centrally managed vcenters, we can go with remote cluster deployment.
If customer want to implement DR solution to withstand region failures, we need to go with VCF multi-region deployment using VCF federation/NSX-T Federation.
3. How many vCenters in VCF ?
In VCF environment, every workload domain will have one dedicated vCenter.
All the workload domain vCenters are associated with each other in enhanced linked mode in ring topology.
ELM supports maximum 15 vCenters so maximum workload domains that can be created in VCF is 15. One management domain and 14 VI domains.
To understand more about SSO ring topology how to see the participating nodes, if the ring formed correctly or not, we can use vCenter command line tool called cmsso-util, Please check this blog to understand more about this sso ring topology.
4. How the Host overlay IP addresses are assigned in VCF?
We have two supported methods
- DHCP
- NSX-T static IP pools
DHCP:
DHCP makes it simpler to assign IP addresses to TEP interfaces automatically, if your environment is dynamically evolving, new nodes are being added regularly then maintaining a DHCP server would be easier approach instead of creating static IP pools. But it creates external dependency on DHCP server.
Static IP pools:
These will be created in NSX-T manager. It eliminates the need of external DHCP server.
It is a new option included with VCF 4.2 onwards.
Every time you are adding new nodes or creating a wld, you specify a network pool and specify the range of IP addresses etc. This is a good option when your environment is small, and not making regular node adds/new wld.
Note: if you use Static IP pool, you cannot stretch the cluster. Stretched cluster is not possible with Host Overlay Static IP addressing.
5. How Management, vSAN, vMotion IP addresses are assigned?
VCF admin/implementation engineer will manually assign Management IP address statically and hostname before starting bringup or before adding the hosts to VCF inventory
VCF can support 3 VDS maximum.
In recent VCF deployment parameter workbooks, we see two VDS profiles.
VDS profile-1:
- By default single VDS with two uplinks (vmnic0, vmnic1)
- Single VDS manages all types of traffic Management, vSAN, vMotion, Host Overlay
- No traffic separation
VDS profile-2:
- 2 VDS with 2 uplinks each
- VDS-1 for Management, vMotion, host Overlay or VDS-1 for Management, vMotion, vSAN
- VDS-2 for VSAN (for example) or VDS-2 for host Overlay
- Management domain supports only one principal storage option that is vSAN.
- VI domains supports non VSAN storage options as their principal storage option.
- NFS, vVOLs, VMFS on FC
- Customers can connect their external storage to their WLD clusters as supplemental storage option.
- Both Management domain and VI domains supports FC, NFS, vVols (supports multiple protocols FC,NFS, iSCSI) as secondary storage options.
- vLCM baselines means the legacy vsphere update manager - VUM. It is by default.
- When we configure wld we get option to choose vLCM baselines or vLCM images.
- Management domain always VUM, and workload domains supports both options baselines and images.
- vLCM images contains the hardware FW, Driver, BIOS etc. (Base image-ESXi, Vendor add on - drivers, Firmware/Driver add on like BIOS, HBA firmware etc)
- We cannot change this option from baseline to images or vice versa once the workload domains created and all the clusters in that workload domain use the same upgrade method.
VCF supports both online method and offline method.
If SDDC Manager have access to VMware depot (internet connectivity) then online method will be easiest. If no internet connectivity (dark site), then customers can use offline method.
I will write another blog post to explain online and offline LCM options.
Generally it is the below sequence but it depends on VCF versions, SDDC Manager will take care of this component update order and dependencies.
Management domain:
- VCF software - sddc manager, sddc manager services etc
- vRealize components
- NSX-T (UC, Edge node, host transport nodes, nsx-t managers)
- vCenter
- ESXi
VI domain:
- NSX-T (UC, Edge node, host transport nodes, nsx-t managers)
- vCenter
- ESXi
There are two types of bundles: upgrade bundles, install bundles
For example you have VCF with a management domain, VI domain running at 4.3 and 4.4 released.
- Then there will be two types of bundles one is upgrade bundle to upgrade the existing workload domains.
- And install bundles that is to deploy new VI domain, if you want to create a new domain it use the install bundle.