VxRail 7.0.XXX Planning
VxRail 7.0.XXX Planning
VxRail 7.0.XXX Planning
241
PLANNING
DOWNLOADABLE CONTENT
DOWNLOADABLE CONTENT
Table of Contents
Glossary..................................................................................................38
ToR Switches
Management Switch
VxRail Cluster
VxRail Manager
vCenter Server
VDS
vSAN
ESXi
VxRail Nodes
The deployment process includes planning, predeployment steps, cluster build, and
postdeployment steps.
Movie:
https://players.brightcove.net/6057277730001/default_default/index.html?videoId=1
710762632297803170
For more information, read the Dell VxRail 7.0.x Deployment Guide.
vCenter Server:
• A centralized platform for managing VMware server virtualization and vSAN
• The primary point of management for VxRail clusters using the VxRail Plugin
VxRail clusters are deployed with one of two vCenter options; Use the VxRail
vCenter Server or Join an existing vCenter Server.
For detailed information, see the VxRail vCenter Server Planning Guide.
ESXi ESXi
ESXi ESXi
Data Center
Planning Considerations:
• Includes a nontransferable standard vCenter license
• Uses VxRail Lifecycle Management for vCenter software upgrades
• Does not support vCenter High Availability
• Deployed with a noncustomizable vSphere Single Sign-on (SSO) domain called
vsphere.local
• Supports internal or external DNS server options
• Supports VxRail-deployed Virtual Distributed Switches
• Supports Enhanced Link Mode with other VxRail-managed vCenter Servers
Prerequisites:
• If internal DNS is selected, an IP address and DNS name must be available
during deployment. These values are added to the DNS service running on the
VxRail Manager VM.
• If external DNS is selected, forward and reverse DNS entries must be
preconfigured on a DNS server in the customer environment.
The Join an existing vCenter Server option can be further divided into two
options, each with its own planning considerations.
To learn more about the deployment options, select each tab below.
Data Center
Planning Considerations:
• Includes a nontransferable standard vCenter license
• Uses VxRail Lifecycle Management for software upgrades
• Does not support vCenter High Availability
• Uses existing noncustomizable vSphere SSO domain called vsphere.local
• Requires an external DNS server
• Supports a new VxRail-deployed or existing Virtual Distributed Switches
Prerequisites:
• Preconfigured forward and reverse DNS entries
• Compatible vCenter and VxRail versions
• Existing SSO domain
• Preconfigured VxRail management and administrator user accounts in vCenter
• Preconfigured data center in vCenter
A new VxRail cluster is deployed with an existing vCenter server that runs on
infrastructure that is external to a VxRail cluster. This option is used for single or
multiple cluster deployments.
ESXi ESXi
ESXi ESXi
Data Center
Planning Considerations:
• Requires a separate vCenter license
• Uses customer-defined procedures for software upgrades
• Supports vCenter HA
• Supports a custom vSphere SSO domain
• Requires an external DNS server
• Supports a new VxRail-deployed or existing Virtual Distributed Switches
Prerequisites:
• Preconfigured forward and reverse DNS entries
• Compatible vCenter and VxRail versions
• Existing SSO domain
• Preconfigured VxRail management and administrator user accounts in vCenter
• Preconfigured data center in vCenter
VxRail nodes connect to one or more ToR network switches and are configured as
a VxRail cluster. The physical switch infrastructure supports all the network traffic
for the VxRail cluster. VxRail logical networking depends on the ToR switches for
physical connectivity between the nodes, and to the core production network.
VxRail network traffic is configured for Layer 2 networking. For supported VxRail
network speeds and interfaces, see the Dell VxRail Spec Sheet. Each node has an
Integrated Dell Remote Access Controller (iDRAC) which requires a connection to
a management switch.
ToR Switches
Management Switch
VxRail Cluster
VxRail nodes with four 10 GbE interfaces and two ToR switches.
Planning Considerations:
• VxRail is compatible with most Ethernet switches.
• Layer 2 switches are required. If the customer requires, Layer 3 switches may
be used.
• Two or more ToR switches are required for increased performance and high
availability.
• A minimum of 10 GbE is required for All-flash and NVMe clusters, but 25 GbE
or faster is recommended.
PCIe NIC
Planning Considerations:
• Using network ports across the integrated and PCIe NICs is required for high
availability.
• Network ports that are selected during the VxRail initial build process are
reserved exclusively for VxRail usage.
• Ports on the integrated and PCIe NICs can run at different speeds. For
instance, the integrated NIC ports can support 10 GbE and the ports on the
PCIe NIC support 25 GbE.
• Network adapter ports must match the type and speed of the ToR switches.
During the initial build of a VxRail Cluster, a predefined or custom network profile is
selected. These profiles determine the number of VDSs, number of uplink port
groups, and uplink failover policy for each distributed port group to be configured.
The VxRail configuration portal groups VxRail networking into six types of traffic.
Each traffic type is associated with a distributed port group that exists on one or
two VDSs. VxRail supports IPv6 for discovery traffic only. All other types of traffic
must be IPv4.
ToR Switches
VxRail Node
Uplink 1 Uplink 2
VDS
A VDS predefined network profile contains default configuration settings for a VDS.
The settings reflect VxRail cluster deployment best practices for common
configurations. The profile is used to create a VDS at the time of cluster
deployment.
Planning Considerations:
The uplink port group configuration for the VDS depends on the number and type
of NICs that are used for VxRail traffic. Two uplinks are configured for each
distributed port group, one active and one standby. For the VDS predefined
network profiles, the NICs must be cabled in a specific manner.
To learn more about the cabling and port group configurations, select each tab
below.
ToR Switches
P0 - vmnic0 P1 - vmnic1
Integrated NIC
Uplink 1 Uplink 2
VDS
Failover Pair
Teaming Policy
ToR Switches
Integrated NIC
Teaming Policy
ToR Switches
Integrated
PCIe NIC
NIC
Failover Pairs
Teaming Policy
The cabling for this configuration is different than the 4x10GbE configuration. The
integrated NIC Port 0 is cabled to Switch 1 and PCIe NIC Port 0 is cabled to Switch
2. These two ports form a failover pair. The integrated NIC Port 1 is cabled to
Switch 2, while PCIe NIC Port 1 is cabled to Switch 1. These two ports form an HA
failover pair as well.
Planning Considerations:
• With a custom uplink assignment, resource-intense networks like vSAN can be
assigned to higher speed uplinks.
• If the customer requires enhanced security in vCenter, management and
nonmanagement port groups can be split across two VDSs.
• For resource-intense networks like vSAN, a higher MTU (jumbo frames) can be
set to improve performance.
• Unlike the predefined network profile, the custom network profile supports
active/active teaming policies for distributed port groups.
• Link aggregation can be configured on the new VDS after the VxRail cluster
build is complete.
• VMkernel MTU must be less than or equal to the VDS MTU.
The custom network profile with an existing VDS requires that a VDS be configured
before the VxRail cluster is deployed. This profile is only available if the VxRail
cluster is joined to an existing vCenter Server. This custom profile is the only profile
that supports Link Aggregation Groups (LAG).
Planning Considerations:
• Port groups for each VxRail traffic type must be created on the VDS.
• Names for the VDS, distributed port groups, and uplinks are case-sensitive and
must match the names in vCenter.
• The port group for the Guest VM network is not required for the VxRail initial
build and can be created later.
• LAG for nonmanagement distributed port groups is supported.
• Static LAG and Link Aggregation Control Protocol (LACP) for dynamic LAG are
supported.
DNS Considerations
Domain Name System (DNS) is required to deploy the VxRail cluster and for
ongoing operations. VxRail supports internal or external DNS options. The internal
DNS service runs on the VxRail Manager VM.
Planning Considerations:
• If a VxRail-managed vCenter is used, the internal or external DNS option is
supported.
• If joining an existing vCenter, only an external DNS option is supported.
• VxRail components that are connected to the Management or vCenter Server
port groups must be able to access the DNS server. These components include
VxRail Manager, vCenter Server, Log Insight (optional), and each ESXi host in
the VxRail cluster.
• For the external DNS option, forward and reverse lookup records must be
created for each component before cluster deployment.
• The internal DNS option only supports naming services for the VxRail cluster
components and does not support naming services outside of the cluster.
• vMotion and vSAN IP addresses do not require DNS entries.
• If an internal DNS is deployed, it can be migrated to an external DNS or a
forwarder to an upstream DNS may be added later.
multicast network protocols to discover VxRail nodes. The IPv6 multicast traffic is
sent over internal management network that is used for discovery only. The default
VLAN ID assigned to the internal management network is 3939 and can be
changed.
Each VxRail node runs the VMware Loudmouth service and an election service to
support build process automation. VxRail Manager also runs the VMware
Loudmouth service.
Planning Considerations:
• The internal management VLAN (discovery) must be configured on the ToR
switches for node ports and Inter-Switch Links (ISLs).
• ToR switches must be configured to allow IPv6 multicast network protocols.
• IPv6 multicast network protocols must be allowed in the customer environment.
For a VxRail deployment, the IP address discovery option is a manual method for
detecting cluster nodes. Each VxRail node management IP address is manually set
in the ESXi shell on the node. On the primary node, the VxRail Manager VM is
manually configured with an IP address. The external management network is used
for IP address discovery.
The IP address discovery option is used when the customer has concerns about
using either IPv6 or multicast network protocols. Since nodes are available over the
external management network, the default ESXi root password can be changed to
address security concerns.
Planning Considerations:
• IPv6 multicast network protocols are not required.
1. Why would the "Join an existing vCenter Server" option be selected for
vCenter deployment? Select three (3).
a. To deploy a second VxRail cluster and minimize the number of vCenter
servers.
b. To deploy a VxRail cluster in a customer's VMware environment.
c. To deploy a VxRail cluster and use a preconfigured virtual distributed
switch.
d. To deploy a stand-alone single VxRail cluster.
e. To deploy the first of many VxRail clusters with a VxRail-managed vCenter.
Which configuration options must be selected to build this second VxRail cluster
configuration? Select three (3).
a. VDS Predefined Network Profile Configuration
b. VDS Custom Network Profile
c. Use the VxRail vCenter Server
d. Join an Existing vCenter Server
e. Automatic Discovery
f. IP Address Discovery
f. IP Address Discovery
The VxRail Configuration Portal is a central interface to plan and prepare VxRail
cluster deployments. Project and cluster configuration information is entered using
wizards. Checklists, cluster reports, and configuration files are then available once
the configurations are saved.
Movie:
https://players.brightcove.net/6057277730001/default_default/index.html?videoId=1
710762632463449907
This example shows a new project with customer and service provider contacts.
The project has been associated with a cluster.
A VxRail Configuration Portal cluster configuration contains settings that are used
to deploy a VxRail Cluster. Project team members enter configuration information
using a wizard. The information is validated, and a configuration file is created that
can be used for deployment. A cluster configuration is associated with only one
project.
This example shows the cluster configuration categories that are covered in the
wizard.
To learn about Creating a Cluster in the VxRail Configuration Portal, watch the
video below.
Movie:
To learn about how to access the HTML version of the report or export a Microsoft
Word version, select each tab below.
To access the configuration report, select the My Clusters tab, select the cluster in
the list, and click the VIEW REPORT button.
On the HTML version of the report, click the EXPORT CONFIGURATION REPORT
button to export a Microsoft Word document version.
Pre-deployment Checklist
The pre-deployment checklist provides the tasks that must be completed before
beginning a VxRail cluster deployment. The checklist also provides links to online
help which contains additional links to resources used to complete the checklist.
Example: Checklist
The cluster configuration file is a JSON file that contains the complete set of
configuration properties for a cluster. The file can be used to build a cluster.
The cluster configuration file can be downloaded from two other locations in the
portal.
Click the Save Progress and Exit button in the course menu or below to
record this content as complete.
Go to the next learning or assessment, if applicable.
Number of VDS
Distributed port
groups assigned
to each VDS
VDS MTU
Number of NICs
assigned to the
Host NICs
assigned to each
Active/standby
teaming and
load-balancing
VDS MTU
Active/active
teaming and
load-balancing
Number of VDS
Is LAG configured?
To learn more about the election service and how to force a node to become the
primary, review KB 5801.
Bearer Authentication
HTTP authentication method that uses a token instead of a username and
password combination.
Fault Domain
A fault domain consist of one or more vSAN hosts grouped according to their
physical location in the data center. Fault domains enable vSAN to tolerate failures
of entire physical racks and failures of a single host, capacity device, network link,
or a network switch dedicated to a fault domain. When fault domains are not
configured, hosts behave as individual fault domains.
LZ4
LZ4 is a lossless data compression algorithm that is focused on compression and
decompression speed.
NSX-T Federation
RAID 1 Mirroring
Mirroring can accommodate an FTT setting of 1, 2, or 3. If FTM is set to mirroring,
for the N number of failures that are tolerated, N+1 copies of the object are created.
Mirroring requires witness components. The number of witnesses is equal to the
FTT setting. RAID 1 with FTT setting of 1 requires a minimum of 3 hosts. RAID 1
with FTT setting of 2 requires a minimum of 5 hosts and, RAID 1 with FTT setting of
3 requires a minimum of 7 hosts.
RESTful API
A useful analogy for APIs is human language. Language enables humans to
communicate with each other by providing structure for how that communication
takes place. Application Programming Interfaces (APIs) function in a similar
manner.
APIs establish rules that enable a client and server to communicate data that would
otherwise be unmanageable. Representation State Transfer (REST) API is a
specific design methodology for API communications. REST API is an architectural
style for client/server communications because it promotes scalability, simplicity of
interface, and independence of communications.
SmartFabric Service
VxRail systems that are configured with SmartFabric Services automatically detect
VxRail nodes during expansion. VxRail nodes advertise NIC port information,
system name, and description via the LLDP protocol to the SmartFabric. The
SmartFabric uses the internal management network to apply network state
changes to the top-of-rack (ToR) switches. SmartFabric configuration includes
creating and configuring L2 VLANs and bonded port channels.
STP is running by default on all ports of the switch. STP makes each port wait up to
50 s before data can be sent on the port. This delay in turn can cause problems
with some applications or protocols including VxRail nodes.
Cisco devices implemented PortFast to alleviate problems that are related to this
delay. The terminology might differ between different vendor devices. PortFast
causes a port to enter the forwarding state almost immediately by dramatically
decreasing the time of the listening and learning states. PortFast minimizes the
time that it takes for the server to come online, thus preventing problems with
applications such as DHCP and DNS. The use of PortFast should only be
implemented when the port on the switch is directly connected to a server and
never to another switch.
VxRail-managed vCenter
VxRail-managed vCenter
A vCenter that is deployed during an initial VxRail cluster build and is upgraded
using VxRail Manager.