Dell VxRail Deploy flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

VxRail minimum cluster size

A standard VxRail cluster requires a minimum of three nodes so that vSAN can tolerate one failure. Two-node clusters need a witness appliance at a third site.

vSAN RAID-5 node minimum

Erasure-coded RAID-5 in vSAN needs at least four nodes in the cluster. Below four nodes, failure tolerance is provided by RAID-1 mirroring instead.

vSAN RAID-6 node minimum

RAID-6 erasure coding in vSAN requires a minimum of six nodes. It tolerates two failures but consumes more capacity overhead than RAID-5.

Witness appliance purpose

A witness appliance holds metadata only, not data. It provides quorum for two-node and stretched clusters so the cluster survives loss of one site or node.

VxRail Manager role

VxRail Manager is a VM running on the cluster that provides deployment wizards, lifecycle bundle upgrades, health monitoring, and the vCenter plugin for the VxRail system.

NTP requirement

All VxRail components, including ESXi hosts, vCenter, and VxRail Manager, must synchronize to reachable NTP sources. Clock skew breaks certificates and authentication during deployment.

Deployment client prerequisites

The browser host running the VxRail deployment wizard must resolve VxRail Manager and ESXi FQDNs, reach their management addresses, and trust the certificates presented.

External vCenter support

VxRail can deploy into a customer-supplied external vCenter instead of the embedded VxRail vCenter. The external vCenter must be a supported version with adequate permissions.

iDRAC first connection

Connect and configure the iDRAC management port before powering nodes into service. It provides remote console and virtual media for the rest of the build.

Redundant PDU cabling

Each VxRail node has two power supplies. Connect one PSU to the A feed and one to the B feed so a single PDU or circuit failure does not power off the node.

Dedicated boot device

VxRail nodes boot from a separate device such as a BOSS card or M.2 RAID-1 mirror, keeping vSAN cache and capacity drives dedicated to the datastore.

vSAN disk group cache

Each vSAN disk group has one cache-tier drive and one or more capacity drives. All-flash configurations use the cache drive as a write buffer.

Firmware bundle discipline

VxRail validates a specific firmware and software bundle for each model. Upgrade components only through the VxRail bundle, not individually, to preserve support.

Pre-deployment hardware check

Before starting the VxRail wizard, verify in iDRAC that power supplies, fans, memory, and all drives report healthy. A failed component halts or degrades the automated build.

Uplink redundancy

VxRail nodes connect to a pair of top-of-rack switches using supported optics or DAC cables. Uplinks are teamed with LACP or active/standby for path redundancy.

Rack airflow direction

Install VxRail nodes so air flows front to back, with the cold aisle at the front. Maintain clearance above and below the chassis to avoid hot spots.

VxRail deployment wizard

The initial VxRail Manager wizard collects vCenter credentials, network settings, DNS, NTP, and the software bundle, then automates the ESXi and vSAN cluster build.

VDS default networking

VxRail 7.x deployments use the vSphere Distributed Switch by default, with physical NICs mapped to uplinks and port groups carrying VLAN IDs per traffic type.

Jumbo frame consistency

vSAN and vMotion port groups commonly use MTU 9000. Physical switches, the distributed switch, and every VMkernel on the path must match or large frames are dropped.

VMkernel IP planning

Each ESXi node needs a separate VMkernel address for vSAN, vMotion, and management traffic. Reserve contiguous IP ranges before running the deployment wizard.

Witness traffic separation

Two-node and stretched clusters require a dedicated witness VMkernel and VLAN so witness traffic is isolated from vSAN and management traffic across sites.

VxRail Manager bootstrap address

VxRail Manager initially answers on a temporary bootstrap address. The wizard sets the permanent IP and FQDN, which must match forward and reverse DNS records.

Post-build validation

After the wizard completes, confirm all nodes joined the cluster, vSAN health passes in vCenter, and the VxRail plugin reports no configuration drift.

External vCenter permissions

Deploying to an external vCenter requires an account with sufficient privileges to create datacenters, clusters, and distributed switches, plus the VxRail-specific roles.

VxRail bundle upgrade order

A VxRail bundle upgrade runs VxRail Manager pre-checks, then upgrades vCenter if embedded, then ESXi nodes one at a time, then VxRail Manager itself.

Pre-upgrade health check

Before upgrading, confirm no degraded vSAN objects, no failed disks, and enough free capacity for rebuilds during rolling node upgrades.

Maintenance mode data migration

Full data migration evacuates all data from a node before service. Ensure accessibility mode only moves what is needed. Full migration requires spare capacity.

VxRail log bundle

The VxRail Manager log bundle collects ESXi, vCenter, and VxRail Manager logs into one archive. It is the standard first artifact for support cases.

vSAN health network check

The vSAN health network check tests connectivity and MTU between hosts. Failures usually indicate a VLAN mismatch or inconsistent jumbo frame configuration.

Configuration drift reconcile

When the VxRail plugin reports drift, use the reconcile or resync operation in VxRail Manager. Manual edits in vCenter can worsen drift and complicate support.

VxRail Manager unreachable

If VxRail Manager is unreachable, confirm the VM is powered on in vCenter, verify its IP and DNS records, and check that the plugin service is running.

Rebuild reserve capacity

Keep enough free vSAN capacity to rebuild a failed node or disk. Monitor capacity alerts and expand the cluster before usable space falls below the rebuild reserve.