Networks and clusters

Use the Networks tab to inspect port groups, switch wiring, and VLAN coverage. Host and cluster inspectors show how those resources fit together.

Networks and switches

The Networks tab (6) groups port groups by distributed switch, standard switch name across hosts, and opaque (NSX) networks. Distributed uplink port groups are omitted.

Column	Meaning
VLAN	VLAN, trunk range, or PVLAN; none for untagged standard port groups
VMS	Attached VMs; - until wiring loads
NOTES	Switch/port-group findings: promiscuous, standby on 1 host, on 1 of 3 hosts, unused
VMK	VMkernel adapters
HOSTS	Hosts on switch; joined/total when cluster hosts are missing

space folds a switch; t toggles the tree and plain list. Filtering keeps matching port groups under their switch and unfolds matching switches.

Enter on a port group shows teaming, active/standby uplinks, security policy, and VMkernel adapters. On a switch, it opens two pages:

Key	Page	Controls and content
0	Overview	Properties, uplink-by-host grid, findings, port groups. Identical host uplinks share a row. ▲ marks speeds differing from the majority; down means no link.
1	Wiring	Physical NICs → uplinks → switch → port groups, with NIC counts across hosts. j/k select a port group; heavy links show active uplinks, dotted links standby. Enter opens it.

Narrow terminals reduce the wiring diagram's labels and columns.

Findings name the host and affected component:

Finding	Consequence
Cluster host missing from switch	VMs using the switch cannot move there
Uplink missing a physical NIC	One fewer traffic path
NIC has no link	Traffic has failed over or has no path
Uplink speeds differ across hosts	Throughput depends on placement
Port group has no working uplink or only standby	Traffic is down or has no failover left
VMkernel MTU exceeds switch MTU	Jumbo-frame traffic such as vMotion/vSAN fails
Promiscuous mode allowed	VMs can see other VMs' traffic
Standard port-group VLAN, switch MTU, or port-group presence differs across hosts	Networking or vMotion depends on placement

Port groups load with inventory. Wiring loads separately while Networks or a switch workspace is open, and after network inventory reloads. It adds hosts, uplinks, NICs, VMkernel adapters, and VM counts. reading wiring… marks a pending read; failures show a reason. These reads need no privileges beyond ordinary inventory.

VLAN map

Press v on Networks to map VLANs across the vCenters in scope; a includes all loaded vCenters. Each vCenter is a column. Untagged networks come first, trunks last.

Cell	Meaning
● 288	Present, with 288 VMs
vmk 56	Present, with 56 VMkernel adapters and no VMs
○	Present, unused
·	Absent
4/6	Four of six clusters can reach it

Standard and distributed port groups sharing a VLAN share a row. Notes flag VLANs found on one vCenter or with no shared name across vCenters (names differ). Enter lists clusters, attached VMs, and trunks carrying the VLAN.

Press p to pick source and target clusters for a live network comparison. Matching uses VLAN first, preferring the same name, then name. The picker suggests a same-named cluster at another site as the target.

The target column marks missing networks ✕: a blocker when VMs use them, an advisory when unused. Matched networks show the target name and differences in MTU, teaming, or security. Enter lists source VMs on that network; x clears the pair.

Host network page

A host inspector offers 0 Summary and 1 Network when topology is available. Network connects each physical NIC to its uplink, switch, port groups, local VMs, and VMkernel adapters. Missing cluster switches get their own line; unclaimed NICs appear last. j/k select a switch, Enter opens it, and Esc returns to the host.

The page uses the same on-demand topology as switch views. It flags missing cluster switches, unassigned uplinks, down NICs, MTU mismatches, and unequal speeds. A linked NIC claimed by no switch means the cable is connected, but the host has not joined a switch using it.

Cluster workspace

Enter on a cluster starts on Summary. Choose pages with 0–2; ←/→ changes cluster while keeping the page.

Key	Page	Shows
0	Summary	Live health/configuration issues; HA, monitoring, failover capacity/admission control; effective hosts and EVC; capacity, DRS, contents
1	Hosts & VMs	Host state, CPU/memory use, VM count. space expands a host's VMs; Enter opens its Hosts-tab row.
2	Storage	Datastore mounts, free space, and hosts missing shared mounts

Summary marks fields ✓, ▲, or ✕ for conditions such as insufficient HA failover capacity, disabled admission control, unavailable hosts, manual DRS, or missing mounts. Capacity shows used/total. Effective capacity subtracts unavailable hosts and hypervisor overhead; it has no usage percentage because host usage includes the hypervisor. Standalone hosts omit Resilience, EVC, and DRS. A cluster whose failover capacity is below its admission-control reserve shows the warning glyph ◐ in the Clusters list even when vCenter shows it green, because vCenter refuses new power-ons there; vsfleet analyze health reports it as ha-failover-capacity-low.

Storage puts incomplete shared mounts first and names missing hosts: VMs cannot run or restart there. A datastore mounted by exactly one host is treated as local and is not flagged. The datastore inspector's Hosts field also lists mount coverage by cluster. For files, use the datastore browser; for network coverage, use the switch workspace.

Cluster pages use already-loaded inventory. Older captures and imported RVTools workbooks lack health, HA, DRS, EVC, and mount evidence; fields show - and Storage reports unread mounts.