Academy/vSphere Foundation 9.0 Support (2V0-18.25)/Lab: ESXi Network Diagnostics & Path Failover
This lab targets VCF 9.0

Lab: ESXi Network Diagnostics & Path Failover

VCF 9.0Intermediatevcp-foundation⏱ 105 min

VVF 9.0 support lab (2V0-18.25). Built for a Holodeck VVF instance.

Objectives

  • List iSCSI paths and the path selection policy with esxcli storage core path / nmp device
  • Disable one path, prove failover to the remaining path and restore it
  • Test VMkernel connectivity with vmkping against interface names, including a don't-fragment jumbo test

Prerequisites

Holodeck VVF 9.0 instance; SSH enabled on one ESX host. Prerequisite you build: New-HoloDeckInstance -VVF builds vSAN storage — its only storage parameter is -vSANMode ESA/OSA, default OSA (Holodeck Command Reference, vmware.github.io/Holodeck) — so this lab needs an iSCSI target you provide: a Linux VM running an iSCSI target service on the lab network, presenting at least one LUN that the nested hosts reach on two VMkernel paths (software iSCSI adapter with two bound VMkernel ports). Not yet tested in Holodeck — record your build steps (VM sizing, network placement, target service and LUN configuration, host-side adapter and port binding) in the lab log.

Lab Environment

Holodeck vSphere Foundation 9.0 instance (New-HoloDeckInstance -VVF; -vSANMode OSA by default, ESA where the physical layer supports it): vCenter managing one vSAN cluster of nested ESX hosts, plus VCF Operations. VVF has no SDDC Manager and no NSX. For this lab, add the learner-built Linux iSCSI target VM on the lab network (see prerequisites; not yet tested in Holodeck).

Credentials

SystemUsernamePassword
vCenteradministrator@vsphere.localHolodeck documented default (Post Deployment table)
VCF OperationsadminHolodeck documented default (Post Deployment table)
Nested ESX hostsrootHolodeck documented default (Post Deployment table)

Tasks

Task 1 iSCSI path failover and VMkernel connectivity

ESXi network diagnostics require understanding the packet path from VM through vNIC → port group → VDS → vmnic → physical switch. Each hop can introduce failures. Systematic diagnosis follows the packet path.
Step 1

SSH to an ESX host with iSCSI storage attached.

Step 2

List the paths of one iSCSI device: esxcli storage core path list -d <naa.id>. Note the runtime names (vmhbaN:C#:T#:L#) and states; esxcli storage nmp device list -d <naa.id> shows the Working Paths and the path selection policy.

Step 3

Disable one path (simulate a cable fault): esxcli storage core path set --state off -p <vmhbaN:C0:T0:L0>.

Step 4

Verify failover: esxcli storage core path list -d <naa.id> shows the disabled path as off; esxcli storage nmp device list -d <naa.id> shows Working Paths reduced to the remaining path(s). A disabled path shows off, not dead — dead is what a real link loss produces.

Step 5

Confirm the LUN is still accessible through the alternate path: esxcli iscsi session list (sessions to the target remain) and esxcli storage core path list (the active path carries I/O). iscsiadm is a Linux tool and does not exist on ESX.

Step 6

Re-enable the path: esxcli storage core path set --state active -p <vmhbaN:C0:T0:L0>, then confirm all paths are active again.

Step 7

Network checks with the VMkernel interface name, not a port-group name: vmkping -I vmk0 -c 4 <gateway> for the management path; for a jumbo-frame path test vmkping -I <vmkN> -d -s 8972 <peer> (VCF 9.0.1 docs p.1642).

Validation Gate

Check: Demonstrate iSCSI path failover and recovery with esxcli (core path list/set, nmp device list, iscsi session list) and verify VMkernel connectivity with vmkping against interface names.

Expected: The disabled path shows off, Working Paths drop to the remaining path(s) with I/O continuing, the path returns to active after re-enabling, and vmkping succeeds (including the -d -s 8972 jumbo test where configured). VM-connectivity diagnosis (esxcli network nic stats get, pktcap-uw capture, VDS port-group VLAN checks) is a follow-on exercise, not part of this gate.

Common Errors

Using ping as the only diagnostic tool for network issues
Fix: Ping tests ICMP connectivity but doesn't diagnose: VLAN issues (use 'esxcli network vswitch dvs vmware list' to verify VLAN), MTU issues (use 'vmkping -d -s 8972' for jumbo frame test), NIC teaming issues (use 'esxcli network nic stats get -n vmnicX' for per-NIC traffic counters), or DNS issues (use 'nslookup' from ESXi shell).
Not checking physical NIC link status and speed negotiation
Fix: A NIC showing 'Up' in vCenter may be negotiated at 1Gbps instead of 25Gbps due to cable or switch port issue. Check: 'esxcli network nic get -n vmnic0' shows link status, speed, duplex. Mismatched speed causes performance issues that look like network congestion.
Forgetting to test path failover after NIC replacement
Fix: After replacing a failed NIC or cable, verify failover works correctly: (a) confirm new NIC is in the correct teaming group, (b) test failover by disconnecting the primary NIC and verifying traffic switches to standby, (c) restore primary and verify failback. Untested failover is not failover — it's hope.
Not verifying VLAN tagging consistency between VDS and physical switch
Fix: A VM on VLAN 100 port group with a physical switch port trunking only VLAN 200 results in silent packet drops — no errors, just no connectivity. Use packet capture (pktcap-uw) on the ESXi host to verify VLAN tags on packets leaving the vmnic match physical switch expectations.

Final Validation

Gate: every check below passes with the evidence recorded in the lab log.

✓ Disabled path state → esxcli storage core path list shows it off; Working Paths drop to the remaining path

✓ I/O continuity → esxcli iscsi session list still shows sessions; the LUN stays accessible

✓ Path restored → All paths active again after --state active

✓ vmkping -I <vmk> tests → Succeed, including -d -s 8972 where jumbo frames are configured

Cleanup / Restore

• Re-enable any path left off

• Remove the lab iSCSI target if it is no longer needed

Design Reflection (VCDX)

Network troubleshooting methodology demonstrates operational maturity. VCDX designs should include diagnostic procedures for common network failure scenarios.

Requirements

  • Diagnose network connectivity issues systematically
  • Verify VLAN, MTU, NIC teaming, and physical connectivity
  • Test failover behavior after changes

Constraints

  • pktcap-uw captures can impact host performance at high rates
  • Some diagnostics require ESXi SSH access (may be disabled in lockdown mode)

Assumptions

  • SSH access is enabled for troubleshooting
  • Physical switch configuration is accessible for cross-reference

Risks

  • Silent packet drops from VLAN mismatch — no error messages
  • Performance degradation from NIC speed mismatch going unnoticed

⚠ Known Pitfalls

Assuming ping failure = network down. Ping tests ICMP only — the issue might be VLAN, MTU, or application-layer.
Not verifying NIC speed negotiation after cable replacement — a 25GbE NIC negotiating at 1Gbps causes subtle performance issues.

References

Was this page useful?
Type to search. ↑ ↓ to move, Enter to open, Esc to close.