Academy/VCP-VCF 9.0 Administrator (2V0-17.25)/Certificate Management Across VCF 9.0 Components with VCF Operations
This lab targets VCF 9.0

Certificate Management Across VCF 9.0 Components with VCF Operations

VCF 9.0Advancedadminarchitectvcdx⏱ 120 min

In VCF 9.0 certificate lifecycle for SDDC Manager, vCenter, NSX Manager and ESX hosts is managed from the VCF Operations console (Fleet Management > Certificates): view certificates and alerts, configure a CA (Microsoft CA; OpenSSL for VCF Instance components), generate CSRs, replace with a configured-CA or imported external certificate, renew, and activate auto-renewal (60 days before expiry). The SDDC Manager certificate UI/API is deprecated in 9.0 (release notes). Default signing CA at deployment is VMCA (Table 721). VCF 5.2.x used SDDC Manager for the same functions — see Extension 2. VCF 9.0.1 docs pp.2918-2927 (Managing Certificates in VMware Cloud Foundation), p.1129 Table 721, p.1271 (manual ESX certificate), p.98 (release notes, certificate management).

Objectives

  • Map the VCF 9.0 certificate architecture (default VMCA signing, VCF Operations fleet management locker, enterprise CA) and the trust boundaries across SDDC Manager, vCenter, NSX Manager and ESX hosts
  • Audit certificate inventory, expiry timelines and issuer trust relationships for VCF Management and VCF Instance components in VCF Operations
  • Configure a Certificate Authority in VCF Operations, generate CSRs and replace certificates for SDDC Manager, vCenter and NSX Manager
  • Replace ESX host and NSX Manager certificates from VCF Operations (Fleet Management > Certificates, Show ESX Hosts), and identify what stays outside VCF Operations (NSX Edge nodes, NSX Global Manager, external integrations) with its own manual procedure
  • Activate certificate auto-renewal and use certificate alerts and notifications so that no VCF certificate expires undetected

Prerequisites

VCF 9.0 management domain and at least 1 workload domain fully deployed and healthy. VCF Operations (with fleet management), SDDC Manager, vCenter, NSX Manager and ESX hosts online. At least one domain with 4-6 ESX hosts and an NSX Manager cluster (NSX Edge nodes optional). A test PKI capable of issuing certificates: a Microsoft CA with the Certificate Authority and Web Enrollment roles on the same server and a VMware template (preferred — supported for VCF Management and VCF Instance components), or the OpenSSL CA on the SDDC Manager appliance (VCF Instance components only). Time synchronised between the CA and SDDC Manager.

Prior labs: holodeck-02

Required skills:

  • Certificate fundamentals (X.509 chain validation, trust roots, CSR generation)
  • VCF Operations console navigation (Fleet Management)
  • vCenter and NSX certificate management screens
  • ESX host certificate files (/etc/vmware/ssl/rui.crt and rui.key) and management agent restarts
  • OpenSSL or equivalent CLI tools (openssl x509 -text, openssl s_client, openssl verify)
  • Basic REST API calls (curl or Postman) for programmatic checks

Lab Environment

VCF 9.0 management domain (SDDC Manager, vCenter, NSX Manager cluster, VCF Operations with fleet management) + 1 workload domain (4-6 ESX hosts, NSX Manager cluster, optional Edge cluster). VCF Operations manages certificate lifecycle for both VCF Management components and VCF Instance components. A Microsoft CA (or the OpenSSL CA on SDDC Manager) issues certificates. Optional: external integration endpoint (e.g., Kubernetes API server, syslog collector) that trusts VCF certificates.

Mermaid source
graph TB
  CA["Microsoft CA (certsrv, VMware template)<br/>or OpenSSL CA on SDDC Manager"]
  Ops["VCF Operations console<br/>Fleet Management > Certificates"]
  Ops -->|Configure CA / Generate CSR / Replace / Renew / Auto-renew| CA
  Ops --> SDDC["SDDC Manager<br/>SDDC M SSL"]
  Ops --> VC["vCenter<br/>machine SSL"]
  Ops --> NSX["NSX Manager<br/>LM and VIP"]
  Ops -->|Show ESX Hosts| ESX["ESX hosts<br/>ESX SSL"]
  Ops --> Mgmt["VCF Management components<br/>Fleet Management, VCF Operations, VIDB, Automation"]
  Manual["Manual lifecycle"] --> GM["NSX Global Manager"]
  Manual --> Ext["External integrations<br/>K8s API, syslog, LB"]
  Edge["NSX Edge nodes<br/>Table 721: Not applicable"]
  Alerts["Certificate Alerts +<br/>notification rules"] -.-> Ops

Credentials

SystemUsernamePassword
SDDC Manageradministrator@vsphere.localFrom VCF deployment
vCenter Serveradministrator@vsphere.localSame as SDDC Manager
NSX ManageradminFrom VCF bring-up spec
VCF OperationsadminDeployed with the VCF instance (fleet management)
Microsoft CA service accountsvc-vcf-ca (least-privileged AD account: Request Certificates, Issue and Manage Certificates; Read + Enroll on the VMware template)Provide your own — set when creating the service account

Tasks

Task 1 Map and audit the VCF 9.0 certificate architecture and trust chain

Security

Before you rotate a single certificate, understand the architecture. In production you have 50+ certificates across 20+ components. Which are still on the default VMCA? Which come from the VCF Operations fleet management locker? Which point to your enterprise CA? Which trust boundaries break if you install the wrong chain? VCF 9.0 gives you one inventory for all of them: Fleet Management > Certificates in the VCF Operations console.

Step 1

Log in to the VCF Operations console (https://<vcf_operations_fqdn>) as a user with the Administrator role and click Fleet Management > Certificates. Click VCF Management to see the certificates for Fleet Management, VCF Operations, VCF Identity Broker, VCF Automation and the logs/networks components. Then click VCF Instances and your instance name to see SDDC Manager, vCenter and NSX Manager; switch on the Show ESX Hosts toggle to add the ESX hosts. Build a table: Component | Issuer | Subject | Expiry Date | Days Until Expiry.

Certificate inventory table covering VCF Management components and the VCF Instance components (SDDC Manager, vCenter, NSX Manager, ESX hosts) with issuer and expiry
Use the filters (type, status, component). Table 721 (p.1129) tells you the default signing CA you should expect: management-domain VMCA for SDDC Manager and NSX Local Manager, local workload-domain VMCA for vCenter and ESX, the fleet management locker for VCF Operations, Automation, logs and networks.
Step 2

Open the SDDC Manager certificate details from the inventory and note the chain (issuer, issuer's issuer, root). Download or copy the certificate and inspect it with OpenSSL: openssl x509 -in <cert-file> -text -noout | grep -A 5 'Issuer:\|Subject:'. Alternatively read it live: openssl s_client -connect sddc-manager.vcf.sddc.lab:443 -showcerts </dev/null.

Certificate chain documented: leaf (SDDC Manager) → management-domain VMCA (or your CA if already replaced)
Step 3

Now map trust boundaries. In vCenter (via vSphere Client), navigate to Administration > Certificates > Trusted Root CA Certificates. Verify: Does vCenter trust the CA that issued SDDC Manager's certificate? List all trusted root CAs in vCenter. This is a trust boundary: if SDDC Manager is signed by a CA that vCenter doesn't trust, mTLS between them will fail.

List of trusted root CAs in vCenter (expect the VMCA root(s) for the management and workload domains and any enterprise CA already imported)
Step 4

In NSX Manager (https://nsx-manager.vcf.sddc.lab), navigate to System > Certificates > General. View NSX Manager's certificate issuer and compare to vCenter's trusted roots. Document: Does vCenter trust NSX Manager's CA? Does NSX Manager trust the CA that issued vCenter's certificate (for management plane callbacks)?

Trust confirmation: vCenter -> NSX mutual trust documented
Step 5

For ESX hosts, confirm the inventory from the host side on one host (e.g., esxi-1.vcf.sddc.lab): SSH to it and run openssl x509 -in /etc/vmware/ssl/rui.crt -text -noout | grep -A 2 'Issuer:\|Subject:\|Not After'. Note the issuer (local workload-domain VMCA by default) and expiry. Compare with the VCF Operations inventory entry for that host (Show ESX Hosts).

ESX certificate issuer and expiry documented and matching the VCF Operations inventory; vCenter trusts the issuing VMCA
Step 6

Create a certificate architecture map showing: (a) leaf certificates (SDDC Manager, vCenter, NSX Manager, ESX hosts, VCF Management components), (b) their issuers (VMCA, fleet management locker, enterprise CA), (c) which are lifecycle-managed by VCF Operations and which are manual (NSX Global Manager; NSX Edges not applicable; external integrations) per Table 721, (d) trust paths. Highlight gaps (for example an external integration that pins the current VMCA root).

Comprehensive certificate architecture map with trust paths and any identified gaps

Validation Gate

Check: (1) Certificate inventory table from VCF Operations: VCF Management components and all VCF Instance components (SDDC Manager, vCenter, NSX Manager, ESX hosts) with issuer, subject, expiry, days-to-expiry, (2) SDDC Manager chain verified with OpenSSL, (3) vCenter trusted roots listed, (4) ESX certificate verified on one host and matched to the inventory, (5) architecture map marks what VCF Operations manages and what stays manual (Table 721)

Expected: Complete certificate architecture map ready for rotation planning. You can now answer: 'If I rotate SDDC Manager to a new CA, what other components must I update to maintain trust?'

Common Errors

Fleet Management > Certificates shows no VCF Instance, or ESX hosts are missing from the list
Cause: The user lacks the Administrator role in VCF Operations, the VCF Instance is not yet connected to fleet management, or the Show ESX Hosts toggle is off
Fix: Log in as a VCF Operations Administrator. Confirm the instance appears under Fleet Management > Lifecycle > VCF Instances. In the Certificates view switch on Show ESX Hosts.
OpenSSL command shows 'Unable to load certificate' when inspecting downloaded cert
Cause: Certificate was downloaded in DER format, but OpenSSL expects PEM (text-based). Or file extension mismatch.
Fix: Specify the format: openssl x509 -inform DER -in <cert-file> -text -noout. Or convert: openssl x509 -inform DER -in <cert.der> -out <cert.pem>
NSX Manager certificate issuer shows 'CN=NSX,O=VMware' but vCenter's trusted roots don't include this issuer
Cause: NSX Manager is using a self-signed certificate (issuer = subject, or issued by NSX's built-in CA). vCenter may not yet trust it, or needs NSX CA imported.
Fix: Export the NSX Manager CA/root certificate (System > Certificates) and import it into vCenter's trusted roots, or — the cleaner path in this lab — replace the NSX Manager certificate from VCF Operations with one issued by your configured CA so that vCenter, SDDC Manager and NSX share one trust root.

Task 2 Configure the Certificate Authority in VCF Operations and generate CSRs

Security

VCF Operations replaces the per-component CSR workflow with one CA configuration per scope: VCF Management components support Microsoft CA only; VCF Instance components support Microsoft CA and OpenSSL. Once a CA is configured, CSR generation and replacement are menu actions on the inventory. In production this still needs the PKI team (template, least-privileged service account) and change control, but the operational steps collapse to a few clicks.

Step 1

Prepare the Microsoft CA (VCF 9.0.1 docs pp.2921-2924): install the Certificate Authority and Certificate Authority Web Enrollment roles on the same server; duplicate the Web Server template as 'VMware' (Compatibility: Windows Server 2008 R2 / Windows 7; Extensions: remove Server Authentication from Application Policies, enable Basic Constraints, set Key Usage 'Signature is proof of origin (nonrepudiation)'; Subject Name: Supply in the request); publish it with Certificate Template to Issue. Create the AD service account svc-vcf-ca and grant it Request Certificates and Issue and Manage Certificates on the CA, and Read + Enroll on the VMware template.

Microsoft CA ready: Web Enrollment reachable at https://<ca-fqdn>/certsrv, VMware template published, svc-vcf-ca with least privilege
If you have no Microsoft CA, use OpenSSL for the VCF Instance components (Table 930: CN = SDDC Manager FQDN, country, locality, organisation, OU, state). VCF Management components need Microsoft CA.
Step 2

In the VCF Operations console click Fleet Management > Certificates, click VCF Instances and your instance name, then click Configure CA. For Microsoft CA enter the CA Server URL (https://<ca-fqdn>/certsrv), the svc-vcf-ca user name and password, and the Template Name (VMware). Click Save. Repeat under VCF Management if you also want to replace the management components' certificates with Microsoft CA certificates.

CA configured for the VCF Instance (and optionally for VCF Management)
The CA is configured separately for VCF Management and for each VCF Instance (p.2920). Verify time sync between the CA and SDDC Manager first.
Step 3

Generate the SDDC Manager CSR: in the VCF Instance certificate list select SDDC Manager, open the three-dot menu and click Generate CSR. Enter Common Name (sddc-manager.vcf.sddc.lab), Organization, Organizational Unit, Country, State/Province, Locality, optional Email, Host, Subject Alternative Name(s) and Key Size (2048 or 4096). Click Save and wait until the CSR is generated.

SDDC Manager row shows the CSR generated
Step 4

Repeat Generate CSR for vCenter and for NSX Manager (the NSX entry covers the Local Manager and VIP certificate). Use each component's FQDN as Common Name and include the VIP and node FQDNs for NSX in the Subject Alternative Name field.

CSRs generated for SDDC Manager, vCenter and NSX Manager
Step 5

External-CA path (if your organisation signs CSRs outside VCF): for each component click the three-dot menu > Download CSR, submit it to your CA, then click Import Certificate, name it, and either paste the server certificate and the CA chain (intermediate then root, PEM) or browse to a chain file (.crt, .cer, .pem, .p7b, .p7c). Click Validate, then Save. Only PEM-encoded certificates are supported (p.2927).

Imported certificates validated and saved for each component you signed externally
If you configured a CA in step 2 you can skip this step — Replace With Configured CA Certificate signs and installs in one action (Task 3).
Step 6

Verify any externally issued certificate before use: openssl x509 -in <cert.crt> -text -noout | grep -A 2 'Subject:\|Issuer:' and openssl verify -CAfile <chain.pem> <cert.crt>. Confirm the subject matches the CSR and the issuer is your CA.

Subjects match the CSRs; issuer is your CA; chain verifies

Validation Gate

Check: (1) CA configured in VCF Operations for the VCF Instance (Microsoft CA or OpenSSL), (2) CSRs generated for SDDC Manager, vCenter and NSX Manager from Fleet Management > Certificates, (3) externally signed certificates (if used) imported and validated, (4) OpenSSL verification confirms subject and issuer

Expected: VCF Operations holds a configured CA and pending CSRs for the three management components; you are ready to replace.

Common Errors

Configure CA fails to save or later CSR signing fails against the Microsoft CA
Cause: The Web Enrollment role is on a different server from the CA role, the URL does not end in /certsrv, the service account lacks Request/Issue permissions or Enroll on the template, or time is not synchronised
Fix: Place both roles on the same server (p.2921 note), use https://<ca>/certsrv, re-check the svc-vcf-ca permissions on the CA and the VMware template (p.2924), and align NTP between the CA and SDDC Manager.
Generate CSR for an ESX host fails on key size
Cause: ESX hosts can only use the 2048-bit key size (p.2926)
Fix: Regenerate the ESX CSR with Key Size 2048.
Import Certificate validation fails
Cause: The pasted chain is not PEM, the intermediate/root order is wrong, or the certificate does not match the generated CSR
Fix: Paste PEM blocks only, server certificate first, then intermediate(s), then root (p.2927). Make sure the certificate was issued from the CSR VCF Operations generated for that component.

Task 3 Replace the SDDC Manager, vCenter and NSX Manager certificates from VCF Operations

Availability

VCF Operations is the certificate lifecycle layer for VCF 9.0: it drives the replacement on each component and shows the result in the same inventory. This task replaces the three VCF Instance management components with certificates from your configured CA (or with the imported external certificates) and verifies trust afterwards.

Step 1

In the VCF Operations console click Fleet Management > Certificates > VCF Instances > your instance. Confirm the three CSRs from Task 2 are present on SDDC Manager, vCenter and NSX Manager.

Inventory shows the generated CSRs
Step 2

Replace NSX Manager first: open its three-dot menu and click Replace With Configured CA Certificate, select your CA and click Confirm (configured-CA path), or click Replace With Imported Certificate and select the imported certificate (external-CA path). Follow the task in the Tasks panel.

NSX Manager certificate replaced; inventory shows the new issuer and expiry
Order is a design choice, not a product rule: VCF Operations runs one replacement at a time. Start with the component whose brief service restart affects the fewest dependants.
Step 3

Repeat for vCenter (Replace With Configured CA Certificate or Replace With Imported Certificate). Expect a short vSphere Client interruption while the vCenter machine SSL certificate is applied and services restart.

vCenter certificate replaced; vSphere Client reachable again with the new certificate
Step 4

Repeat for SDDC Manager. Expect a brief SDDC Manager UI interruption.

SDDC Manager certificate replaced
Step 5

Verify from outside: openssl s_client -connect <fqdn>:443 -showcerts </dev/null | openssl x509 -noout -issuer -enddate for each of the three FQDNs. Confirm the issuer is your CA and the expiry is the new one.

All three endpoints present certificates from your CA
Step 6

Verify trust between components: log in to SDDC Manager and the vSphere Client, check that the NSX Manager integration and vCenter inventory show healthy, and review Fleet Management > Certificates > Certificate Alerts for any alert raised during replacement.

No trust errors; no open certificate alerts for the replaced components

Validation Gate

Check: (1) SDDC Manager, vCenter and NSX Manager show the new issuer and expiry in Fleet Management > Certificates, (2) openssl s_client confirms the new certificates on all three FQDNs, (3) no open Certificate Alerts and no trust errors between components

Expected: Three management components replaced from VCF Operations with your CA's certificates.

Common Errors

Replace With Configured CA Certificate is greyed out or fails immediately
Cause: No CA is configured for this scope, or the CSR has not finished generating
Fix: Configure CA for the VCF Instance (Task 2 step 2) and wait for the CSR task to complete before replacing (p.2926).
Replacement task fails on vCenter and the vSphere Client is unreachable for more than 10 minutes
Cause: vCenter services did not come back after the machine SSL change
Fix: Check the task error in VCF Operations, then on the vCenter appliance run service-control --status --all and service-control --start --all; retry the replacement once vCenter is healthy.
After NSX Manager replacement, SDDC Manager shows the NSX Manager as unreachable
Cause: The new NSX certificate was issued by a CA SDDC Manager does not yet trust (external-CA path with an incomplete chain)
Fix: Re-import the certificate with the full chain (intermediate and root) and replace again; with the configured-CA path VCF Operations handles the chain.

Task 4 Replace ESX host certificates from VCF Operations and inventory what stays manual

Manageability

In VCF 9.0 ESX host certificates are lifecycle-managed from VCF Operations like the other components (Fleet Management > Certificates with Show ESX Hosts; Table 721 lists ESX with life cycle 'VCF Operations'). What stays outside: NSX Global Manager (manual), NSX Edge nodes (Table 721: not applicable — Edges use the NSX Manager-issued certificates) and external integration endpoints (Kubernetes API servers, load balancers, syslog collectors) that their owners rotate. The production gotcha is finishing the VCF Operations work and forgetting the external dependants.

Step 1

In Fleet Management > Certificates > VCF Instances > your workload domain, switch on Show ESX Hosts. Record each host's issuer and expiry (default: local workload-domain VMCA).

ESX hosts listed with their current certificates
Step 2

Generate the CSR for esxi-1: three-dot menu on the host > Generate CSR; Common Name = esxi-1.vcf.sddc.lab, Key Size = 2048 (ESX hosts only support 2048-bit keys, p.2926). Wait for the CSR task to finish.

CSR generated for esxi-1
Step 3

Replace the esxi-1 certificate: three-dot menu > Replace With Configured CA Certificate (select your CA, Confirm) or, for an externally signed certificate, Download CSR > Import Certificate > Replace With Imported Certificate. Watch the host in vCenter: it briefly reconnects as hostd and vpxa restart.

esxi-1 shows the new issuer and expiry in VCF Operations; host connected in vCenter
Step 4

Repeat Generate CSR and Replace for the remaining hosts one at a time. Alternatively, for hosts still on VMCA, use Renew Certificate (three-dot menu) to renew without changing the issuer.

All ESX hosts in the domain show certificates from your CA (or renewed VMCA certificates)
Stagger the hosts; each replacement restarts the host management agents.
Step 5

Break-glass manual method (documented, p.1271) — use only when VCF Operations cannot reach a host: in the VMware Host Client start the TSM-SSH service; in the ESX Shell, in /etc/vmware/ssl run mv rui.crt orig.rui.crt and mv rui.key orig.rui.key; copy in the CA-signed certificate and key as rui.crt and rui.key; run /etc/init.d/hostd restart and /etc/init.d/vpxa restart; stop TSM-SSH. Record that this host was changed outside VCF Operations so the inventory is re-checked.

Manual procedure documented as the exception path
Do not use the manual method on hosts VCF Operations manages unless the automated path is unavailable; keep the orig.* files until the host reconnects.
Step 6

Inventory what stays manual: (a) NSX Global Manager (self-signed by default, manual lifecycle — Table 721), (b) NSX Edge nodes (not applicable in Table 721; Edges use NSX Manager-issued certificates and need no action here), (c) external integrations that trust or pin VCF certificates (Kubernetes API servers, load balancers, syslog collectors, backup). For each, record owner, how it learns about a new VCF certificate, and the rotation procedure.

Manual-scope inventory with owners and procedures

Validation Gate

Check: (1) All ESX hosts in the domain show the new issuer and expiry in Fleet Management > Certificates (Show ESX Hosts) and are connected in vCenter, (2) break-glass manual procedure documented (rename rui.crt/rui.key, copy in, restart hostd and vpxa), (3) manual-scope inventory lists NSX Global Manager, NSX Edge status and every external integration with owner and procedure

Expected: ESX certificates lifecycle-managed from VCF Operations; everything outside VCF certificate management identified. VCF 9.0.1 docs pp.2918-2927 (Managing Certificates in VMware Cloud Foundation), p.1129 Table 721, p.1271 (manual ESX certificate), p.98 (release notes, certificate management).

Common Errors

Generate CSR for a host fails
Cause: Key Size left at 4096; ESX supports 2048 only (p.2926)
Fix: Regenerate with Key Size 2048.
After replacement a host stays disconnected in vCenter for more than 5 minutes
Cause: The host management agents did not restart cleanly, or the certificate chain was incomplete (external-CA path)
Fix: Check the VCF Operations task error; in vCenter right-click the host > Connection > Connect. If still failing, use the break-glass method (step 5) to restore orig.rui.crt/orig.rui.key, restart hostd and vpxa, then retry from VCF Operations with the complete chain.
Host replaced by the manual method shows the old certificate in VCF Operations
Cause: The inventory refreshes on its own schedule
Fix: Allow the inventory to refresh and re-check; confirm from the host with openssl x509 -in /etc/vmware/ssl/rui.crt -noout -issuer -enddate.

Task 5 Activate auto-renewal and set up certificate alerts and notifications

Manageability

Rotation is an event; expiry prevention is forever. VCF 9.0 ships both halves: automatic renewal of certificates issued by VMCA, Microsoft CA or OpenSSL for components that support non-disruptive updates (60 days before expiry), and Certificate Alerts in the VCF Operations console that you can route to email or an incident tool with a notification rule. The design work that remains is for the manual scope you inventoried in Task 4.

Step 1

In the VCF Operations console click Fleet Management > Certificates > VCF Instances > your instance (or domain). Use the Activate Auto-renewal toggle, review the information and click Confirm. Repeat under VCF Management for the management components. A configured CA is required for Microsoft CA or OpenSSL renewal; VMCA-issued certificates renew without one (pp.2924-2925).

Auto-renewal shows active for the VCF Instance and for VCF Management (status can take a moment to appear)
Automatic renewal happens 60 days before expiry for components in the Non-Disruptive Certificate table (ESX SSL, vCenter machine SSL, NSX LM and VIP, SDDC Manager SSL, VIDB, VCF Operations, Automation, logs, networks, Fleet Management).
Step 2

Open Fleet Management > Certificates > Certificate Alerts and review the alerts VCF Operations raises for certificates (expiry, replacement outcomes). Note the components and timings currently alerting.

Certificate Alerts view reviewed; current alerts documented
Step 3

Route certificate alerts to people: in Infrastructure Operations > Configurations > Outbound Settings confirm a Standard Email (and optionally Webhook/Slack/ServiceNow) instance is running, then in the Notifications tile add a notification rule whose Alert Scope selects the certificate alert definitions and whose outbound method is your email instance.

Notification rule active for certificate alerts
Exact alert-definition names vary by release — pick them from the Alert Definition selector in the rule rather than typing them.
Step 4

For the manual scope from Task 4 (NSX Global Manager, external integrations), implement an external check: a scheduled script that runs openssl s_client -connect <fqdn>:443 </dev/null 2>/dev/null | openssl x509 -noout -enddate against each endpoint and alerts when fewer than 30 days remain (email, Slack or syslog). Schedule it daily (e.g., cron 0 6 * * *).

External expiry check scripted and scheduled for endpoints outside VCF Operations
Step 5

Document the renewal design: what VCF Operations renews automatically (and 60 days before expiry), what needs Renew Certificate or Replace by an operator, and the procedure and owner for each manual-scope endpoint. As a VCDX reflection, compare the 60-day fixed lead time with an 80%-of-lifetime policy and state which your organisation's PKI standard requires.

Renewal design documented with automatic, operator-driven and manual scopes
Step 6

Test the notification path: trigger a test from the Standard Email outbound instance and, if your lab allows, use Renew Certificate on one ESX host and confirm the resulting task and any certificate alert reach your mailbox.

Test notification received; renewal task visible in VCF Operations

Validation Gate

Check: (1) Auto-renewal active for the VCF Instance and VCF Management, (2) Certificate Alerts reviewed, (3) notification rule routes certificate alerts to email (and optionally an incident tool), (4) external expiry check scheduled for manual-scope endpoints, (5) renewal design documented

Expected: No VCF certificate can expire undetected: VCF Operations renews the managed ones and alerts on all; the manual scope is covered by the external check.

Common Errors

Activate Auto-renewal fails or stays inactive for Microsoft CA/OpenSSL certificates
Cause: No CA configured for that scope (p.2924)
Fix: Configure CA for the VCF Instance (or VCF Management, Microsoft CA only), then activate auto-renewal again.
Certificate alerts appear in the console but no email arrives
Cause: The outbound instance is stopped or the notification rule's criteria exclude the certificate alert definitions
Fix: Test the Standard Email instance in Outbound Settings; edit the rule and include the certificate alert definitions and criticalities.
External check script reports the wrong expiry for a VIP
Cause: openssl s_client returned the node certificate instead of the VIP certificate because SNI was not sent
Fix: Add -servername <fqdn> to the openssl s_client command.
Operator assumes Microsoft CA certificates cannot auto-renew
Cause: Outdated assumption from earlier releases
Fix: VCF 9.0 supports auto-renewal for VMCA, Microsoft CA and OpenSSL-issued certificates on components that support non-disruptive updates (release notes p.98; pp.2924-2925).

Final Validation

You have audited VCF 9.0's certificate architecture from the VCF Operations inventory, configured a Certificate Authority in VCF Operations, replaced the SDDC Manager, vCenter, NSX Manager and ESX host certificates from the default VMCA to your enterprise CA, identified what stays manual (NSX Global Manager, external integrations), and activated auto-renewal with alert notifications. The VCF stack is now under your organisation's PKI governance.

✓ Certificate architecture mapped from Fleet Management > Certificates: VCF Management and VCF Instance components with issuer, expiry and trust paths; manual scope identified (Table 721) → Inventory table and architecture map complete

✓ CA configured in VCF Operations; CSRs generated for SDDC Manager, vCenter and NSX Manager → CSRs present on the three components; CA saved for the VCF Instance

✓ SDDC Manager, vCenter and NSX Manager replaced from VCF Operations → New issuer and expiry in the inventory; openssl s_client confirms; no trust errors

✓ ESX host certificates replaced (or renewed) from VCF Operations with Show ESX Hosts; break-glass manual method documented → All hosts show the new certificates and are connected in vCenter

✓ Auto-renewal active and certificate alerts routed by a notification rule → Activate Auto-renewal confirmed for VCF Instance and VCF Management; test notification received

✓ Renewal design documented: automatic (60 days before expiry), operator-driven and manual scopes with owners → Design document with integration points and responsible teams

Cleanup / Restore

Snapshot: management-domain-post-cert-rotation

• Take a snapshot of SDDC Manager, vCenter, and NSX Manager post-rotation: 'management-domain-post-cert-rotation'

• Document and retain: the new issuer and expiry for each component as shown in Fleet Management > Certificates — attach to your lab notebook

• Leave auto-renewal active (it is the desired steady state); remove the test notification rule if you do not want lab alerts in your mailbox

• If you need to revert certificate changes, restore to the 'management-domain-pre-cert-rotation' snapshot (note: this reverts to the default VMCA-issued certificates)

Design Reflection (VCDX)

A VCDX panelist examining a multi-domain VCF certificate strategy would ask: How do you handle certificate replacement across 50+ workload domains without cascading outages, now that VCF Operations does it per VCF Instance? What is your renewal lead time — the product's 60-day auto-renewal, or a stricter PKI standard — and how does it fit your change calendar? If the enterprise root expires, what is the order for intermediates and leaves? What stays manual (NSX Global Manager, external integrations) and who owns it? What is your break-glass procedure if the CA is compromised? Be ready to defend 'why an enterprise CA instead of staying on VMCA' and 'what is the audit trail for every replacement'.

Requirements

  • All VCF components (SDDC Manager, vCenter, NSX Manager, ESXi hosts) must use valid X.509 certificates with CN and SAN matching their FQDNs
  • All component-to-component communication (SDDC Manager → vCenter, NSX Manager ↔ SDDC Manager, ESXi ↔ vCenter) must validate certificates to prevent man-in-the-middle attacks
  • Certificates must be rotated before expiry (ideally 30-90 days in advance) with zero disruption to operational workloads
  • Audit trail: every certificate replacement and renewal must be traceable (VCF Operations tasks and Certificate Alerts for managed components; change records for the manual scope)
  • Certificate expiry must be detected proactively (alerting at 30 days remaining) and never allowed to lapse undetected

Constraints

  • Replacing the SDDC Manager or vCenter certificate restarts services and causes a brief UI interruption — schedule in a maintenance window; ESX replacement restarts hostd and vpxa and briefly disconnects the host from vCenter (guest VMs unaffected)
  • VCF Management components support Microsoft CA only; VCF Instance components support Microsoft CA and OpenSSL; the CA is configured separately per scope and per VCF Instance (p.2920)
  • ESX hosts accept 2048-bit keys only (p.2926); NSX Global Manager certificates are managed manually and NSX Edge certificates are not applicable in VCF certificate management (Table 721)
  • Trust chain must be complete when importing external certificates (server, intermediate, root in PEM); the configured-CA path handles the chain
  • External integrations that trust or pin VCF certificates (Kubernetes API, load balancers, syslog collectors) break if not coordinated with the replacement

Assumptions

  • A Microsoft CA with Web Enrollment on the same server and a VMware template is reachable from VCF Operations, or the OpenSSL CA on SDDC Manager is acceptable for VCF Instance components
  • VCF Operations with fleet management is deployed and the VCF Instance is connected; the SDDC Manager certificate UI/API is deprecated in 9.0 and not used
  • All component FQDNs (sddc-manager.vcf.sddc.lab, vcenter.vcf.sddc.lab, etc.) resolve consistently and match certificate CN/SAN
  • Change control process allows 30-90 day advance scheduling for cert rotations
  • Operations team has PKI/certificate experience or will be trained on CA CSR submission and cert installation
  • Monitoring infrastructure (syslog server, email relay, or alerting system) is available for cert expiry alerts

Risks

  • Certificate expiry: if auto-renewal is not activated and alerts are not routed, a certificate expires and the component becomes unreachable or untrusted (vCenter expiry isolates its hosts). MITIGATION: activate auto-renewal for VCF Management and every VCF Instance, route Certificate Alerts to email and the incident tool, and run the external check for the manual scope.
  • Trust chain breakage: an externally signed certificate imported without its intermediate/root fails validation or breaks component trust after replacement. MITIGATION: prefer the configured-CA path; for external CAs import the full PEM chain and Validate before Replace; test on one component first.
  • Certificate mismatch: leaf cert subject (CN, SAN) doesn't match component's FQDN, or private key doesn't match public cert. SSL/TLS handshake fails. IMPACT: component unreachable or untrusted. MITIGATION: validate CSR subject before CA signing; validate cert post-installation (openssl verify, openssl x509 -subject).
  • Replacement sequencing: replacing several components back-to-back during one window compounds the brief restarts. MITIGATION: one component at a time from VCF Operations, verify each with openssl s_client and the Certificate Alerts view before the next.
  • External integrations break: if Kubernetes API server has cert pinning for vCenter's CA cert, rotating vCenter to new CA breaks K8s-vCenter mTLS until K8s is reconfigured. IMPACT: K8s worker nodes may disconnect from vCenter. MITIGATION: maintain inventory of external integrations with cert dependencies; notify teams 2 weeks before rotation; coordinate update of external system certs in parallel with VCF rotation.

Self-Assessment Discussion Prompts

  1. You have 20 workload domains across 3 data centers. Each domain has its own NSX Manager and vCenter. Do you use a single root CA for all 60 components, or separate CAs per data center? What are the governance and operational trade-offs?
  2. Your internal CA root cert expires in 2 years. All 50 leaf certs (SDDC Manager, vCenters, NSX, ESXi) are issued by intermediates under that root. Design the renewal sequence: when do you rotate the root? When do you rotate intermediates? When do you rotate leaf certs? What's the critical path?
  3. A security audit discovers that 5 ESX hosts were replaced by the manual rui.crt/rui.key method with certificates from an unauthorised CA. How do you detect this drift in the VCF Operations inventory, and how do you remediate from Fleet Management > Certificates without guest downtime?
  4. Your monitoring system detects that NSX Manager's cert will expire in 7 days, but you're in production deployment crunch and can't schedule a maintenance window for 2 weeks. What do you do? (This is a real-world scenario.)
  5. VCF Operations auto-renews managed certificates 60 days before expiry. Design the renewal for everything it does not cover (NSX Global Manager, external integrations) so that it needs no human intervention. What are the failure modes if the CA is unavailable? If a renewal fails mid-way?
  6. You are moving from the default VMCA-issued certificates to enterprise CA certificates. Which external integrations break (systems that imported the VMCA root, pinned fingerprints, partner APIs, public-facing load balancers) and how do you sequence their updates?

Extensions

Implement certificate pinning for critical management APIs

Certificate pinning locks a specific certificate (or public-key hash) into a client so that even a valid certificate from a different issuer is rejected. Implement pinning on one external path — a monitoring script or integration that calls the SDDC Manager API — and then run Renew Certificate on SDDC Manager from VCF Operations. Observe the break, document the pin-update procedure, and decide where pinning is worth the operational cost in a platform whose certificates auto-renew every 60 days before expiry.

harder

Compare VCF 5.2.x certificate management with 9.0.x workflow

In VCF 5.2 certificate management (CA configuration, CSR generation, replacement) lived in SDDC Manager (Administration > Security > Certificate Authority; Inventory > Workload Domains > domain > Certificates). VCF 9.0 moves it to VCF Operations (Fleet Management > Certificates) and deprecates the SDDC Manager UI/API, adds auto-renewal, and extends the scope to the VCF Management components. Research and document: (a) the 5.2 procedure, (b) how 5.2 handled chains and ESX hosts, (c) whether 5.2 alerted on approaching expiry, (d) operator time to replace one domain in 5.2 vs 9.0. If you have a 5.2 lab, perform the 5.2 replacement and time it.

same

Design and implement certificate automation for external integrations

Extend your VCF cert infrastructure to external systems: Kubernetes API server, syslog collector, backup system (e.g., Veeam), etc. Generate certs for these systems from your internal CA. Implement a certificate delivery mechanism (e.g., REST API endpoint on a management workstation that external systems can poll daily for updated certs). Test end-to-end: external system requests cert, receives updated cert, loads it (without restart if possible). This simulates the complexity of managing certs across a heterogeneous infrastructure.

harder

Simulate a certificate compromise and execute emergency remediation

Assume your internal CA's private key was compromised (or suspected). All certs issued by that CA are now untrusted. Design and execute an emergency cert rotation: (a) revoke all VCF certs, (b) generate new certs from a new CA (or new intermediate), (c) rotate all components within a 4-hour window to minimize exposure. Document the incident response procedure, decision tree for when to escalate, and communication plan for external stakeholders. This is a real-world security incident that tests your operational readiness.

harder

References

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