19 articles in this category
Which VCF components come as a pair and which do not, why a cluster is three nodes when a body gets by with two kidneys, and what vSphere HA is really doing underneath all of it.
VCF Installer lets you leave VCF Automation out of a converge, finish the platform, and then bring your existing VMware Aria Automation in and upgrade it. The option is documented in three places, and the detail worth knowing is which menu you come back through.
Fourteen resolved issues, five of them in the upgrade path. One of them lifts a datastore restriction, one formalises the Migration Service Engine ordering rule, and one was silently throwing away your proxy configuration while reporting success.
Forty three prechecks run before a VCF Automation upgrade, and most of them are testing things you can check yourself a week earlier. Here is what they look for, what the exit codes mean, and the thresholds worth knowing before you book the window.
Deploying a VCF instance asks you two questions about management services: which components run here, and how is the cluster built. The documentation answers both in two tables and two diagrams. Here is the same thing as a story.
VCF 9.1 adds password policies, so you can set length, complexity, expiration and lockout once and apply them across the fleet. VCF Operations keeps no copy of your passwords, and CyberArk integration gives you somewhere to store them.
Sizing for VCF Operations Orchestrator is a short answer, one fixed appliance spec and exactly three nodes. The longer part is building the cluster, which is a load balancer, three OVA deployments and five commands over SSH.
Every published VCF Automation configuration maximum from 9.0.0 to 9.1.1 on one page, Medium and Large profiles side by side. Those numbers are the All Apps figures, so the vRealize Automation limits that a VM Apps organization keeps are here too.
Two dropdowns in the planning workbook decide the VCF Automation footprint. Small deploys a single node, Medium and Large deploy a three-node cluster, and the numbers follow from that. Figures taken from the VCF planning and preparation workbooks.
If the Migration Service Engine was patched before VCF Automation, both components end up stuck with a PackageDeploymentsCompleteWaitTimeout error. Broadcom KB 455559 has the recovery, and this is what it involves.
Patching the VCF Services Runtime component to 9.1.1.0, from prechecks through a five hour rolling replacement of every node, with the kubectl commands to follow it from the control plane.
A patching order gotcha in VCF 9.1.1.0. If you patch the migration service engine before VCF Automation, the VCD_MIGRATOR upgrade runs for six hours and then fails on vmsp_upgrade.
The order in which to patch the VCF management components when going from 9.1.0.x to 9.1.1.0. Two components go in sequence, then the rest follow in no fixed order, with two that must run on their own.
A deep dive into what actually happens when VCF Automation is upgraded to 9.1.0.0, covering the inputs you need, the import, configure and precheck phases, and the new cluster that gets built during the upgrade.
Walking through the SDDC Lifecycle patch to 9.1.1.0, from setting the target version and running prechecks through to the component showing the new version.
Patching the VCF Operations component to 9.1.1.0, including the prechecks, the UI downtime to expect during the patch, and monitoring progress from the VCF Operations Admin UI.
A walkthrough of patching the VCF management services fleet lifecycle component from 9.1.0.0 to 9.1.1.0, with screenshots of the UI steps and what happens on the control plane node.
VCF 9.1.1.0 has been released as a maintenance release with an updated BoM. Here are the release notes links plus my notes on the known and resolved issues in VCF Automation and VCF Operations.
Fleet Management node 9.0 convergence