Not content with pulling disks and SSDs out of nodes, I also did some testing of powering off a node. I used a small cluster, two SimpliVity nodes, with 80 workload VMs. At the start, the VMs were split between the two nodes. Then I powered off one node using it’s built-in IPMI management controller. As you would hope, vSphere HA did its job and after a few minutes all the VMs were running on the one surviving node.
I’ve been doing a lot of work with SimpliVity and wanted to share a few videos of what happens when things go wrong with the hardware in their hyperconverged platform. This video looks at the results of hard disk and SSD failure in a SimpliVity node. A cluster running 120 VMs had no outage with two hard disks and one SSD failed.
Steven and I talk about some of what he learned deploying vMSC with HP 3Par storage.
We look at what vSphere Metro Storage Cluster is and how Steven has implemented it.
Steven blogs at http://ssbkang.com/
What is zeroing?
Should I be lazy or eager?
What about thin provisioning?