Lenovo Support in France After End of Life

One Failure, Two Losses

In a traditional estate a failed server costs compute and leaves storage untouched, because they are different equipment. A hyperconverged node is both: it ran workloads and it contributed capacity to the cluster. Losing it removes compute and triggers a data rebuild in the same instant, and the two consequences compound rather than sit side by side.

The Rebuild Lands at the Worst Moment

The surviving nodes absorb the failed node workloads and simultaneously begin reconstructing its data across themselves. They are doing more work with fewer resources while also moving large volumes of data between themselves. That is why a hyperconverged cluster feels slowest immediately after a failure rather than during it.

Spare Capacity Is the Mechanism, Not Waste

A cluster sized with no headroom for the loss of a node cannot complete the rebuild, and remains degraded until the node is physically replaced. Headroom that looks like unused capacity on a report is the specific thing that lets the cluster recover without intervention, and it is frequently the first thing reclaimed when capacity is tight.

Two Things to Have Ready

Drives matched to the tiers in use, because that is what a rebuild consumes, and confirmation that a node is obtainable, because that is what ends the degraded state. ICD supplies both for ThinkAgile and ThinkSystem generations past vendor support, shipped to France in 3 to 5 days, tested and covered for one to three years.