What actually happens?
Traditional backups copy entire files or virtual disks, which takes time and slows down your system. Nutanix uses a Redirect-on-Write (RoW) architecture combined with instantaneous metadata pointers.
1๏ธโฃ ๐๐ฉ๐ฆ ๐๐ฐ๐ช๐ฏ๐ต๐ฆ๐ณ๐ด ๐๐ณ๐ฆ๐ฆ๐ป๐ช๐ฏ๐จ (๐๐ฏ๐ด๐ต๐ข๐ฏ๐ต):
- Instead of copying your data blocks, Nutanix copies the metadata map (the index or map that tells the system where your data blocks live on the physical disks).
* The Process: The system takes a snapshot of this map, which takes milliseconds because metadata is incredibly small.
* The Result: The original data blocks become read-only for that specific point in time.
2๏ธโฃ ๐๐ฉ๐ฆ ๐๐ฆ๐ฅ๐ช๐ณ๐ฆ๐ค๐ต-๐ฐ๐ฏ-๐๐ณ๐ช๐ต๐ฆ ๐๐ฆ๐ค๐ฉ๐ข๐ฏ๐ช๐ด๐ฎ (๐๐ฐ ๐๐ท๐ฆ๐ณ๐ฉ๐ฆ๐ข๐ฅ):
- The moment after the recovery point is created, your Virtual Machine (VM) keeps running and writing new data.
* Old Data: Stays exactly where it was, safely frozen as your recovery point.
* New Data: Instead of overwriting old data, any new changes are written to completely new locations on the storage array. The VMs active metadata map is updated to point to these new locations.
* Because no data was moved or copied during the creation phase, there is zero impact on VM performance and no waiting time.
3๏ธโฃ ๐๐ฉ๐บ ๐๐ฆ๐ค๐ญ๐ข๐ช๐ฎ๐ข๐ฃ๐ญ๐ฆ ๐๐ฑ๐ข๐ค๐ฆ ๐๐ฉ๐ฐ๐ธ๐ด 0 ๐:
- The Reclaimable Space column displays the unique data blocks that would be physically deleted and freed up from the storage pool if only that specific snapshot were deleted.
- Because Nutanix snapshots use a Redirect-on-Write (RoW) and pointers architecture, consecutive recovery points share overlapping blocks. If the data inside the VM hasnt changed much between snapshots, or if those blocks are still referenced by other active recovery points in the chain, deleting just this snapshot yields 0 bytes of extra capacity back to the cluster
๐ง๐ต๐ฎ๐ ๐บ๐ฎ๐ธ๐ฒ๐ ๐๐ต๐ฒ ๐ถ๐ป๐ถ๐๐ถ๐ฎ๐น ๐๐ป๐ฎ๐ฝ๐๐ต๐ผ๐ ๐ผ๐ฝ๐ฒ๐ฟ๐ฎ๐๐ถ๐ผ๐ป ๐ฒ๐
๐๐ฟ๐ฒ๐บ๐ฒ๐น๐ ๐ณ๐ฎ๐๐, ๐ฟ๐ฒ๐ด๐ฎ๐ฟ๐ฑ๐น๐ฒ๐๐ ๐ผ๐ณ ๐๐ต๐ฒ๐๐ต๐ฒ๐ฟ ๐๐ต๐ฒ ๐ฉ๐ ๐ถ๐ ๐ฑ๐ฌ ๐๐, ๐ฑ๐ฌ๐ฌ ๐๐, ๐ผ๐ฟ ๐๐ฒ๐๐ฒ๐ฟ๐ฎ๐น ๐ง๐.