I just have a question related to the disk DWPD that I have a hard time finding the answer in google.
Consider this scenario:
- 1 server with only 1 nvme disk of 1TB (no SATA, no Optane, no SD cards, just 1 disk)
- VMWare ESX 6.7
- Let’s say I create 1 VM with an allocated disk space of 200GB
My question is:
If I write all day long on those 200GB, will the nvme disk internal firmware be “smart enough” to spread the writes across all disk space (200GB assigned + 800GB unallocated), or will those write be always happen on the same “physical” 200GB cells?
I ask because if the answer is “it will write on the same 200GB” then the TBW will be reached “5 times faster” (writes go to the same 200GB instead to the total 1000GB) and I’ll will need to take that into consideration while assigning the disk space for each VM
Hope I’m explaining myself.
Yes, the physical media will be used approximately evenly. This is called wear leveling and solid state drives have been doing it from almost the beginning. They would not have been viable as storage without it.
This work is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.