vmfs question

[bellocarico]

A question on the VMFS vlock size.

If I have a 500Gb datastore wirth 2Mb block size, does that means that writing a smallish file (e.g. 1K) on a VM that uses that datastore will commit 2Mb on the disk no matter what?

If that's correct followin the same logic does a 2TB datastore with 8Mb block size commit minimum 8Mb at the time?

Thanks

[DSTAVERT · Accepted solution]

Sub blocks are used to allocate smaller chunks of data. There is no penalty in using 8MB datastores as far as efficient use of storage.

[DSTAVERT]

http://www.yellow-bricks.com/2009/05/14/block-sizes-and-growing-your-vmfs/

[chriswahl]

I don't think so, there is a way that it can slice up smaller blocks out of the block size for those files. Although, there isn't all that many files on a VMFS datastore anyway (vmx, vmdks, a few others) so it wouldn't matter that much either way.

I default all datastores to 8MB blocks as it is the most efficient way to ensure no issues later, and that you can always use the max size supported on VMFS3.

Edit: DSTAVERT beat me to it

[bellocarico]

Brillian! Thanks 🙂

[RR9]

if you are not having any disk of 2TB then there is no need to have 8MB block size. Then its better to have smaller block size and small size  datastore.

its better to have multiple datastores of different block size.