Increase Datastore capacity, what with block size ?

[MagicRecon]

Hi,

I'm trying to extend my 2TB datastore, I've added a new 2TB lun from the same 4TB FC SAN and there is no problem adding this to the existing datastore.

But the existing datastore is formatted in VMFS3  with a block size of 4MB to get a max of 512GB file size. But when i try to extend the datastore the option to select the block size is set to 1MB and greyed out.
Does the extend automaticly take over the original block size or wil it get a 1MB block size ? There are some >256GB files on the existing datastore and i think i would get some strange behavior if half the datastore is 4MB and the other half is 1MB blocksize.

Can anyone confirm that the blocksize wil be the same for the extend ?

I can also make a second datastore of 2TB an manualy divide the VM's over the 2 datastores. are there any advantages or disadvantages to this aproach ?

[a_p_ · Accepted solution]

Don't worry, adding an extent to an existing datastore will certainly not change the block size. Anyway, if it is an option for you to use two datastores, I would do that. With the two separate datastores you will not only be able to distribute the load on each of these LUNs but also avoid complexity. With ESXi 5 (VMFS-5) a single LUN can grow up to ~64TB without the need of extents.

Btw, a 4MB block size allows 1TB minus 512 bytes.

André

[MagicRecon]

Ok thanks for the info. I've created a VMFS 5 datastore on the new LUN and i am migrating the VM's as we speak. when all machies are migrated i'll upgrade the old VMFS 3 datastore to VMFS 5

I will still have 2 datastores instead of one. because i don't have the space to clear out the 4TB at once. but at least there vmfs5