First the facts:
If you assign a VMFS volume with 1MB blocksize the thin provisioned disk will grow in 1MB increments. Think of this in terms of disk I/O. If you where to increase the size of a VM by placing a 8GB file on the system you would end up with the following:
8MB Block size: 1,000 grows vs. 1MB Block size: 8,000 grows (Which do you think has less over head?)I found this paragraph on many blogs but not sure who wrote it:
"If you create a thin provisioned disk on a datastore with a 1MB blocksize the thin provisioned disk will grow with increments of 1MB. Hopefully you can see where I’m going. A thin provisioned disk on a datastore with an 8MB blocksize will grow in 8MB increments. Each time the thin-provisioned disk grows a SCSI reservation takes place because of meta data changes. As you can imagine an 8MB blocksize will decrease the amount of meta data changes needed, which means less SCSI reservations. Less SCSI reservations equals better performance in my book."I have to admit that really in the world today why wouldn't you just choose to go with the 8MB block size. If you're using Jumbo Frames on iSCSI wouldn't you also think that bigger is better? If you have proof that this is not the case please let me know. I want to get to the bottom of this.
It was my personal experience when I moved all of the virtual machines to the same EqualLogic with 4MB or 8MB block size they performed better and I was able to put more machines on a datastore. If you find the same performance increase please let me know.