I’m sure the topic of VMFS block size considerations has been discussed plenty of times and a quick search on the internet will yield plenty of blogs and articles about it but I’ve noticed that I almost always get asked this question whenever I’m doing knowledge transfer to clients.
Other than the obvious reasons such as the maximum size limit of the vmdk (http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1003565), space usage (most argue this is negligible), performance (http://www.yellow-bricks.com/2009/03/24/an-8mb-vmfs-blocksize-doesnt-increase-performance/) and thin disk provisioning considerations, there’s one other that I noticed that doesn’t get brought up as much and that’s LUN to LUN copy performance. While the ever improving SAN performance may make this a non-factor, I think it’s worth mentioning that if you’re performing a copy operating between two LUNs with the same block size, performance tends to be faster because the blocks of data no longer needs to be resized.
I’ve also written another post about VMDK partition alignment (GUI vs Command Line) which may be of interest for those who may want to know more about the differences between using the GUI and command line: http://terenceluk.blogspot.com/2011/01/vmdk-partition-alignment-gui-vs-command.html
No comments:
Post a Comment