Good day.
I`m trying to get optimal formula for calculation of datastore capacity, where linked-clone pool resides.
There is one in Sizing Storage for VMware View Linked Clones » myvirtualcloud.net:
(n° VMs * (Delta Footprint + Log + User Data Disk) + Replica + Master Replica +Overhead)
Andre takes vm with 20 Gb hard disk, 2Gb RAM and 2 Gb user persistent disk and he gets this (for 100vm):
(100 * (6144MB + 100MB + 2048MB) + 20GB + 124GB)
I don`t get few things:
1. Why Delta Footprint is 3*RAM, not 2*RAM (I mean there are 2 Gb vswp and 2 Gb on disposable disk what i missed)? Also I saw 2 vswp files in linked clone vm directory, for example: vmx-vcli-03-1188161764-1.vswp and vcli-03-46d1e8e4.vswp, 1480 and 1277 Mb respectively, I don`t get why there are two of them, not one?
2. Why Replica + Master Replica + Overhead = 20GB + 124GB? Why he do not calculate hypothetical linked-clone growth? What is overhead? It says here http://www.vmware.com/files/pdf/view_storage_considerations.pdf :
Replica Overhead
As a best practice, allow for 2 times the replica size on each LUN you are going to provision.
But i provide separate datastore for replica and i do not see replicas on different LUNs (may be it is hidden, but i can see 1 replica on the datastore that i provide during pool deployment).
I thought I want to use this formula (in case when replica and master image resides on different datastore):
N * ((RAM-reservation) + Log + User disk + disposable disk + Hypothetical linked-clone growth)
So what is wrong with my formula, what is overhead and why Andre do not calculate linked-clone growth in his formula?
Thanks in advance and sorry for my bad english.
Update: also i don`t get what is:
Linked Clone Base Size
Assume a base linked clone size of 50 percent of the replica.
Why linked-clone growth calculates in percentage from linked clone base size, not from replica disk size?
