Hyper size

IMHO, the factors to take into account regarding hyper size are as follows (not in any particular order) :

- We are taking regular thick devices here, not save area / Data devs etc...

- Does the hyper size maximize the useable space on the physical disk (note that a small amount of free space left over can be good for GK's and Quorum devices). Also see below regarding VDEVS / SAVE area.

- What is the total amount of storage you want to address vs. how many FA ports you have / # FA's each device is mapped to. There is a finite number of devices (including meta members) that an FA port can have a mapping for. The larger the hyper the higher the total amount of capacity a given FA port can address. Small hypers can max out the available addressing before the box is full of physical disk.

- What is the granularity that the customer wants to provision (increments) in capacity.

- Are you spreading the across the back end wide enough to cope with host workload (regarding meta devices)

- How much local replication do you intend to use. Consider every symdev you create multiplied the number of replicas you intend to have - have you reached any symdev / max hypers per physical disk limits?

- If you intend to use Virtual device, it is best to spread the SAVE area as wide as possible, across all disks if possible, so work out your desired save area / # physicals to determine how much space to use per disk for save area (1 hyper per physical is best of possible). The remaining space on the physical disk is for your regular hypers.

If you take the above into account your hyper size is sure to be truly magical

Hope this helps.

OF course this will help a lot to a novice like me!!!!   Thanks a lot!!!!  Do we have a detailed literature or white paper on this.

I am not aware of any whitepapers specific to this area. There is some useful information in the Solutions Enabler Symmetrix Array Controls CLI product guide. Also some of the limits described previously are detailed in the DMX4 product guide.