Pillar Axiom 300 Storage System
 
Pillar Axiom 500 Storage System
 
Pillar Axiom 600 Storage System
 

PLA-Support>Sun Systems>DISK>Pillar Axiom>SN-DK: Ax600
 

In this Document

Goal

Fix

Applies to:

Goal

Thin Provisioning: Operating system allocation

Growth increments using Thin Provisioning

Fix

Thin Provisioning is the idea that you can create a LUN or File System in the Axiom that incorporates more size than is physically allocated to it.



Growth is automatic and more storage may need to be added when reacing 90% capacity.
When a new 3.x LUN or File System is created, metadata is allocated for the size of that storage space and an automatic growth increment is set. The default increment set is 1% of the total storage object size. As of 3.1, the increment is set to either 1% of the size or the original LUN size growth before the update. Then the first few logical block addresses of the disk are allocated. After this it is up to the host.

Host Access

Once the host has gained access to the LUN or File System, the operating system will control the read and write flow. If the writes to the disk are enough to cause infill across the LUN, it will use the full capacity and allocate the entire space. This is done over time as the host writes data to all the LBA ranges.

Using all Allocated Space

What happens next is entirely up to the file system. There are some highly virtualized file systems that only allocate inodes as they are needed. If the inodes are not created, then you don’t need to create all the directory and indirect blocks that point to them, which means the LUN doesn’t need to allocate storage by infill until something is written to an LBA range. Other file systems splatter multiple copies of the file system superblock every so many increments across the file system, again, more storage.

For Windows systems, either format option will not allocate the entire storage range.

For Linux systems, it is recommended to use a volume manager to prevent the OS from allocating the entire range. The manager will only touch the space necessary as the file system grows. ZFS and VMFS are two examples of volume managers that work well.

For More Information

www.cuddletech.com/veritas/vxcrashkourse/ar01s02.html

This solution has no attachment