HP LeftHand P4000 SAN Solutions - Manual (Page 142)

Thin provisioning
Thin provisioning reserves less space on the SAN than is presented to application servers. The
LeftHand OS software allocates space as needed when data is written to the volume. Thin
provisioning also allows storage clusters to provision more storage to application servers than
physically exits in the cluster. When a cluster is over-provisioned, thin provisioning carries the risk
that an application server will fail a write if the storage cluster has run out of disk space. The
LeftHand OS software adds utilization and over-provisioned events as the cluster approaches 100%
utilization. You can add capacity to the cluster or delete unneeded snapshots to accommodate
additional volume growth.
NOTE: Paying attention to the space utilization events on over-provisioned storage clusters is
critical to prevent a write failure on a thin volume.
Best practice for setting volume size
Create the volume with the size that you currently need. Later, if you need to make the volume
bigger, increase the volume size in the CMC and then expand the disk on the server. In Microsoft
Windows, you expand a basic disk using Windows Logical Disk Manager and Diskpart. For
detailed instructions, see
“Changing the volume size on the server” (page 153) .
Planning data protection
Data protection results from creating data redundancy for volumes on the SAN. Configure data
protection levels, called Network RAID, when you create a volume. Because data is stored
redundantly on different storage systems, all data protection levels are tied to the number of
available storage systems in a cluster. Network RAID achieves data protection using can use either
replication or parity.
•
Data protection with replication - Data protection using replication is achieved using Network
RAID-10, Network RAID-10+1, or Network RAID-10+2, which store two, three, or four mirrored
copies of the data.
•
Data protection with parity - Data protection using parity is achieved using Network RAID-5
and Network RAID-6, which store data and parity dynamically according to the number of
storage systems in a cluster. Under some workloads that write the data infrequently, Network
RAID-5 and Network RAID-6 provide better capacity utilization and similar high availability
to Network RAID-10 and Network RAID-10+1.
Data protection level
Seven data protection levels are available, depending upon the number of available storage
systems in the cluster.
142 Provisioning storage