As a unified storage array, the Unity XT allocates block LUNs and file systems on the same storage pool.
Block LUNs are required for FC connections. Every SAP HANA node requires two LUNs, one for data and a second for log.
When creating LUNs for the SAP HANA persistence, follow these configuration guidelines:
Use the Unity XT Unisphere UI to create LUNs. From the dashboard, select Storage > Block > LUNs, and then click the + icon.
A NAS server is a file server that uses the SMB protocol, NFS protocol, or both to share data with network hosts. It also catalogs, organizes, and optimizes read and write operations to the associated file systems. Before you can provision file system storage, ensure that a NAS server that is appropriate for managing the storage type is running on the system.
Creating NAS servers for the SAP HANA persistence requires you to specify the following information:
Use the Unity XT UI to configure NAS servers. From the dashboard, select File > NAS Servers, and then click the + sign to create NAS servers.
Run the following NAS services command to enable large transfer sizes. See the Unity XT documentation for information about how to set the parameter.
svc_nas ALL -param -facility nfs -modify v3xfersize -value 1048576
File systems are required for NAS (NFS) connectivity. Before you can create a file system, you must create NAS servers, as described in Create NAS servers.
When creating file systems for the SAP HANA persistence, follow these configuration guidelines:
Configure file systems through the Unity XT UI. From the dashboard, select Storage > File > File Systems, and then click the + icon to create file systems.
You can create the NFS shares during file system creation or, for an existing file system, afterward. From the dashboard, select Storage > File > NFS Shares, and then click the
For maximum performance and scalability, thick devices are recommended for SAP HANA production systems. Thin provisioning is required to take advantage of the Unity XT data services, but thin provisioning and snapshots have an impact on performance and scalability. Using data reduction and advanced deduplication is not recommended in SAP HANA production environments.
When you provision storage on a Unity XT storage array, whether as block LUNs or file systems, the following data services are available for space efficiency and operational requirements:
Thin provisioning can improve storage efficiency while reducing the time and effort that are required for monitoring and rebalancing existing pool resources. Organizations can purchase less storage capacity up front and increase available drive capacity by adding drives on an on-demand basis and according to actual usage, instead of based on the requests or predictions of connected hosts. With thin provisioning, multiple storage resources can subscribe to common storage capacity within a pool, while the system allocates only a portion of the physical capacity that is requested by each storage resource. The remaining storage is available for other storage resources.
Thin provisioning is enabled by default on a Unity XT storage array. However, we recommend turning off thin provisioning for the persistence (data and log) of production SAP HANA systems. For example, thin provisioning can be used for /hana/shared or nonproduction SAP HANA persistence.
If you are using thin provisioning for production SAP HANA persistence, reduce the total number of SAP HANA nodes, as specified in Table 2, by approximately 20 percent.
Applying data reduction saves space on the system. Data reduction includes savings from deduplicating blocks that contain internally defined data patterns including zero detection and compression. Data reduction is available for thin LUNs, thin file systems, and thin NFS and VMFS datastores that are created in all-flash pools. The thin file system must be created on a Unity XT system that is running OE version 5.0.x or later.
The advanced deduplication option helps increase the storage efficiency of systems beyond what is achieved with data reduction on a LUN, file system, or datastore. Advanced deduplication enables you to reduce the amount of storage that is required for user data by keeping only a small number of copies (often just one copy) of a data block. You can enable advanced deduplication only if data reduction is enabled first.
These data services have a significant impact on performance and so are not supported for the persistence (data and log) of production SAP HANA nodes. SAP HANA has already implemented highly efficient space-saving algorithms for the database persistence, which reduces the benefit of space savings on a storage array.
Data reduction and advanced deduplication services can be used for nonperformance critical devices like the /hana/shared file system.
A snapshot is a point-in-time copy of data stored in a specific storage resource on the storage system. [KA3][KW4]Periodically creating snapshots of storage resources provides an effective technique for meeting data protection and recovery requirements.
Snapshots are supported on LUNs and file systems. When using snapshots in SAP HANA environments, ensure that array snapshots are synchronized with the database internal storage snapshots.
The impact on the storage performance for the SAP HANA persistence depends on several factors, including the data change rate, the number of snapshots, and the retention policies. We strongly recommend engaging Dell EMC Professional Services to ensure that the SAP HANA performance requirements are met when you implement snapshots.