This post has been republished via RSS; it originally appeared at: New blog articles in Microsoft Community Hub.
Table of Contents
This article describes the value and use cases for Azure NetApp Files (ANF) volume provisioning using Azure NetApp Files application volume group (AVG) for SAP HANA. The article refers to a series of videos, which explain the pre-requisites as well as the different deployment scenarios in detail.
Co-authors: Nils Bauer (SAP Technical Marketing Engineer, NetApp)
Since the general availability of Azure NetApp Files application volume group feature for SAP HANA it has become the standard provisioning method for SAP HANA when using Azure NetApp Files.
It is well understood that SAP HANA applications are performance demanding and require sub-millisecond latency and high throughput numbers. The placement of the ANF volumes used for SAP HANA, relative to Azure compute, is critical to fulfil these requirements. With the normal workflow provisioning ANF volumes, the achievable latency and performance is often not optimal, and manual interaction with the Microsoft Azure team was required.
With application volume group for SAP HANA an application specific provisioning workflow has been developed, which covers three main areas:
- it optimizes ANF volume-to-compute placement to provide lowest latency and best performance,
- it greatly simplifies the volume provisioning by configuring all required HANA volumes in a single step and by pre-configuring the volume capacity and throughput numbers, using deployment best practices,
- it supports the main SAP HANA configuration options including high availability and disaster recovery.
To understand why the application volume group feature is crucial for HANA deployments on ANF, it is critical to understand the relation between the various infrastructure components at the compute, network, and storage layers.
Compute as well as Azure NetApp Files resources are organized in clusters which are connected to the highspeed Azure network. To optimize latency, the Azure VMs as well as the ANF bare-metal fleet resources hosting the HANA volumes must be ideally in close proximity to each other. For this purpose, Azure proximity placement groups are leveraged.
In addition to the proximity of the HANA VM and the ANF volumes, the application volume group logic places data and log volumes of a HANA system on different ANF resources within the ANF bare-metal fleet, resulting in separate storage endpoint for those volumes. These storage endpoints are configured with individual IP addresses to ensure direct access to the data with optimal latency. Using different IP addresses for the HANA data and log volume is also crucial for the NFS client performance on the Linux host; this way the Linux NFS client manages an individual IO queue for each individual IP address and therefor volume. Having a different queue for data and log provides best latency and performance on the Linux host.
Lastly the application volume group workflow also optionally provisions log and data backup volumes. These volumes are then provisioned separated of the data and log volume, outside of the proximity placement group.
Below is an example of a volume provisioning workflow using application volume group. The picture shows the workflow for an SAP HANA single host system.
First, the user must provide a couple of input values, that describe the HANA system as well as the ANF environment. In addition, the proximity placement group that will be used must be provided. Then, based on the RAM size of the HANA system, capacity and throughput numbers are calculated and proposed for the volumes to be provisioned.
In total 5 volumes will be provisioned for a HANA single host system:
- an SAP HANA data, log and shared will be provisioned within the provided proximity placement group,
- a naming convention including the SID of the HANA system will be used for the volume names,
- data and log backup volumes will be provisioned outside of the proximity placement group, since they are not latency sensitive.
Three different storage endpoints with individual IP addresses will be provisioned for the volumes.
As mentioned, Azure NetApp Files application volume group for SAP HANA can be used to provision the ANF volumes for the most common HANA system layouts. This includes HANA single host as well as multiple host systems. For single and multiple-host systems, secondary system volumes can be provisioned either in a HANA system replication (HSR) setup or for disaster recovery using ANF cross-region replication (CRR), as shown in the following image.
A series of short videos explains the application volume group feature and the usage for the different SAP HANA deployments. Depending on the HANA system type and the configuration of high availability with HANA system replication or disaster recover with ANF cross-region replication, the AVG deployment process is either a single step or a multiple steps configuration.
Any configuration always starts with the volume provisioning for a HANA single host system. This single host system can then be extended in a second step to a multiple host system, a HANA system replication setup or a DR solution using ANF cross region replication.
Additional configuration steps need to be done if a multiple host system should be configured with HSR or ANF cross-region replication. For all these configuration options, a demo video which shows and explains the workflow using AVG is available below.
The video in this section explains the technical background of the provisioning workflow and prerequisites for the application volume group feature.
The video in this section explains the required preparations and the volume provisioning workflow for an SAP HANA single host system using the Azure NetApp Files application volume group feature.
The video in this section explains the provisioning workflow for an SAP HANA multiple host system using Azure NetApp Files application volume group.
The video in this section explains the volume provisioning workflow for an SAP HANA system with HANA System Replication using the Azure NetApp Files application volume group feature.
Videos for a disaster recovery configuration with ANF cross region replication and for multiple steps configurations will be made available as well. This blog will then be updated with the new content.
Application volume group (AVG) for SAP HANA enables you to deploy all volumes required to install and operate an SAP HANA database according to best practices in a single one-step and optimized workflow. The application volume group feature includes the use of proximity placement group (PPG) with VMs to achieve automated, low-latency deployments. Application volume group for SAP HANA has implemented many technical improvements that simplify and standardize the entire process to help you streamline volume deployments for SAP HANA - instead of creating the SAP HANA volumes (data, log, shared, log-backup, file-backup) individually, the new application volume group for SAP HANA creates these volumes in a single 'atomic' operation (GUI, RP, API).
Azure NetApp Files application volume group will shorten SAP HANA landscape deployment time and increase overall application performance and stability, including the use of multiple storage endpoints. The application volume group feature supports both single-node (scale-up) and multi-node (scale-out) standardized and optimized HANA deployments. The application volume group feature also proposes optimized sizing, standard naming conventions, and includes support for both HANA system replication (HSR) for high availability and Azure NetApp Files cross region replication (CRR) for regional disaster recovery with storage-based replication.
This article provides guidance and videos on how to best leverage the Azure NetApp Files application volume group (AVG) feature in the most common SAP HANA deployment scenarios.