Редактиране

Споделяне чрез

Considerations for Azure Virtual Machines DBMS deployment for SAP workload

  • Статия

This guide is part of the documentation on how to implement and deploy SAP software on Microsoft Azure. Before you read this guide, read the Planning and implementation guide and articles the planning guide points you to. This document covers the generic deployment aspects of SAP-related DBMS systems on Microsoft Azure virtual machines (VMs) by using the Azure infrastructure as a service (IaaS) capabilities.

The paper complements the SAP installation documentation and SAP Notes, which represent the primary resources for installations and deployments of SAP software on given platforms.

In this document, considerations of running SAP-related DBMS systems in Azure VMs are introduced. There are few references to specific DBMS systems in this document. Instead, the specific DBMS systems are handled in other database system specific documents.

Resources

There are other articles available on SAP workload on Azure. Start with SAP workload on Azure: Get started and then choose your area of interest.

The following SAP Notes are related to SAP on Azure in regard to the area covered in this document.

Note number Title
1928533 SAP applications on Azure: Supported products and Azure VM types
2015553 SAP on Microsoft Azure: Support prerequisites
1999351 Troubleshooting enhanced Azure monitoring for SAP
2178632 Key monitoring metrics for SAP on Microsoft Azure
1409604 Virtualization on Windows: Enhanced monitoring
2191498 SAP on Linux with Azure: Enhanced monitoring
2039619 SAP applications on Microsoft Azure using the Oracle database: Supported products and versions
2233094 DB6: SAP applications on Azure using IBM DB2 for Linux, UNIX, and Windows: Additional information
2243692 Linux on Microsoft Azure (IaaS) VM: SAP license issues
2578899 SUSE Linux Enterprise Server 15: Installation Note
1984787 SUSE LINUX Enterprise Server 12: Installation notes
2772999 Red Hat Enterprise Linux 8.x: Installation and Configuration
2002167 Red Hat Enterprise Linux 7.x: Installation and upgrade
2069760 Oracle Linux 7.x SAP installation and upgrade
1597355 Swap-space recommendation for Linux
2799900 Central Technical Note for Oracle Database 19c
2171857 Oracle Database 12c: File system support on Linux
1114181 Oracle Database 11g: File system support on Linux
2969063 Microcode Validation Failed in HCMT on Azure
3246210 Azure - HCMT Fails During Some Disk Performance Tests

For information on all the SAP Notes for Linux, see the SAP community wiki.

You need a working knowledge of Microsoft Azure architecture and how Microsoft Azure virtual machines are deployed and operated. For more information, see Azure documentation.

In general, the Windows, Linux, and DBMS installation and configuration are essentially the same as any virtual machine or bare metal machine you install on-premises. There are some architecture and system management implementation decisions that are different when you use Azure IaaS. This document explains the specific architectural and system management differences to be prepared for when you use Azure IaaS.

Storage structure of a VM for RDBMS deployments

To follow this chapter, read and understand the information presented in:

For Azure block storage, the usage of Azure managed disks is mandatory. For details about Azure managed disks read the article Introduction to managed disks for Azure VMs.

In a basic configuration, we usually recommend a deployment structure where the operating system, DBMS, and eventual SAP binaries are separate from the database files. We recommend having separate Azure disks for:

  • The operating system (base VHD or OS VHD)
  • Database management system executables
  • SAP executables like /usr/sap
  • DBMS data files
  • DBMS redo log files

A configuration that separates these components into five different volumes can result in higher resiliency since excessive usage on one volume doesn't necessarily interfere with the usage of other volumes as long as VM storage quota and limits aren't exceeded.

The DBMS data and transaction/redo log files are stored in Azure supported block storage or Azure NetApp Files. Azure Files or Azure Premium Files isn't supported as storage for DBSM data and/or redo log files with SAP workload. They're stored in separate disks and attached as logical disks to the original Azure operating system image VM. For Linux deployments, different recommendations are documented. Read the article Azure Storage types for SAP workload for the capabilities and the support of the different storage types for your scenario. Specifically for SAP HANA start with the article SAP HANA Azure virtual machine storage configurations.

When you plan your disk layout, find the best balance between these items:

  • The number of data files.
  • The number of disks that contain the files.
  • The IOPS quotas of a single disk or NFS share.
  • The data throughput per disk or NFS share.
  • The number of additional data disks possible per VM size.
  • The overall storage or network throughput a VM can provide.
  • The latency different Azure Storage types can provide.
  • VM storage IOPS and throughput quota.
  • VM network quota in case you're using NFS - traffic to NFS shares is counting against the VM's network quota and NOT the storage quota.
  • VM SLAs.

Azure enforces an IOPS quota per data disk or NFS share. These quotas are different for disks hosted on the different Azure block storage solutions or shares. I/O latency is also different between these different storage types as well.

Each of the different VM types has a limited number of data disks that you can attach. Another restriction is that only certain VM types can use, for example, premium storage. Typically, you decide to use a certain VM type based on CPU and memory requirements. You also need to consider the IOPS, latency, and disk throughput requirements that usually are scaled with the number of disks or the type of premium storage disks v1. The number of IOPS and the throughput to be achieved by each disk might dictate disk size, especially with premium storage v1. With premium storage v2 or Ultra disk, you can select provisioned IOPS and throughput independent of the disk capacity.

Note

For DBMS deployments, we highly recommend Azure premium storage (v1 and v2), Ultra disk or Azure NetApp Files based NFS shares for any data, transaction log, or redo files. It doesn't matter whether you want to deploy production or nonproduction systems. Latency of Azure standard HDD or SSD isn't acceptable for any type of production system.

Note

To maximize Azure's single VM SLA, all disks that are attached must be Azure premium storage (v1 or v2) or Azure Ultra disk type, which includes the base VHD (Azure premium storage).

Note

Hosting main database files, such as data and log files, of SAP databases on storage hardware that's located in co-located third-party data centers adjacent to Azure data centers isn't supported. Storage provided through software appliances hosted in Azure VMs, are also not supported for this use case. For SAP DBMS workloads, only storage that's represented as native Azure service is supported for the data and transaction log files of SAP databases in general. Different DBMS might support different Azure storage types. For more details check the article Azure Storage types for SAP workload

The placement of the database files and the log and redo files and the type of Azure Storage you use, is defined by IOPS, latency, and throughput requirements. Specifically for Azure premium storage v1, to achieve enough IOPS, you might be forced to use multiple disks or use a larger premium storage disk. If you use multiple disks, build a software stripe across the disks that contain the data files or the log and redo files. In such cases, the IOPS and the disk throughput SLAs of the underlying premium storage disks or the maximum achievable IOPS of standard storage disks are accumulative for the resulting stripe set.

If your IOPS requirement exceeds what a single VHD can provide, balance the IOPS that is needed for the database files across a number of VHDs. The easiest way to distribute the IOPS load across disks is to build a software stripe over the different disks. Then place a number of data files of the SAP DBMS on the LUNs carved out of the software stripe. The number of disks in the stripe is driven by IOPS demands, disk throughput demands, and volume demands.

Windows storage striping Windows

We recommend that you use Windows Storage Spaces to create stripe sets across multiple Azure VHDs. Use at least Windows Server 2012 R2 or Windows Server 2016.

Linux storage striping Linux

Only MDADM and Logical Volume Manager (LVM) are supported to build a software RAID on Linux. For more information, see:

For Azure premium storage v2 and Ultra disk, striping may not necessary since you can define IOPS and disk throughput independent of the size of the disk.

Note

Because Azure Storage keeps three images of the VHDs, it doesn't make sense to configure a redundancy when you stripe. You only need to configure striping so that the I/Os are distributed over the different VHDs.

Managed or nonmanaged disks

An Azure storage account is an administrative construct and also a subject of limitations. For information on capabilities and limitations, see Azure Storage scalability and performance targets. For standard storage, remember that there's a limit on the IOPS per storage account. See the row that contains Total Request Rate in the article Azure Storage scalability and performance targets. There's also an initial limit on the number of storage accounts per Azure subscription. As of 2017, Azure introduced the concepts of Azure Managed Disks that relief you of taking care of any storage account administration. Using Azure managed disks is the default to deploy for SAP workload in Azure.

Important

Given the advantages of Azure Managed Disks, it is mandatory that you use Azure Managed Disks for your DBMS deployments and SAP deployments in general.

If you happen to have SAP workload that isn't yet using managed disks, to convert from unmanaged to managed disks, see:

Caching for VMs and data disks

When you mount disks to VMs, you can choose whether the I/O traffic between the VM and those disks located in Azure storage is cached.

The following recommendations assume these I/O characteristics for standard DBMS:

  • It's mostly a read workload against data files of a database. These reads are performance critical for the DBMS system.
  • Writing against the data files occurs in bursts based on checkpoints or a constant stream. Averaged over a day, there are fewer writes than reads. Opposite to reads from data files, these writes are asynchronous and don't hold up any user transactions.
  • There are hardly any reads from the transaction log or redo files. Exceptions are large I/Os when you perform transaction log backups.
  • The main load against transaction or redo log files is writes. Dependent on the nature of the workload, you can have I/Os as small as 4 KB or, in other cases, I/O sizes of 1 MB or more.
  • All writes must be persisted on disk in a reliable fashion.

For Azure premium storage v1, the following caching options exist:

  • None
  • Read
  • Read/write
  • None + Write Accelerator, which is only for Azure M-Series VMs
  • Read + Write Accelerator, which is only for Azure M-Series VMs

For premium storage v1, we recommend that you use Read caching for data files of the SAP database and choose No caching for the disks of log file(s).

Note

With some of the new M(b)v3 VM types, the usage of read cached Premium SSD v1 storage could result in lower read and write IOPS rates and throughput than you would get if you don't use read cache.

For M-Series deployments, we recommend that you use Azure Write Accelerator only for the disks of your log files. For details, restrictions, and deployment of Azure Write Accelerator, see Enable Write Accelerator.

For premium storage v2, Ultra disk and Azure NetApp Files, no caching options are offered.

Azure nonpersistent disks

Azure VMs offer nonpersistent disks after a VM is deployed. If a VM reboots, all content on those drives can be wiped out. It's a given that data files and log and redo files of databases should under no circumstances be located on those nonpersisted drives. There might be exceptions for some databases, where these nonpersisted drives could be suitable for tempdb and temp tablespaces.

For more information, see Understand the temporary drive on Windows VMs in Azure.

Windows nonpersisted disk Windows

Drive D in an Azure VM is a nonpersisted drive, which is backed by some local disks on the Azure compute node. Because it's nonpersisted, any changes made to the content on drive D are lost when the VM is rebooted. Changes include files that were stored, directories that were created, and applications that were installed.

Linuxnonpersisted disk Linux

Linux Azure VMs automatically mount a drive at /mnt/resource that's a nonpersisted drive backed by local disks on the Azure compute node. Because it's nonpersisted, any changes made to content in /mnt/resource are lost when the VM is rebooted. Changes include files that were stored, directories that were created, and applications that were installed.

Microsoft Azure Storage resiliency

Microsoft Azure Storage stores the base VHD, with OS and attached disks or blobs, on at least three separate storage nodes. This type of storage is called locally redundant storage (LRS). LRS is the default for all types of storage in Azure.

There are other redundancy methods. For more information, see Azure Storage replication.

Note

Azure premium storage v1 and v2, Ultra disk and Azure NetApp Files are the recommended type of storage for DBMS VMs and disks that store database and log and redo files. With exception of premium storage v1, the only available redundancy method for these storage types is LRS. As a result, you need to configure database methods to enable database data replication into another Azure region or availability zone. Database methods include SQL Server Always On, Oracle Data Guard, and HANA System Replication.

VM node resiliency

Azure offers several different SLAs for VMs. For more information, see the most recent release of SLA for Virtual Machines. Because the DBMS layer is critical to availability in an SAP system, you need to understand different deployment types and maintenance events. For more information on these concepts, see Manage the availability of virtual machines in Azure.

The minimum recommendation for production DBMS scenarios with an SAP workload is to:

  • Deploy two VMs using the chosen deployment type in the same Azure region.
  • Run these two VMs in the same Azure virtual network and have NICs attached out of the same subnets.
  • Use database methods to keep a hot standby with the second VM. Methods can be SQL Server Always On, Oracle Data Guard, or HANA System Replication.

You also can deploy a third VM in another Azure region and use the same database methods to supply an asynchronous replica in another Azure region.

Azure network considerations

In large-scale SAP deployments, use the blueprint of Azure Virtual Datacenter. Use it for your virtual network configuration and permissions and role assignments to different parts of your organization.

These best practices are the result of thousands of customer deployments:

  • The virtual networks the SAP application is deployed into don't have access to the internet.
  • The database VMs run in the same virtual network as the application layer, separated in a different subnet from the SAP application layer.
  • The VMs within the virtual network have a static allocation of the private IP address. For more information, see IP address types and allocation methods in Azure.
  • Routing restrictions to and from the DBMS VMs aren't set with firewalls installed on the local DBMS VMs. Instead, traffic routing is defined with network security groups (NSGs).
  • To separate and isolate traffic to the DBMS VM, assign different NICs to the VM. Every NIC gets a different IP address, and every NIC is assigned to a different virtual network subnet. Every subnet has different NSG rules. The isolation or separation of network traffic is a measure for routing. It's not used to set quotas for network throughput.

Note

Assigning static IP addresses through Azure means to assign them to individual virtual NICs. Don't assign static IP addresses within the guest OS to a virtual NIC. Some Azure services like Azure Backup rely on the fact that at least the primary virtual NIC in the guest OS is set to DHCP and not to static IP addresses. For more information, see Troubleshoot Azure virtual machine backup. To assign multiple static IP addresses to a VM, assign multiple virtual NICs to a VM.

Warning

Configuring network virtual appliances in the communication path between the SAP application and the DBMS layer of a SAP NetWeaver-, Hybris-, or S/4HANA-based SAP system isn't supported. This restriction is for functionality and performance reasons. The communication path between the SAP application layer and the DBMS layer must be a direct one. The restriction doesn't include application security group (ASG) and NSG rules if those ASG and NSG rules allow a direct communication path. This also includes traffic to NFS shares that host DBMS data and redo log files.

Other scenarios where network virtual appliances aren't supported are in:

Network virtual appliances in communication paths can easily double the network latency between two communication partners. They also can restrict throughput in critical paths between the SAP application layer and the DBMS layer. In some customer scenarios, network virtual appliances can cause Pacemaker Linux clusters to fail. These are cases where communications between the Linux Pacemaker cluster nodes communicate to their SBD device through a network virtual appliance.

Important

Another design that's not supported is the segregation of the SAP application layer and the DBMS layer into different Azure virtual networks that aren't peered with each other. We recommend that you segregate the SAP application layer and DBMS layer by using subnets within an Azure virtual network instead of by using different Azure virtual networks.

If you decide not to follow the recommendation and instead segregate the two layers into different virtual networks, the two virtual networks must be peered.

Be aware that network traffic between two peered Azure virtual networks is subject to transfer costs. Huge data volume that consists of many terabytes is exchanged between the SAP application layer and the DBMS layer. You can accumulate substantial costs if the SAP application layer and DBMS layer are segregated between two peered Azure virtual networks.

Use Azure Load Balancer to redirect traffic

The use of private virtual IP addresses used in functionalities like SQL Server Always On or HANA System Replication requires the configuration of an Azure load balancer. The load balancer uses probe ports to determine the active DBMS node and route the traffic exclusively to that active database node.

If there's a failover of the database node, there's no need for the SAP application to reconfigure. Instead, the most common SAP application architectures reconnect against the private virtual IP address. Meanwhile, the load balancer reacts to the node failover by redirecting the traffic against the private virtual IP address to the second node.

Azure offers two different load balancer SKUs: a basic SKU and a standard SKU. Based on the advantages in setup and functionality, you should use the Standard SKU of the Azure load balancer. One of the large advantages of the Standard version of the load balancer is that the data traffic isn't routed through the load balancer itself.

An example how you can configure an internal load balancer can be found in the article Tutorial: Configure a SQL Server availability group on Azure Virtual Machines manually

Note

There are differences in behavior of the basic and standard SKU related to the access of public IP addresses. The way how to work around the restrictions of the Standard SKU to access public IP addresses is described in the document Public endpoint connectivity for Virtual Machines using Azure Standard Load Balancer in SAP high-availability scenarios

Deployment of host monitoring

For production use of SAP applications in Azure virtual machines, SAP requires the ability to get host monitoring data from the physical hosts that run the Azure virtual machines. A specific SAP Host Agent patch level is required that enables this capability in SAPOSCOL and SAP Host Agent. The exact patch level is documented in SAP Note 1409604.

For more information on the deployment of components that deliver host data to SAPOSCOL and SAP Host Agent and the life-cycle management of those components, start with the article Implement the Azure VM extension for SAP solutions.

Next steps

For more information on a particular DBMS, see: