Resource Partitioning on Hitachi Virtual Storage Platform Gx00 for SAP HANA Tailored Datacenter Integration Lab Validation Report By Kengo Miyazawa and Stephan Kreitz July 2016 Feedback Hitachi Data Systems welcomes your feedback. Please share your thoughts by sending an email message to [email protected]. To assist the routing of this message, use the paper number in the subject and the title of this white paper in the text. Contents Introduction .......................................................................................................................... 2 Hitachi Virtual Storage Platform Gx00 Models ............................................................. 3 SAP HANA Certified Enterprise Storage Gx00 Models .................................................................. 3 Hitachi Storage Partitioning ............................................................................................................ 4 Hitachi Dynamic Provisioning ......................................................................................................... 6 Hitachi Compute Blade Server Deployment Models ...................................................... 7 Hitachi Compute Blade 2500 .......................................................................................................... 7 Virtualized SAP HANA Deployments............................................................................................... 7 SAP HANA Tailored Datacenter Integration .................................................................... 9 Storage Setup for SAP HANA TDI .................................................................................................. 9 Hitachi Storage Partitioning Use Cases........................................................................... 11 Separation of Production and Non-production SAP HANA Nodes .............................................. 11 Separation of Production and Non-production SAP HANA Nodes, and SAP Application Servers ........................................................................................................ 13 Engineering Validation ....................................................................................................... 15 Hardware Setup ............................................................................................................................ 15 Test Setup..................................................................................................................................... 18 Test Summary ............................................................................................................................... 20 Conclusion........................................................................................................................... 21 1 Resource Partitioning on Hitachi Virtual Storage Platform Gx00 for SAP HANA Tailored Datacenter Integration Lab Validation Report Throughout the enterprise, different end users seek access to different data sets, depending on their department, role, function, and so on. As concerns over data privacy and compliance issues have grown, organizations have addressed them by deploying multiple storage systems to physically separate storage resources. While multiple storage systems may address data privacy concerns, purchasing and managing these separate storage systems is expensive and inefficient. Purchasing separate storage increases capital expenditures (CAPEX), and operating expenses (OPEX) of disparate storage systems takes up valuable rack space and increases energy costs. As storage needs grow, each storage system must be grown separately, limiting enterprise-wide scale-up and scale-out options, and adding expense. As storage systems are added, administration costs rise as well, because separate storage systems often have their own unique administration tools, user lists, and authentication and access control mechanisms. As a result, with so many different systems to manage, using multiple storage systems creates more opportunities for security to be compromised. Rather than using multiple storage solutions for physical separation of data, partitioning allows for multiple tenants to safely coexist on a single storage system. Partitioning allows multitenancy without the risk of activities in one storage region affecting performance, while allowing for availability or privacy in others. A storage system that supports resource partitioning provides many benefits: Multitenancy protects privacy and compliance without requiring separate storage systems. Efficient storage scalability gives the ability to scale-up to meet business needs throughout the enterprise. Integration with enterprise authentication and access control systems such as Microsoft® Active Directory® and LDAP (Lightweight Directory Access Protocol) reduce administration overhead. Unified management of a single storage system provides better security. Improved datacenter efficiency eliminates unnecessary capex and operating expenditures associated with purchasing and operating several different storage systems. In a modern SAP landscape, SAP HANA builds the foundation for SAP applications as a high performance in-memory database or even as application platform. Business critical applications like SAP HANA require data security as well as performance guarantees that are provided by using Hitachi enterprise storage for SAP HANA and Hitachi Compute Blade server technology for running full SAP landscapes. 1 2 Introduction The purpose of this document is to provide help and guidance to customers who want to achieve maximum separation of their different workloads on their Hitachi Virtual Storage Platform for their SAP HANA landscape in a SAP HANA Tailored Datacenter Integration (TDI) environment. Hitachi storage allows organizations to use partitioning to enable multitenancy in the storage environment in a manner that protects data for applications, storage administrators, applications, and business units. Unified management eases the administrative burden, and by consolidating multiple storage systems, you reduce CAPEX and OPEX while improving security and manageability. This document is applicable for the following SAP HANA certified enterprise storage models from the Hitachi Virtual Storage Platform family: Hitachi Virtual Storage Platform G200 (VSP G200) Hitachi Virtual Storage Platform G400 (VSP G400) Hitachi Virtual Storage Platform G600 (VSP G600) Hitachi Virtual Storage Platform G800 (VSP G800) In the case of SAP HANA scale-out deployments, two Hitachi NAS Platform 4060 (HNAS) servers can be used to provide a highly available NFS-based shared file system for SAP HANA configuration files, binaries, and trace files. Using the Virtual Storage Platform family of enterprise storage products with SAP HANA has the following benefits: Increased performance for SAP HANA data load Scalable deployments of SAP HANA Disaster recovery with minimal performance impact to the production instance Workload separation using Hitachi storage partitioning features Consolidation using Hitachi Dynamic Provisioning (HDP) pools This technical paper assumes you have familiarity with the following: Storage area network (SAN) based storage systems General storage concepts General network knowledge SAP HANA Platform SAP HANA TDI Common IT storage practices Note — Testing of this configuration was performed in a lab environment. Many things affect production environments beyond prediction or duplication in a lab environment. Follow the recommended practice of conducting proof-of-concept testing for acceptable results in a non-production, isolated test environment that matches your production environment before your production implementation of this solution. 2 3 Hitachi Virtual Storage Platform Gx00 Models Hitachi Virtual Storage Platform Gx00 models are based on industry-leading enterprise storage technology. With flashoptimized performance, these systems provide advanced capabilities previously available only in high-end storage arrays. With the Virtual Storage Platform Gx00 models, you can build a high performance, software-defined infrastructure to transform data into valuable information. Hitachi Storage Virtualization Operating System provides storage virtualization, high availability, superior performance, and advanced data protection for all Virtual Storage Platform Gx00 models. This proven, mature software provides common features to consolidate assets, reclaim space, extend life, and reduce migration effort. New management software improves ease of use to save time and reduce complexity. The infrastructure of Storage Virtualization Operating System creates a management framework for improved IT response to business demands. SAP HANA Certified Enterprise Storage Gx00 Models VSP G200 VSP G200 entry level storage system includes advanced SVOS capabilities in this affordable platform. It offers a choice of internal capacity expansion of up to 12 or 24 drives, the ability to add more external storage trays, and easy setup and management for virtual applications. Advanced SVOS virtualization capabilities protect business-critical data. VSP G400 VSP G400 midrange platform helps reduce operational costs with an efficient unified architecture. Consolidate file, block, and object data for extensive cost savings. Simplify operations for all data types with one interface and consistent workflows. Reduce storage expense by reclaiming space and increasing utilization. Midsized organizations can choose this model to be their primary storage system. VSP G600 VSP G600 extends the performance and capacity scalability of VSP G400 for organizations powering additional virtualized applications. Deploy more storage solutions with this validated application integration. Manage more data types from a simple, unified interface. Deliver central, consistent storage services to all data consumers. Accelerate multiple applications with this flash optimized architecture available tiered or all flash. VSP G800 VSP G800 entry enterprise platform is designed to consolidate the business-critical data from multiple applications and virtual servers. Implement dual datacenter replication for high availability business continuation. Securely partition information assets for multiple tenants. Utilize crash-consistent snapshots for application-aware backup, recovery and failover. Designed for organizations seeking enterprise functionality for regional deployments. VSP G1000 VSP G1000 enterprise platform enables the continuous operations, self-managing policy-driven management, and agile IT that today’s new breed of cloud applications demand. Global storage virtualization enables an always-on infrastructure with enterprise-wide scalability. An ideal solution for applications that require zero recovery point and recovery time objectives, VSP G1000 redefines mission-critical storage virtualization and resets expectations for the datacenter. Out of the five members of the VSP family (VSP G1000, VSP G800, VSP G600, VSP G400, and VSP G200), validation of this reference architecture was done with SAP HANA Platform nodes on VSP G600, VSP G400, and VSP G200. These three storage platforms have similar high availability features and run on the same family microcode as VSP G1000 (VSP G1000 was released and certified by SAP for use as enterprise storage for SAP HANA before the VSP midrange models). The VSP G1000 hardware architecture is more resilient than the VSP midrange models. 3 4 Figure 1 shows VSP storage arrays in the order of increasing performance and storage capacity. Figure 1 Hitachi Storage Partitioning Hitachi Virtual Storage Platform Gx00 storage systems can connect to multiple hosts and can be shared by multiple users, divisions in a company, or by multiple companies, which can result in conflicts among users. For example, if a host issues many I/O requests or reads or writes a large amount of data, the I/O performance of other hosts may be affected. Even more, many storage administrators from different organizations can access the storage system. Managing the entire storage system can become complex and difficult. Private data might be accessed by other users, or a volume in one organization might be destroyed by mistake by a storage administrator in another organization. To avoid such problems, use Hitachi Resource Partition Manager software to set up resource groups that allow you to manage one storage system as if it were multiple virtual private storage systems. The storage administrators in each resource group can access only their assigned resources and cannot access other resources. Configuring resource groups prevents the risk of data leakage or data destruction by another storage administrator in another resource group. Resources such as logical devices (LDEVs), parity groups, external volumes, ports, or host groups can be assigned to a resource group. These resources can be combined to flexibly compose a virtual private storage system. 4 5 Figure 2 shows an overview of multiple hosts connected to a single VSP without any partitioning, sharing all resources of multi-processing units (MPU) and the cache. Figure 2 Using resource groups to partition the resources into up to four separate partitions, it is possible to ensure performance and isolation of different hosts or host groups as shown in Figure 3. Figure 3 5 6 Hitachi Dynamic Provisioning On Hitachi storage systems, Hitachi Dynamic Provisioning provides wide striping and thin provisioning functionalities. Using Dynamic Provisioning is like using a host-based logical volume manager (LVM), but without incurring host processing overhead. It provides one or more wide-striping pools across many RAID groups. Each pool has one or more dynamic provisioning virtual volumes (DP-VOLs) of a logical size you specify (up to 60 TB created against it without allocating any physical space initially). Deploying Dynamic Provisioning avoids the routine issue of hot spots that occur on logical devices (LDEVs). These occur within individual RAID groups when the host workload exceeds the IOPS or throughput capacity of that RAID group. Dynamic Provisioning distributes the host workload across many RAID groups, which provides a smoothing effect that dramatically reduces hot spots. Consistent Performance at a Lower Cost Dynamic provisioning enables you to maintain optimal QoS for your SAP deployment while still lowering overall storage costs. In addition, it transparently spreads the many individual I/O workloads typical of a SAP deployment across multiple physical disks. Such I/O workload balancing eliminates I/O bottlenecks across multiple applications while automatically tuning performance. The bottom line: better overall performance and more predictable performance at a lower overall cost. When to use Hitachi Dynamic Provisioning Dynamic Provisioning is a best fit in an open-systems environment in the following scenarios: When the aggregation of storage pool capacity usage across many volumes provides the best opportunity for performance optimization. For stable environments and large consistently growing files or volumes. When device addressing constraints are a concern. 6 7 Hitachi Compute Blade Server Deployment Models In combination with Hitachi Compute Blade server technology, the Hitachi VSP family offers a strong hardware setup for running SAP landscapes. Hitachi Compute Blade 2500 Hitachi Compute Blade 2500 delivers enterprise computing power and performance with unprecedented scalability and configuration flexibility. Lower your costs and protect your investment. Flexible I/O architecture and logical partitioning allow configurations to match application needs exactly with Hitachi Compute Blade 2500. Multiple applications easily and securely co-exist in the same chassis. Add server management and system monitoring at no cost with Hitachi Compute Systems Manager. Seamlessly integrate with Hitachi Command Suite in Hitachi storage environments. Hitachi Compute Blade 520X server blades provide a high performance server platform for running SAP HANA. Using symmetric multi-processing (SMP) technology to combine multiple server blade resources into a single server, configurations from 2-socket servers up to 8-socket servers are available. 520H server blades are available for entry level SAP HANA systems in TDI environments. For SAP HANA appliances as well as for TDI configurations, pre-configured and tested configurations are available from Hitachi Data Systems running SAP HANA on bare metal and virtualized environments. For SAP Application servers and the various other components of a SAP landscape, the full range of Compute Blade server blades are available to fit the special needs of the specific applications. Virtualized SAP HANA Deployments In addition to bare metal configurations and solutions, Hitachi Data Systems offers pre-configured solutions for virtualized HANA systems based on either Hitachi logical partitioning technology or the VMware ESXi server. Hitachi Unified Compute Platform for the SAP HANA Platform with Logical Partitioning The logical partitioning feature from Hitachi partitions the physical resources of one single-blade or multi-blade server logically into independent server environments called LPARs (logical partitions). CPUs, memory, and I/O devices can be assigned in dedicated mode to LPARs to create fully isolated server environments for SAP HANA single node installations within each LPAR running on the same physical host without any noisy neighbor effects between the different server environments. 520X B2 server blades must be set to LP mode in order to enable logical partitioning capabilities. LPARs using one, two or four CPU sockets in dedicated mode, accessing the CPU sockets’ local memory bank, can be created for running production SAP HANA instances. Combined with dedicated HBA cards for each LPAR, and separated controller ports and RAID groups in Hitachi Virtual Storage Platform, the LPAR manager enables the consolidation of multiple production SAP HANA instances on the same physical server using a single Virtual Storage Platform. LPAR in combination with VSP resource groups offers the capability to provide end-to-end partitioning of server and storage resources creating fully separated environments across the whole hardware stack. Unified Compute Platform for SAP HANA Running on VMware vSphere VMware vSphere is a virtualization platform that provides a datacenter infrastructure. It features vSphere Distributed Resource Scheduler (DRS), high availability, and fault tolerance. 7 8 VMware vSphere has the following components: ESXi A hypervisor that loads directly onto a physical server. It partitions one physical machine into many virtual machines that share hardware resources. vCenter Server vCenter Server can manage the vSphere environment through a single user interface. With vCenter, additional features are available such as vMotion, Storage vMotion, Storage Distributed Resource Scheduler, High Availability, and Fault Tolerance. VMware vSphere 5.5 An enterprise virtualization solution to create a dynamic and flexible datacenter with integrated management and reporting capability for a high level of server, service, and client uptime. The Unified Compute Platform for the SAP HANA Platform running on VMware vSphere solution combines the benefits of the UCP for SAP HANA appliance with the flexibility and manageability of the VMware vSphere solution. 8 9 SAP HANA Tailored Datacenter Integration Unlike a SAP HANA appliance in which all hardware components are pre-configured by the hardware vendor, SAP HANA tailored datacenter integration (TDI) deployments are customized solutions where the customer can choose any of the certified SAP HANA server vendors along with any certified SAP HANA enterprise storage to implement SAP HANA. This provides customers an opportunity to leverage their existing hardware and reduce TCO. Using this reference architecture, SAP HANA solutions can be deployed using any certified SAP HANA server vendors and Hitachi Virtual Storage Platform Gx00. A list of SAP certified servers that are available for SAP HANA appliances can be found in the SAP Certified and Supported SAP HANA® Hardware Directory. SAP only allows using homogeneous compute server hardware from a single hardware partner in a SAP HANA tailored datacenter integration. Also, if a certification provided by SAP is for a specific operating system, only that operating system can be used for SAP HANA implementation. Engineering validation for this solution has been performed using Hitachi Data Systems server blades. Every SAP certified enterprise storage platform must meet TDI storage KPI requirements set by SAP. Testing showed that the storage design of Hitachi Virtual Storage Platform for the SAP HANA Platform meets the TDI requirements from SAP up to a maximum number of nodes as listed in the SAP Certified and Supported SAP HANA® Hardware Directory. It is not mandatory for customers to use the same storage design as was used for storage KPI testing, and this is demonstrated in this reference architecture guide. Refer to the SAP HANA Tailored Datacenter Integration FAQ for more details about TDI. During validation, the scalability and storage KPI testing was performed using the SAP HWCCT tool (Please refer to SAP Note 1943937 - Hardware Configuration Check Tool - Central Note). Note - Since the release of SAP HANA TDI in November 2013 several versions of the HWCCT have been published. To check whether or not the hardware configuration of your SAP HANA TDI infrastructure meets KPIs from SAP, it is crucial that you use the same version of the HWCCT used during the certification of the hardware (compute servers and storage system) for your tests. SAP Note 1943937 describes how to determine the right version of the HWCCT for your tests. Storage Setup for SAP HANA TDI SAP HANA appliance configurations often use dedicated disks and RAID groups for the SAP HANA data and log volumes as well as for the operating system and the HANA shared volume. This separation of disks grants isolation of the volumes at the cost of flexibility and efficient usage of the available disk space, often leaving some space on the RAID groups fully unused. In SAP HANA TDI environments, Hitachi enterprise storage for SAP HANA provides much more flexibility, efficiency, and security by offering the support of a range of storage features and products that are not available for all SAP HANA appliances. Intelligent provisioning of disk space is required to ensure that all parts of an SAP landscape have exactly the amount of space they need to function correctly. Hitachi Dynamic Provisioning enables you to make sure the right amount and the right type of storage is online to maximize performance and efficiency in your environment. To further enhance performance, Hitachi Dynamic Tiering guarantees that critical production data is available on the fastest and nearest storage tier. It stores less frequently accessed data on more cost-effective storage media. Hitachi TrueCopy (synchronous) or Hitachi Universal Replicator (asynchronous) software ensure that all data is protected both on-site and off-site, and provide disaster recovery for SAP environments. 9 10 SAP HANA TDI setups, Hitachi Dynamic Provisioning (HDP) pools and the Hitachi VSP enterprise storage family For provide the flexibility to size volumes for multiple SAP HANA nodes and an efficient use of the available disk space. With reduced disk requirements HDP pools grant disk performance for a good number of production SAP HANA systems at lower cost following some basic best practices: Create an HDP pool with a minimum of two parity groups whenever possible A parity group should be dedicated to one Pool only, and should not be used for other purposes if one of its logical devices is a Pool Volume HDP pools should be configured as RAID-6 Distribute the parity groups across at least two disk trays Create four volumes (VVOLs) for each HANA log volume and distribute them across the various multi-processing units (MPUs) Create four volumes for each HANA data volume and distribute them across the various MPUs Use full allocation to provision volumes whenever possible SAP HANA Tailored Datacenter Integration on Hitachi Virtual Storage Platform Gx00 using Hitachi Dynamic Provisioning Pools provides a reference architecture for SAP HANA nodes and the VSP Gx00 family. For non-production SAP HANA systems, HDP pools allow maximum allocation of the available disk space for consolidation of many QA, test, and development systems of different sizes. Since there are no performance requirements for non-production systems, Hitachi Dynamic Provisioning provides the flexibility a test and development landscape requires, often in combination with virtualization solutions at the server level such as Hitachi logical partitioning technology. 10 11 Hitachi Storage Partitioning Use Cases Hitachi storage partitioning can be applied to many different scenarios separating workloads, data, and even drive types and configurations. In this paper two use cases that have also been verified in a lab environment will be explained in more detail. The focus of these use cases is SAP HANA production systems. For the test setup described in the next section, a VSP G400 was used. Because of this, the overview and numbers in this section are based on VSP G400 as well. Increasing the number of HANA nodes on single partitions apply for the larger members of the VSP Gx00 family. Separation of Production and Non-production SAP HANA Nodes This use case describes the consolidation of multiple production SAP HANA nodes and a larger number of nonproduction SAP HANA nodes on a single storage array. Figure 4 shows the setup without any storage partitioning configured. Even with separated RAID groups or HDP pools, the workload of the QA and development systems will affect the performance of the production HANA nodes. Figure 4 11 12 Figure 5 shows the same setup after configuring storage partitioning and thus providing full isolation of the production HANA systems. With this isolation, enough resources are available for the production nodes, yet it is still possible to operate a larger number of QA and development systems on the second partition. Figure 5 Running QA and development systems within a separate partition provides the flexibility required for such systems by having the benefit and efficiency of HDP pools on the storage side as well as Hitachi logical partitioning or VMware server virtualization. This is possible while still dedicating the bigger part of the storage resources to production SAP HANA systems running on bare metal Hitachi Compute Blade 2500 or even in a virtualized server environment according to best practices for virtualized SAP HANA systems. Tests have shown that also for production systems HDP pools allow consolidation at the storage level reducing cost and disk overhead while still granting high performance to fulfill all requirements of production SAP HANA systems in a TDI environment. 12 13 Separation of Production and Non-production SAP HANA Nodes, and SAP Application Servers This use case adds a third partition to the scenario providing a partition for SAP Application Servers and other SAP chassis components in a SAP landscape. Figure 6 shows the overview without storage partitioning. Figure 6 13 14 workload of the QA and development systems will have an impact not only on the performance of the SAP HANA The systems but also on the SAP Application Servers connecting to these HANA systems. Configuring storage partitioning, as shown in Figure 7, will remove the impact between the different application workloads of SAP HANA, classic SAP applications, and QA and development systems containing both SAP HANA and non- SAP HANA systems. Figure 7 Tests with both dedicated RAID groups and HDP pools for the production SAP HANA and SAP Application servers are described in the next section. 14 15 Engineering Validation The main focus of the validation testing in the Hitachi Data Systems labs was on the production SAP HANA nodes fulfilling TDI KPIs according to the SAP HANA Enterprise Storage certification. HWCCT revision 97.3 was used for the scalability testing on VSP G400 because the same version was originally used for the Enterprise Storage certification with the HANAHWC-ES 1.0 scenario. The cache write pending and MPU utilization were also considered, and the cutoff was set at a maximum of 60% usage for storage best practices. Testing without storage partitioning shows the impact of the additional systems connected to the VSP G400 on the performance of the production HANA nodes. Configuring storage partitioning allows the production nodes to achieve KPIs without any impact from the additional system. Hardware Setup The use cases described in the last section have been tested with two different disk layouts for the first partition with the SAP HANA production nodes. Disk layout 1 uses dedicated RAID groups for each SAP HANA node, and disk layout 2 uses two HDP pools for all SAP HANA nodes The first HDP pool provides volumes for the operating system, the HANA shared volumes, and the HANA data volumes. The second HDP pool provides the HANA log volumes. The detailed disk layout for each test case is found in the following sections. Table 1 shows the hardware configuration for the test setup. Table 1. Hardware Configuration Hardware Server Test Case 1 Test Case 2 Test Case 3 Test Case 4 Server Chassis CB 2500 1 1 1 1 Blade 520X B2 4 4 4 4 CPU E7-8880v3 Haswell-EX 8 8 8 8 Memory 128 GB memory 4 LPAR (HANA 4 LPAR (HANA 3 LPAR (HANA Prod) Prod) Prod) 1 LPAR (AppServer) 3 LPAR (HANA Prod) 1 LPAR (AppServer) 64 GB memory 6 LPAR (QA/Dev) 6 LPAR (QA/Dev) 6 LPAR (QA/Dev) 6 LPAR (QA/Dev) 40 GB memory 6 LPAR (QA/Dev) 6 LPAR (QA/Dev) 6 LPAR (QA/Dev) 6 LPAR (QA/Dev) 8 8 8 HBA Fibre Channel 8 16 Gb/sec 2port 15 16 Table 1. Hardware Configuration (Continued) Hardware Storage Test Case 1 Test Case 2 Test Case 3 Test Case 4 Controller VSP G400 1 1 1 1 CHB Fibre Channel 8 16 Gb/sec 2port 8 8 8 Cache 8 GB DIMM 16 16 16 16 DKB 12 GB SAS 4 4 4 4 96 72 104 96 72 48 72 56 8 8 8 8 16 16 24 32 10 3 11 4 1 3 1 4 HDD Total HDDs 2.5inch 600 GB SAS(10krpm) 900 GB 1.2 TB RAID 6D+2P Configuration 14D+2P (RAID6) Figure 8 shows an overview of the Fibre Channel cabling. All 16 server HBA ports are directly connected to the VSP G400 storage ports without any SAN infrastructure. For further consolidation, SAN switches can be used; however, it has to be verified that all KPIs are fulfilled on the production nodes. Figure 8 This setup is the same throughout all the test configurations and has does not change in between the various test runs. 16 17 HANA QA and Development Systems SAP The partition for the QA and development HANA nodes always uses the same number of HANA systems and the same disk configuration. HDP pools are used for the operating system, HANA shared volumes, and HANA data volumes of all SAP HANA systems and for the HANA log volumes. Figure 9 shows an overview of the QA and development setup. Figure 9 Table 2 shows the detailed configuration of the QA and development SAP HANA server providing 12 HANA systems on a 4-socket server using Hitachi logical partitioning. Table 2. QA and Development Server Configuration Hardware Server Chassis CB 2500 1 Server Blade CB520X B2 2 CPU E7-8880v3 Haswell-EX 4 (2.30GHz 18 core) HBA Fibre Channel 16 G2port(FIVE-FX) 4 17 18 Table 3 shows the storage configuration for the QA and development LPARs. Table 3. QA and Development Partition Storage Configuration Hardware Configuration Number of QA/DEV HANA Systems 12 Memory capacity 6 × 64 GB and 6 × 40 GB RAID groups in HDP pools Pool 1: OS, HANA shared, and HANA data volume 1 × RAID6 (14D+2P) Pool 2: HANA log volume 1 × RAID6 (6D+2P) HDD type 1.2 TB SAS 900 GB SAS Performance is not measured on these systems; however, the same HWCCT configuration is used on all of these systems to generate significant workload. Test Setup The test procedure for this validation is focused on the production SAP HANA nodes fulfilling SAP HANA TDI KPIs. HWCCT revision 97.3 was executed in parallel on the production nodes in partition 1 as well as on all other nodes at the same time. Test 1: Use Case 1 – Using Dedicated RAID groups For use case 1 the four production HANA systems in partition 1 are setup with separate dedicated RAID groups for each system. Having dedicated disks removes the impact on concurrent I/O on the RAID group level from different HANA volumes and the other HANA nodes. Table 4 lists the hardware configuration for the systems in partition 1. Table 4. Hardware Configuration for Production SAP HANA Systems Test 1 Hardware Configuration Number of Production HANA Systems 4 Memory capacity 128 GB per system Shared RAID groups RAID groups per system HDD type OS and HANA shared 1 × RAID6 (6D+2P) HANA data volume 1 × RAID6 (6D+2P) HANA log volume 1 × RAID6 (6D+2P) 600 GB SAS See the section SAP HANA QA and Development Systems for the second partition setup. 18 19 2: Use Case 1 – Using HDP Pools Test Test 2 has the same test setup on the server side as test 1, but it uses a consolidated disk layout based on HDP pools instead of dedicated RAID groups for each system. Table 5 shows the setup where all systems in partition 1 use the same HDP pools for their HANA volumes. Table 5. Hardware Configuration for Production HANA Systems Test 2 Hardware Configuration Number of Production HANA Systems 4 Memory capacity 128 GB per system RAID groups in HDP pools Pool 1: OS, HANA shared, and HANA data volume 2 × RAID6 (14D+2P) Pool 2: HANA log volume 2 × RAID6 (6D+2P) HDD type 600 GB SAS Partition 2 remains unchanged throughout the tests. Test 3: Use Case 2 – Using Dedicated RAID groups For Test 3 a third partition is introduced to the test setup. Partition 1 holds three production HANA systems, while the new second partition holds one system that stands for the classic SAP application server. The last partition again holds the same twelve QA and development systems causing the noisy workload in this partition. Table 6 shows the hardware setup of the production HANA system on partition 1. Table 6. Hardware Configuration Production HANA Systems Test 3 Hardware Configuration Number of Production HANA Systems 3 Memory capacity 128 GB per system Shared RAID groups RAID groups per system OS and HANA shared 1 × RAID6 (6D+2P) HANA data volume 1 × RAID6 (6D+2P) HANA log volume 1 × RAID6 (6D+2P) HDD type 600 GB SAS Partition 2 stands for the classic SAP workload partition and uses the storage layout as shown in Table 7. Table 7. SAP Classic Partition RAID Configuration Volume RAID Configuration OS and HANA shared 1 × RAID6 (6D+2P) with 1.2TB SAS HANA data 1 × RAID6 (6D+2P) with 600 GB SAS HANA log 1 × RAID6 (6D+2P) with 600 GB SAS 19 20 system in partition 2 is considered and measured as a production system using HWCCT since it can also be seen as The having additional HANA systems in this partition, even if the workload for non-HANA systems is usually different from HANA systems. Test 4: Use Case 2 – Using HDP Pools Test 4 is based on Test 2. Table 8. Hardware Configuration Production HANA Systems Test 4 Hardware Configuration Hardware Configuration Number of Production HANA Systems 3 Memory capacity 128 GB per system RAID groups in HDP pools Pool 1: OS, HANA shared, and HANA data volume 2 × RAID6 (14D+2P) Pool 2: HANA log volume 2 × RAID6 (6D+2P) HDD type 600 GB SAS This test setup uses HDP pools for the second partition as well as shown in Table 9. Table 9. SAP Classic Partition HDP Pool Configuration Volume HDP Pool Configuration OS, HANA shared, and HANA data HDP pool with 1 × RAID6 (14D+2P) on 1.2TB SAS drives HANA log HDP pool with 1 × RAID6 (6D+2P) on 600 GB SAS drives The same test conditions as Test 4 apply here. Test Summary Testing was done twice for all of the tests, using HWCCT revision 97.3, on all involved systems, for all of the above test cases. The first test was done without partitioning of storage resources, and the second time the described partitioning scheme was used. Analysis of the first test run showed that the QA and development systems generated a heavy I/O load on the storage that lead to the test failing at least some SAP KPIs. On the storage side it was observed that a high cache write pending rate and MPU usage rate was reached multiple times during the complete test run. Running the same test case with partitioning in test run 2 shows the isolation between the partitions and impact was observed between the partitions. Partition 1 for Test 1 and Test 2, and partitions 1 and 2 for Test 3 and Test 4 passed SAP TDI KPIs without any impact from the QA and development partition. Storage resources were constantly below the upper thresholds and within recommended bounds. High cache write pending and MPU usage rates could still be seen on the QA and development partition which is expected behavior and does not cause any issues because there are no performance requirements for the systems on this partition. 20 21 Conclusion The Hitachi Virtual Storage Platform (VSP) family provides a high performance platform for SAP HANA landscapes and adds significant benefits in reliability, security, and consolidation by providing unique features such as Hitachi storage partitioning and Hitachi Dynamic Provisioning pools. With the right grouping of resources within the VSP configuration, performance can be guaranteed where it is really needed while still offering the possibility of consolidation and costreduction, especially for less critical non-production systems. In combination with Hitachi Compute Blade logical partitioning technology, Hitachi storage partitioning offers the capability for an end-to-end partitioning of server and storage resources. Hitachi Dynamic Provisioning pools add additional cost reduction possibilities when consolidating multiple production and non-production systems on a single Hitachi VSP enterprise storage array without impact on the production SAP HANA systems. HDP pools have great potential for consolidation and can offer the same or even improved I/O performance due to better scaling on several RAID groups compared to a dedicated RAID group setup. This can prove especially valuable when running multiple systems on the same HDP pool at different peak times in regards to I/O workload. 21 For More Information Hitachi Data Systems Global Services offers experienced storage consultants, proven methodologies and a comprehensive services portfolio to assist you in implementing Hitachi products and solutions in your environment. For more information, see the Hitachi Data Systems Global Services website. Live and recorded product demonstrations are available for many Hitachi products. To schedule a live demonstration, contact a sales representative. To view a recorded demonstration, see the Hitachi Data Systems Corporate Resources website. Click the Product Demos tab for a list of available recorded demonstrations. Hitachi Data Systems Academy provides best-in-class training on Hitachi products, technology, solutions and certifications. Hitachi Data Systems Academy delivers on-demand web-based training (WBT), classroom-based instructor-led training (ILT) and virtual instructor-led training (vILT) courses. For more information, see the Hitachi Data Systems Services Education website. For more information about Hitachi products and services, contact your sales representative or channel partner or visit the Hitachi Data Systems website. 1 Corporate Headquarters 2845 Lafayette Street Santa Clara, CA 96050-2639 USA www.HDS.com community.HDS.com Regional Contact Information Americas: +1 408 970 1000 or [email protected] Europe, Middle East and Africa: +44 (0) 1753 618000 or [email protected] Asia Pacific: +852 3189 7900 or [email protected] © Hitachi Data Systems Corporation 2016. All rights reserved. HITACHI is a trademark or registered trademark of Hitachi, Ltd. VSP is a trademarks or registered trademark of Hitachi Data Systems Corporation. Microsoft, Active Directory, and Windows Server are trademarks or registered trademarks of Microsoft Corporation. All other trademarks, service marks, and company names are properties of their respective owners. Notice: This document is for informational purposes only, and does not set forth any warranty, expressed or implied, concerning any equipment or service offered or to be offered by Hitachi Data Systems Corporation. AS-521-00. July 2016
© Copyright 2026 Paperzz