Migrating over 200 TB SAP Oracle Database to Azure with Lemongrass Cloud Platform Migrate
Co-Authors Ravi Athmakuri Vinaykumar Jangamshetty Amandeep Kalsi Sachin Ghorpade Overview In this blog post, we will cover SAP on Azure solution and Lemongrass Cloud Platform Migrate solution and how it helped to migrate over 200 TB Oracle on SAP system to Azure with 80% less downtime than SAP standard migration. Customer Environment and Challenges The process of migrating an SAP solution implemented 20 years ago to Azure can present multiple challenges and we will discuss the key challenges faced by customers who want to take advantage of Azure’s scalability, flexibility and performance. Non-Unicode SAP Systems Non-Unicode systems are incompatible with newer SAP environments, especially SAP S/4HANA, which requires Unicode. Non-Unicode systems also limit integration with modern applications and languages. Large Database (over 200 TB) Migrating a very large database requires robust data transfer solutions, especially for environments that are sensitive to downtime. Big Endian Platforms (IBM Power, SUN/HP, Mainframe) Platforms like IBM Power typically run-on Big-Endian architecture, whereas most cloud environments and databases are on Little Endian, which complicates data migration. Most of the public cloud environment and including SAP HANA, runs on Little Endian. (Heterogenous Migrations) Lack of Database Compression Without compression, the database may be significantly larger than necessary, increasing costs and migration complexity. SAP Heterogeneous migration and Key Challenges SAP heterogenous migrations are a common technique used to solve the challenges above. Heterogeneous migrations in SAP involve shifting an SAP system across different operating systems, Hardware platform or databases, often involving a platform change such as migrating from Oracle on AIX to Oracle on Linux. While heterogeneous migrations can deliver significant performance improvements and allow for platform modernization, they are also complex and pose specific challenges. We will discuss those challenges in this section before we move to the next section to address those challenges. Downtime Management Minimizing downtime is essential to avoid business disruption. However, with large datasets and complex dependencies, reducing downtime during migrations can be challenging. Large databases introduce additional complexities, including prolonged data transfer times, which can delay the migration. Extended Data Transfer Times Data transfer time directly affects the overall migration timeline, especially for remote data centers or cloud migrations. When migrating to a cloud platform or a distant data center, network latency can slow down data transfer, leading to increased downtime. Configuring parallel processes during migration can significantly reduce data transfer time. However, improperly configured parallelism can lead to resource contention and migration failures. Ensuring Data Consistency Data consistency is crucial in SAP migrations, as any mismatches can lead to severe application and transactional issues. Employ SAP tools like R3load and the Migration Monitor, which perform integrity checks during data transfer. It’s also best to run multiple post-migration checks and test cycles to validate data consistency. Mock Runs Migrating between platforms can surface unexpected issues, from compatibility problems to unoptimized configurations. Run mock migrations and testing to identify potential bottlenecks, without affecting the running production systems. Next, we’ll explore SAP on Azure and the Lemongrass Cloud Platform (LCP) Migrate solution, focusing on how it modernizes and optimizes SAP systems on Microsoft Azure. SAP on Azure and LCP Migrate addresses key challenges like downtime, large data volumes, network performance, and data consistency. SAP on Azure Solution The architecture and solution for hosting a 200+ TB SAP Oracle Database on Azure discussed in detail in another blog post: “Scalability in the Cloud: Migrating over 200 TB SAP Oracle Database to Azure”. This blog explores SAP on Azure solutions that minimize migration downtime, focusing on the need for strong parallel processing, high storage, and network throughput. Migration Parallel Processing We utilized Azure's Mdsv3 series VMs, featuring Intel’s Sapphire Rapids processors with DDR5 memory. Mdsv3 SKU provides higher SAPS than previous generations, allowing increased parallelization during critical SAP import stages of migration. Storage and Network Throughput Azure VM SKUs are equipped with right-sized storage and network technology to meet the migration downtime requirement. The Mdsv3 VMs incorporate Azure Boost technology, which enhances storage efficiency and performance by offloading tasks to Azure Boost FPGA. This offload improves security, reduces jitter, and lowers latency, while also enhancing network performance with up to 200 Gbps bandwidth and strengthened network security. Storage Leveraged Azure NetApp Files (ANF) for the SAP data and redo file systems to provide required throughput with low latency. ANF enabled us to adjust throughput dynamically to support different phases of the migration, without causing any outages to the operational Oracle system. Network Connectivity For private connectivity between on-premises and Azure, we used Azure ExpressRoute. This solution, facilitated by connectivity providers, bypasses the public Internet, offering enhanced reliability, faster speeds, consistent latency, and superior security compared to standard Internet connections. Azure ExpressRoute played a vital role in shipping data between on-prem to Azure to achieve the downtime requirement. Azure Monitoring Leveraged Azure Monitoring and dashboard to monitor the real time metrics of CPU, IO and Network throughput during the migration process to catch any hot spots so that can be dynamically addressed without interrupting the migration process. LCP Migrate LCP Migrate is a proprietary technology product developed by Lemongrass for migrating very large databases (5TB-250TB) from source on-prem to Microsoft Azure. We dive into LCP Migrate and its role in modernizing and optimizing SAP systems, addressing the critical challenges of downtime, large data volumes, network performance, and data consistency. By implementing LCP Migrate, customers have been able to meet business downtime requirements. Using LCP Migrate, SAP Complex migrations like these can be achieved with minimal downtimes using the incremental approach by moving high volume passive data during uptime (Migrate all stable range data during uptime in the week(s) prior to offline phase) along with light weighted triggers (Record all changes to data since start of online migration) on these and lower volume active data during the cutover (Migrate all changed data and all remaining data during offline phase). Mock system deployed by using LCP Migrate in parallel to production to validate end to end business processes. The LCP Migrate approach is based on the SAP Supported tooling like Migmon, R3load with tuning across these. Historically, a comparative analysis of migration methodologies focused on downtime based on the larger database sizes, as illustrated below: LCP Migrate also offers various solutions like SAP Non-Unicode to Unicode Conversions along with Migrate, Upgrades with Migrate, Plugin service to SUM DMO and many more optimizations. Using LCP Migrate solution, we have made sure that the data transfers during the cutover event to optimal number where the throughput was consistent. LCP Migrate solution incorporates automation tools, including Ansible, scripting, and custom dashboards, to facilitate real-time monitoring of migration activities. Database Compression Constructed a standardized template-based framework for SAP ERP/SCM/BW/CRM on Oracle Database, aligned with SAP on Azure and SAP on Oracle best practices, and seamlessly integrated with LCP Migrate migration and Unicode conversion processes. By leveraging Lemongrass’s refined templates and process, we can attain a compression ratio exceeding 65% on target databases concurrently with the migration. Some of the Advantages which we gain with this is smaller database size, Lower maintenance for disk allocation, HA/DR, Backup maintenance, more efficient use of the Oracle buffer pool due to a higher data density, Improved system performance, and fewer extrusions must be made from the buffer pool and overall improvement in the Business Processes. Operationally, advanced compression is utilized to maximize the database's compression benefits. While basic compression generally requires less CPU than advanced compression, it also applies an ILM policy that periodically compresses data which has been decompressed due to DML activity. This process introduces additional CPU load, negating some of the CPU savings. Therefore, using advanced compression with minimal CPU overhead may be more beneficial for overall performance. Conclusion A technically sound, well-executed and successful migration of a large-scale SAP system to Azure, leveraging the advanced capabilities of LCP Migrate. The LCP Migrate solution proved to be a valuable tool for complex SAP migrations, offering a robust and efficient approach to minimize downtime and maximize system performance. By addressing the challenges of heterogeneous migrations and leveraging innovative technologies, Lemongrass has demonstrated its expertise in delivering successful SAP migrations to Microsoft Azure. References About Lemongrass - Company Overview eBook – Lemongrass SAP on Azure Video – Lemongrass Cloud Platform (LCP) Video – Lemongrass Cloud Platform Migrate (LCP Migrate) Video – Lemongrass Minimum Downtime Operations (MDOTM) Video – Lemongrass and Microsoft Talk All Things RISE with SAP
