Episode #291
Introduction
In episode 291 of our SAP on Azure video podcast we talk about Azure SAP Zone Alignment.
In this video, Ralitza Deltcheva and Sanoop Thrivikraman Nampoothiri talk about how the new Azure SAP Zone Alignment Resource Agent automates SAP application server alignment with the active HANA zone to improve high availability, performance and cost optimization on Azure.
Please checkout “Aligning SAP application servers with the HANA primary zone on Azure (Public Preview) — Part 1”: https://techcommunity.microsoft.com/blog/SAPApplications/aligning-sap-application-servers-with-the-hana-primary-zone-on-azure-public-prev/4490925
Find all the links mentioned here: https://www.saponazurepodcast.de/episode291
Reach out to us for any feedback / questions:
- Goran Condric: https://www.linkedin.com/in/gorancondric/
- Holger Bruchelt: https://www.linkedin.com/in/holger-bruchelt/
#Microsoft #SAP #Azure #SAPonAzure #HighAvailability #AzureZones
Summary created by AI
- SAP Application Server Zone Alignment Automation:
- Goran, Ralitza, and Sanoop discussed the challenges of aligning SAP application servers with the active HANA database zone, introducing a new Azure Pacemaker resource agent developed by Ralitza’s team to automate this process and reduce manual intervention for customers.
- Deployment Patterns Overview: Ralitza explained the two main deployment patterns for SAP on Azure—active-active and active-passive—highlighting the importance of aligning application servers with the database zone to minimize latency and optimize performance. In active-passive setups, only the app servers in the same zone as the database are kept active, while others are stopped or deactivated.
- Manual Versus Automated Alignment: Previously, customers had to manually align SAP application servers with the database zone after a failover, which often led to administrative overhead and potential performance issues. The new resource agent automates this process, ensuring seamless alignment without manual intervention.
- Resource Agent Functionality: The Azure Pacemaker resource agent detects HANA failovers, identifies which application servers are in the same or different zones, and performs actions such as starting, stopping, or deactivating SAP instances as needed. It uses Azure APIs and managed identities for secure operations, and supports both user-assigned and system-assigned managed identities.
- Modes of Operation and Cost Optimization: The agent offers two modes: one prioritizes speed by deactivating instances, and the other prioritizes cost by stopping and deallocating VMs when not needed. Customers can configure the agent to suit their operational preferences, balancing performance and cost.
- Integration and Customization: The agent integrates with Pacemaker clusters, uses standard cluster attributes for status tracking, and provides extensive configuration options, including patterns for VM names, logical grouping for proximity placement groups, and adjustable timeouts for graceful shutdowns. It does not require additional agents on app servers or direct network connectivity between HANA and app servers.
- Technical Demonstration of Resource Agent:
- Sanoop conducted a live demonstration showing the resource agent in action during a HANA failover, illustrating how SAP application servers are automatically aligned with the new primary database zone and how the agent manages instance states and VM operations.
- Demo Environment Setup: Sanoop described the demo environment, which included a HANA scale-up system deployed across two Azure zones, four SAP application servers, and two HANA cluster nodes. The agent was configured to manage active and passive states of the application servers based on the primary database node.
- Failover Process Automation: During the demo, Sanoop triggered a HANA failover, and the agent detected the new primary node, started or deactivated application server instances as appropriate, and ensured that all instances in the same zone as the new primary were active and healthy before deactivating those in the other zone.
- Azure API Utilization: The agent uses Azure VM start, stop, and run command APIs to manage VM and SAP instance states, relying on Azure managed identities for authentication and eliminating the need for additional agents or direct network connectivity between HANA and app servers.
- Cluster Attributes and Monitoring: Sanoop highlighted the use of Pacemaker cluster attributes for real-time monitoring of the agent’s activities, including phase tracking and timeout controls, which provide transparency and troubleshooting capabilities during failover events.
- Supported Scenarios and Limitations:
- Ralitza and Sanoop clarified the current scope of the resource agent, noting its support for HANA scale-up, SAP ABAP instances, and both zonal and proximity placement group deployments, while excluding SAP Java, HANA scale-out, and multi-SID environments from the public preview.
- Supported Configurations: The agent supports SLES 15 SP5 or later, SAP ABAP instances on HANA scale-up, zonal deployments, and proximity placement groups. Customers can provide logical groupings for VMs if zone metadata is unavailable.
- Current Limitations: The public preview does not support SAP Java, HANA scale-out, or multi-SID environments. Network access requirements are limited to outbound access from HANA VMs to Azure API endpoints, with no need for direct connectivity to app servers.
- Configuration Parameters and Customization:
- Sanoop detailed the resource agent’s configuration options, including cluster attributes, VM naming patterns, logical groupings, managed identity settings, and various timeout controls, enabling customers to tailor the agent’s behavior to their operational needs.
- Cluster Attributes and Timeouts: The agent uses cluster attributes to track phases and start times, allowing customers to set timeouts for each phase and control how long operations can run before triggering alerts.
- VM and Instance Configuration: Customers can specify VM names individually or by pattern, provide logical groupings for proximity placement groups, and adjust wait times for starting and stopping SAP instances to ensure smooth transitions during failovers.
- Logging and Troubleshooting: Verbose logging can be enabled for detailed troubleshooting, and all activities are recorded in standard Pacemaker logs for transparency and support.
- Customer Adoption and Feedback:
- Ralitza and Goran encouraged customers operating in active-passive mode to test the resource agent in non-production environments, provide feedback, and suggest improvements via blogs or community channels.
- Testing and Feedback Channels: Customers are invited to try the resource agent in their environments, share feedback, and request enhancements through blog comments, YouTube channels, or other community platforms.
- Documentation and Guides: Comprehensive documentation, including an announcement blog and a technical setup guide, is available to assist customers with configuration and parameterization of the resource agent.