I didn't take the time to check if I have made prior mention of a new project we are enabling here in the lab focused upon SAP Hana HA/DR. This is not a new topic at all and there is a lot of good information out there on the subject of HA/DR but in this case we are building out a physical instance at COIL to demonstrate different SAP Hana HA DR scenarios. From my view it is a new project with respect to the fact that we are finally getting it started this month even though i reality we have been working the plan and planning the work for quite some time.
The pre-sales engineering team approached us about how they might pursue such a project as there has become an increasing desire to more thoroughly address the topic with customers. To be able to prove out some HA DR scenarios can prove challenging as this means you may need to secure hardware and other resources that a customer does not have and yet wants to see proof points on premise. For an HA DR scenario representative of a production environment, it might be unfeasible to spin up something as complex as two 3 node clusters in order to simulate site to site fail overs.
What the team asked is if we could work with some SAP COIL project members and sponsors to build out an HA DR environment here in the lab and then look to run some Data Center Readiness workshops with customers to cover off some of the key HA DR criteria they have for deployment.
The goal for the workshop is to provide a repeatable and programmatic way to effectively address customer’s needs to gain deep understanding of HANA operational capabilities without having to conduct custom POCs.
The proposed framework for the content and deliverable for this Data Center Readiness workshop is similar to existing SAP Hana Technical Academy content. It will comprise of presentation content related to architecture, positioning, roadmap, and presenter-led deep dive demos. There is so far no hands-on exercises for the attendees planned. The preliminary outline for the demos is-
- High Availability – 3-node cluster (2 Active, 1 Standby)
- Show work occurring on HANA.
- Kill one active node to simulate HW failure.
- Show stand-by node coming on-line and serving clients
- Disaster Recovery – Create Active/Passive pairs (2 x 3-node clusters. The active cluster is the same cluster being used for HA scenario)
- Create the replication relationship between two HANA clusters
- Show work occurring in active cluster
- Kill active cluster to simulate a disaster
- Show passive cluster coming on-line, serving clients, and data is consistent.
- Optional Content
- Monitoring and Management
- Studio – this will be used throughout the demo
- Solman/DBA Cockpit/3rd party (optional components)
- Backup and Recovery
- Security (TBD)
- Monitoring and Management
For this project we've worked with Cisco and EMC to secure the necessary hardware and they both look forward to working with the SAP team with producing some really great workshops. As the guy at COIL who works to develop and enable all of these cool projects, I am always in a good mood when I see the hardware standing in the lab.
Here is a shot of both racks that are in the process of being cabled up and configured. We will have two three node clusters, comprised of 4 Cisco B440 M2 blades, some Cisco Nexus switches and EMC VNX 5300 storage (25, 600GB Sata drives). The plan will also look to include some network simulation via Shunra.
Still a lot of cabling and configuration to get after next but we hope to have spin up early April and look forward to seeing the first workshop scheduled. Many thanks to COIL founding sponsor Cisco and to EMC who joined COIL this year as one of its newest COIL Project Members.
As the project develops further, we will keep the blog updated and the JAM and Twitter updates coming.