Setting up Failover Cluster for Active-Passive Microsoft SQL Server & Hyper-V management
About the Project
Arcadium is one of the best Customer Relationship Management (CRM) systems in the world. It interacts with your existing and potential clients, and it can establish a long term relationship. This client is one of our premium clients and using Dedicated servers for their CRM business.
Today, in a cut-throat competition, a business can survive only if you provide due attention to both existing as well as the potential clients. And this can be easily achieved by using an effective CRM tool.
THE PROBLEM
As per the current scenario, the client works with a total of 3 Dedicated server. Among that, 2 Dedicated servers are handling the Active/Passive SQL Cluster with failover solution while one standalone server has Hyper-V service with more than 50 virtual machines.
Standalone CRM Server Configuration :
| Processor | 2 x E5-2650v4 |
| Total number of VMs | 1500+ Forex VPS with OS Licenses and Dedicated IPS addresses on each VMs |
| Memory | 256 GB RAM |
| Storage | 6 TB |
| OS | Windows 2012 DataCenter |
| Backup | Weekly Backup |
SQL Cluster Server Configuration :
| Processor | 2 x E5-2637 v4 |
| Total number of VMs | 1500+ Forex VPS with OS Licenses and Dedicated IPS addresses on each VMs |
| Memory | 256 GB RAM |
| Storage | 275 GB SSD Boot Drive + 2 TB SSD SAN Storag |
| OS | Windows 2016 Standard (8 Cores) |
| SQL Server | SQL Standard (8 Cores) |
Now, the client is concerned with their virtual machines lying on the standalone Dedicated server and he asked us for a high-availability failover cluster solution. In Addition to this, they were looking for proper Backup and disaster recovery as well. They were ready to move his entire setup on a cloud-based solution.
OUR SOLUTION
Accuwebhosting believes in providing the best solutions for our clients. Upon receiving client's requirement, our management has appointed a dedicated team for this project. After gathering all the technical understanding, we proposed for 2 HyperV nodes with Intel Silver Scalable 4216 CPUs that provide 25% faster performance than the current E5-2650v4 Processor.
To achieve high availability his Hyper-V cluster will be configured with N+1 Hyper-v redundancy. This means one of the nodes can be offline for scheduled or unscheduled maintenance without impacting any VMs. In case of unscheduled maintenance or node failure, VMs on the offline node will be migrated to the new node and reboot. However, in the scheduled maintenance, VMs will be migrated to the new node, totally seamless to the end-user.
Upon looking at the SQL Cluster (maintaining on 2 separate nodes), CPU usage was below 15% and only approx 60GB of memory was being used by SQL Server. We recommended him for expanding this cluster and asked to virtualize the SQL server and run the SQL Server cluster inside of Hyper-V nodes. Later, we can add a third node and we would have sufficient space for all the other VMs too.
Proposed Configuration for Arcadium-cluster with Hyper-V management
Configuration of Highly Available Node 1 :
| Processor | 2xIntel Silver 4216 16 cores 2.1Ghz |
| RAM | 384GB RAM DDR4 |
| OS | Windows 2019 Data Center Edition 32 Cores |
| Storage | 6TB All Flash SSD SAN Storage |
| Backup Option | 5TB Rubrik Primary Backup +14TB Rubrik Secondary Backup Offsite Archival |
Configuration of Highly Available Node 2 :
| Processor | 2xIntel Silver 4216 16 cores 2.1Ghz |
| RAM | 384GB RAM DDR4 |
| OS | Windows 2019 Data Center Edition 32 Cores |
| Storage | 6TB Nimble SAN Storage |
Additionally, we will monitor the on-going SAN replication and our team will provide 10 hours of professional services for configuration of cluster and migration of existing workloads and other 10 hours of professional services for configuration of disaster recovery node.
THE RESULTS
Client had accepted our solution and currently running on our high availablability failover cluster. Our team provides him required help related to vm management and server management at regular intervals. As per them, Dedicated server failover cluster is working best for their hyper-v management.