Overall implementation architecture of distributed storage
In the environment, a storage pool space is formed by using 4 DellEMC PowerFlex Ready Node R740XD nodes. The customer also has several servers running ESXi that access the storage space provided by the Ready Node through the network, using a 2-layer PowerFlex storage structure.
In the figure below, all the PowerFlex nodes are SDS, providing storage space; SDC is provided by the user to allow other hosts to access the storage space. The data is distributed across all the SDS nodes, providing the highest performance and redundancy.
The implementation content of the project involves the expansion and upgrade of the second-phase private cloud architecture using Dell hyper-converged distributed node devices.
Step 1: Network Optimization
1. Count and check the network equipment and network connections
2. Organize and plan the network
3. Network Management (Out-of-Band and System Management)
4. Sphere vMotion Network (This is mandatory for vSphere High Availability Clusters)
5. Storage Network (Front-end and Back-end)
6. Business Network
7. Other Networks
8. Set up VLANs to ensure security
9. Collect the latest information on physical and logical network connections
Step 2: Optimize some of the computing nodes of the virtualization platform
1. Select the nodes with relatively higher configurations from the existing computing nodes.
2. Migrate the existing virtual machines from the selected node to other nodes (provided that the vMotion network is configured)
3. Create a new highly available vSphere cluster (HA & DRS)
Step 3: Optimize the deployment of the PowerFlex platform
1. Confirm the optimized PowerFlex network settings
2. Two backend data networks
3. Two front-end data networks
4. A management network
5. Set up VLANs to ensure isolation of security and availability
6. Re-deploy the new PowerFlex cluster (4 nodes)
7. Manually enter the maintenance mode for each node.
8. Remove 4 nodes from the existing cluster
9. Reinstall the operating system (CentOS 7.9) for 4 nodes
10. Install and configure a new PowerFlex cluster based on the network plan
11. Connect the PowerFlex storage to the new vSphere high availability cluster
12. Expand the new cluster to 8 nodes
13. Migrate the virtual machines stored on the remaining 4-node cluster to the local machine or a new cluster.
14. Deactivate the old cluster and delete the original PowerFlex system
15. Reinstall the operating system (CentOS 7.9) for 4 nodes
16. Based on the network planning, add 4 nodes to the new cluster through installation and configuration.
17. Complete the setup of the new cluster