It is difficult to provide specific areas for improvement at this time. We may realize approximately 20% gains in the future, but the previous SAN we used was not very time-consuming, so the improvement is not very significant. We anticipate gains of approximately 10 to 20%. The management interface could be simplified. The cloud console contains two main sections: one for hardware and one for logical volumes on the LUNs being managed. This split is understandable because the hardware console contains HPE Alletra Storage MP B10000 for block storage and HPE Alletra for backup, both accessible in the same console. The separation of the SAN hardware console and the backup console is logical and understandable. We have four compute nodes, and the system was not designed to shut down gracefully with only two compute nodes remaining. Running with just one compute node functions properly, and the system continues to operate at reduced capacity. However, when we experienced a crash involving half of the system, HPE Alletra Storage MP B10000 failed completely. This behavior is documented as normal in the documentation, but we hoped the block storage could manage to run with half the power. This capability will be available in a future release but is not included in the current version. Our testing before moving to production was not entirely successful, but this aligns with the documented specifications. When we perform failover and data center crash tests, the system performs very well. We have deliberately crashed all four compute nodes, and the secondary storage site manages the workload without any virtual machines malfunctioning. This level of reliability is excellent, and it demonstrates why we selected HPE Alletra Storage MP B10000. Improvements in this area would make the solution even stronger.
