Key points
- VMware vSAN combines storage from multiple ESXi hosts into a shared pool, making it easier to scale storage and improve resilience without relying on traditional SAN hardware.
- vSAN changes how databases perform because data is spread across multiple nodes, making network speed, latency, and throughput important factors for read and write performance.
- Storage policies play a major role in database performance by controlling how data is stored, protected, and distributed across the vSAN cluster.
- You can tune vSAN for faster database performance or stronger fault tolerance depending on your workload and business needs.
- vSAN makes it easy to scale database storage by adding more nodes and supports features like deduplication, compression, and erasure coding to improve efficiency.
vSAN is a storage technology that spreads data across multiple nodes, with the goal of improving performance, scalability, and resilience in VMware virtualized environments. As database performance is heavily reliant on latency and throughput, hosting database servers on distributed vSAN is different than local storage, requiring fine-tuning.
This guide explains the factors you need to consider when implementing vSAN for database hosts.
What is a vSAN, and how does it work?
VMware vSAN is a product that combines storage from ESXi hosts in a software-defined, distributed pool. This eliminates the need for costly and complex hardware-based SAN (storage area networks).
vSAN is popular in hyper converged infrastructure (HCI) deployments that tie together compute and storage in software-defined, distributed systems. Nodes can be added when additional processing or data storage is required, rather than having to re-provision individual components. The virtual infrastructure that uses these pooled resources is defined by policies. This provides a modular approach to infrastructure that is highly resilient and scalable, with the flexibility of cloud platforms that allow you to spin-up and spin-down components as they are needed.
Traditionally, vSAN nodes mixed SSDs (for caching and buffering) and traditional spinning disks (for long-term storage and redundancy), however it’s increasingly common for purely SSD-based systems to be implemented.
How vSAN changes storage behavior
As well as the underlying storage hardware, the performance of the distributed storage pool provided by vSAN is heavily dependent on network performance and configuration.
How your database behaves is also a significant factor. Database performance depends on how quickly data can be read and written, and systems can be optimized for read/write performance and resiliency.
Impact on database latency and throughput
The primary factor is latency and throughput – how quickly, and how much data your network infrastructure can move between hosts in the storage pool. Assuming you’ve accounted for the performance of the physical media and network connections used in your storage pool, performance is defined by the storage policies you create and how your workload behaves.
For example, vSAN nodes that are located in the same cabinet with high-speed networking can provide good latency and increased throughput by spreading load across multiple nodes in high traffic environments. With the same virtual resources, however, you could reconfigure your vSAN to prioritize resiliency and fault tolerance to protect vital data by keeping several copies (note that this is not the same as database replication).
The role of storage policies in performance
Storage policies define how vSAN behaves within its virtualized environment, allowing you to specify how storage behaves, from the number of copies to performance parameters like the number of disk stripes per object. Exactly how you configure these will affect read/write speed, resource usage, and how efficiently the storage volume is used.
Your database and priorities for your use case
What ‘efficiency’ means, and whether performance or fault-tolerance is prioritized, will depend on your use case and how your database behaves. One of the biggest advantages of vSAN is that you can tweak your storage configuration to fine-tune it, without having to make hardware changes.
When fine-tuning the performance and resiliency of your database, you can use tools like SQL Server Distributed Replay to benchmark different vSAN storage policies and find the one that works best for you.
How distributed storage improves scalability
Beyond its ability to provide high throughput and its inherent flexibility, vSAN is scalable: if more capacity is needed, you just add more nodes with the required additional storage, and it will be included in the pool. vSAN can optimize capacity, and depending on your configuration, deduplicate, compress, and use erasure-coding to protect data.
This can be cost-effective in scaling environments, ensuring storage is effectively utilized, and able to either prioritize performance through parallelism, or resiliency and data protection by keeping data on separate physical devices transparently.
Challenges of running databases on vSAN
While vSAN solves many problems with traditional SAN deployments, there are challenges: it is highly dependent on an initial investment in fast networking infrastructure and servers. In enterprise use cases, however, this is offset by the later ability to scale by adding new nodes to the existing system.
Configuration complexity can also introduce issues if engineers are not familiar with vSAN policies. Performance or data protection could be compromised due to misconfiguration. For example, it’s important to understand what the theoretical maximum performance of your vSAN system is, to determine whether it is the cause of a bottleneck, as opposed to latency, congestion, or database design issues.
Similarly, you also need to understand the key metrics and conditions for success for your database. Performant storage is generally not a good long-term solution for a poorly optimized database (or queries in the code accessing it).
Note that connecting remote vSAN nodes in the same pool is not a substitute for proper backups, though using stretched vSAN can help with availability as part of a broader strategy.
Monitoring and protecting VMware ESXi and vSAN infrastructure, and the data on it
NinjaOne provides a comprehensive IT management platform, including tools for monitoring and managing VMware ESXi hosts and vSAN infrastructure. NinjaOne can also monitor the performance of SQL servers in real time, giving you the tools you need to monitor database and vSAN performance and health.
Backups are also covered, allowing you to remotely back up database hosts on-site or to the cloud.
Quick-Start Guide
NinjaOne Virtualization Capabilities:
– VMware ESXi Monitoring (versions 6.0–8.1) — Monitor VM hosts and guest machines with metrics like CPU, memory, datastore performance, and network throughput
– Hyper-V Monitoring — Monitor Hyper-V hosts and guest VMs with similar performance metrics
– Datastore Support — Supports NAS and VMFS datastore types for VMware environments
– Performance Conditions — Set alerts for CPU usage, memory, disk active time, disk transfer rates, and network utilization on VMs
What NinjaOne Does NOT Do:
NinjaOne doesn’t provide specialized analysis of how specific storage technologies (like vSAN) impact database workload performance. That’s a specialized infrastructure analysis topic that would require:
– vSAN-specific performance profiling tools
– Database-specific monitoring (SQL Server, Oracle, PostgreSQL, etc.)
– Storage I/O analysis and benchmarking
Recommendation:
If you need to understand vSAN’s impact on database performance, you’d want to consult VMware’s vSAN documentation, performance whitepapers, or use VMware-specific tools like vRealize Operations. NinjaOne can monitor the results of that performance (VM metrics), but not analyze the storage technology’s specific impact.
