SAS RAID Controller for VMware & VMware ESXi: Best Hardware Choices and Practices
A SAS RAID controller is the standard for running VMware ESXi in production. Unlike consumer SATA solutions, SAS controllers deliver enterprise features: dual‑port reliability, high queue depth, and long‑term driver support.
VMware ESXi depends on certified hardware for stable storage integration. Choosing the right SAS RAID controller ensures predictable rebuilds, consistent I/O latency, and seamless compatibility with VMware’s storage stack.
This guide highlights the best hardware RAID options for VMware ESXi, focusing on SAS controllers proven in datacenters for virtualization, database hosting, and mission‑critical workloads.
Why VMware Requires a Certified SAS RAID Controller
VMware ESXi is built for enterprise workloads, and its stability depends on certified hardware. SAS RAID controllers provide the consistency and resi
lience that desktop‑grade or “fake RAID” solutions cannot deliver.
IOPS Consistency vs. Raw Throughput in ESXi Workloads
- IOPS consistency matters more than peak throughput: Virtual machines generate mixed random I/O patterns. Certified SAS controllers are optimized for predictable latency and balanced performance under load.
- VMware integration: ESXi relies on controllers that can sustain high queue depths and manage cache safely, ensuring virtualized workloads remain responsive even during rebuilds.
Why Desktop RAID and FakeRAID Fail Under VMware
- BIOS‑level RAID limitations: FakeRAID depends on host CPU and drivers, which often lack VMware certification. This leads to instability, poor rebuild performance, and unpredictable latency.
- Driver and firmware gaps: Desktop RAID controllers rarely provide long‑term driver maintenance or VMware HCL certification, making them unsuitable for production ESXi environments.
- Result: Arrays may appear functional but fail under stress, risking VM downtime and data loss.
VMware Compatibility Requirements for SAS RAID Controllers
VMware ESXi doesn’t just need a RAID controller that “works” — it requires hardware that meets strict compatibility standards to ensure stability, predictable performance, and long‑term support.
VMware HCL: What Actually Matters (and What Doesn’t)
- What matters: Inclusion on VMware’s Hardware Compatibility List (HCL) means the controller has been tested for ESXi stability, driver support, and firmware reliability.
- What doesn’t: Marketing claims of “VMware ready” without HCL certification. Controllers may function at install but fail under production workloads or kernel updates.
Driver Model, Queue Depth, and Firmware Stability
- Driver model: ESXi relies on stable, vendor‑supplied drivers. Certified SAS controllers ensure driver updates align with VMware releases.
- Queue depth: Enterprise SAS controllers support high queue depths, critical for handling multiple VMs with mixed I/O patterns.
- Firmware stability: Certified controllers undergo regression testing to prevent firmware bugs that can cause datastore corruption or unpredictable rebuilds.
Pass‑Through vs. Hardware Abstraction in ESXi
- Pass‑through mode: HBAs can expose raw disks to ESXi, useful for software‑defined storage (vSAN, ZFS).
- Hardware abstraction: RAID controllers present logical volumes to ESXi, simplifying management but tying recovery to controller firmware.
- Best practice: Choose based on workload — hardware RAID for predictable performance, pass‑through for flexibility and advanced storage stacks.
Best SAS RAID Controller for VMware ESXi (By Use Case)
VMware ESXi requires controllers listed on the VMware Hardware Compatibility List (HCL). The right SAS RAID controller depends on workload scale, performance needs, and budget.
Best Enterprise SAS RAID Controller for VMware Hosts
Broadcom/LSI MegaRAID 9460‑8i / 9560‑16i
- Certified for ESXi 7.x and 8.x.
- High queue depth, RAID 5/6/10 support, and flash‑backed cache for power‑loss protection.
- Ideal for large VMware clusters running databases and virtualization workloads.
Best SAS RAID Controller for Small VMware Clusters
Dell PERC H740P / H745
- Based on Broadcom MegaRAID architecture, fully supported in VMware HCL.
- Strong balance of performance and cost, with predictable rebuild behavior.
- Suitable for SMB clusters or branch office ESXi deployments.
Best Budget‑Conscious SAS RAID Controller That Still Passes HCL
Adaptec SmartRAID 3162‑8i
- Entry‑level SAS RAID controller with VMware certification.
- Supports RAID 0/1/10/5/6, decent cache, and stable driver integration.
- A cost‑effective option for labs or small ESXi setups where HCL compliance is still required.
Recommended Hardware SAS RAID Controllers for VMware
VMware ESXi environments demand controllers that are not only powerful but also certified on the VMware Hardware Compatibility List (HCL). The following SAS RAID controllers are the most reliable options for production deployments.
Broadcom / LSI SAS RAID Controllers (Industry Standard)
Why Broadcom Dominates VMware Deployments
- Broadcom (formerly LSI) MegaRAID controllers are the backbone of most OEM solutions (Dell PERC, HPE Smart Array, Lenovo RAID).
- They offer high queue depth, stable driver support, and long‑term firmware maintenance aligned with VMware releases.
- VMware HCL consistently lists Broadcom SAS controllers as certified for ESXi 7.x and 8.x, making them the safest choice for enterprise workloads.
Cache, BBU, and Write‑Back Behavior Explained
- Write‑back cache accelerates performance but requires protection.
- Battery Backup Units (BBU) preserve cached data during outages, flushing it safely once power returns.
- Flash‑backed cache (FBWC) stores data in non‑volatile flash, eliminating battery maintenance and ensuring longer retention.
- Broadcom controllers integrate these features seamlessly with ESXi, preventing datastore corruption during unexpected shutdowns.
HPE Smart Array and Dell PERC Controllers
OEM Lock‑In vs. Native VMware Stability
- HPE Smart Array and Dell PERC controllers are OEM‑branded versions of Broadcom MegaRAID hardware, tuned for their respective server lines.
- They provide tight integration with vendor management tools (iLO for HPE, iDRAC for Dell) and certified VMware drivers.
- The trade‑off: OEM lock‑in limits cross‑platform flexibility, but stability and support in VMware environments are excellent.
Best RAID Levels for VMware on SAS Storage
Choosing the right RAID level for VMware ESXi hosts is as important as selecting a certified SAS controller. The RAID configuration directly affects VM density, rebuild times, and overall datastore reliability.
Note: what are RAID controllersWhy RAID 10 Remains the Gold Standard for VMware
- Performance + redundancy: RAID 10 combines mirroring and striping, delivering high IOPS and fault tolerance.
- Predictable rebuilds: Rebuilds are faster and less disruptive compared to parity‑based RAID.
- Best fit: High‑density VM workloads, databases, and mission‑critical ESXi clusters where uptime is non‑negotiable.
When RAID 6 Makes Sense (and When It Hurts)
- Pros: Dual‑parity protection allows two simultaneous disk failures, making it attractive for large SAS arrays.
- Cons: Write performance suffers due to parity calculations, and rebuilds can be slow under VMware workloads.
- Best fit: Archival or backup‑focused ESXi datastores where capacity and redundancy outweigh performance.
Why RAID 5 Is a Risky Choice for Modern VM Density
- Single‑parity limitation: Only one disk can fail safely, which is risky in high‑density VM environments.
- Performance degradation: Write penalties and long rebuild times can cripple ESXi performance during disk failures.
- Verdict: RAID 5 is rarely recommended for production VMware hosts; it’s better suited for non‑critical or test environments.
Controller Cache, BBU, and Performance Impact in ESXi
VMware ESXi performance depends heavily on how RAID controllers handle caching and power‑loss protection. Certified SAS controllers balance speed with data safety through write‑back cache and backup units.
Write‑Back Cache: Performance Gain vs. Data Risk
- Performance gain: Write‑back cache accelerates I/O by acknowledging writes before they hit disk. This is critical for VM workloads with high random writes.
- Data risk: Without protection, sudden power loss can corrupt cached data, leading to datastore corruption.
- VMware impact: ESXi hosts require cache protection to ensure VM stability during outages or unexpected reboots.
BBU vs. SuperCap: What VMware Hosts Actually Need
- Battery Backup Unit (BBU): Keeps cache contents alive during power loss, flushing them once power returns. Requires periodic battery replacement.
- SuperCap / Flash‑Backed Cache (FBWC): Uses a capacitor to move cached data into non‑volatile flash. No battery maintenance, longer retention, and faster recovery.
- Best practice for VMware: SuperCap/FBWC is preferred in modern ESXi deployments — it reduces maintenance overhead while ensuring cache safety.
SAS RAID Controller vs. HBA for VMware
VMware ESXi can run on either hardware RAID controllers or HBAs with software‑defined storage. The choice depends on whether predictable performance or flexibility is the priority.
When Hardware RAID Beats HBA + Software Storage
- Predictable performance: Hardware RAID offloads parity and rebuild operations, ensuring stable I/O latency for VM workloads.
- Cache protection: Battery‑backed or flash‑backed cache prevents datastore corruption during power loss.
- Certified drivers: VMware HCL certification guarantees stability across ESXi updates.
- Best fit: Enterprise hosts running databases, high‑density VM clusters, or workloads where uptime is critical.
When HBA Is the Better Choice (vSAN, ZFS, Ceph)
- Software‑defined storage: HBAs expose raw disks, letting VMware vSAN, ZFS, or Ceph handle redundancy and recovery.
- Transparency and flexibility: Admins gain full visibility into disk health, rebuilds, and storage policies without relying on proprietary firmware.
- Scalability: Ideal for hyperconverged infrastructure or environments prioritizing distributed storage.
- Best fit: Modern VMware clusters using vSAN, or setups where ZFS/Ceph provide advanced features like checksumming and self‑healing.
Common VMware SAS RAID Mistakes That Kill Performance
Even with certified hardware, misconfiguration can cripple VMware ESXi performance. These are the most frequent SAS RAID mistakes administrators make.
Wrong Stripe Size for VMFS
- Mismatch issue: VMFS datastores perform best with stripe sizes aligned to typical VM I/O patterns.
- Mistake: Using default or oversized stripe sizes can cause inefficient reads/writes, especially with small random I/O.
- Impact: Latency spikes and wasted throughput under VM density.
Queue Depth Mismatch Between Controller and ESXi
- Controller queue depth: Determines how many outstanding I/O requests can be handled.
- Mistake: Deploying controllers with shallow queue depth or misaligned driver settings.
- Impact: ESXi hosts stall under load, starving VMs of IOPS despite available hardware capacity.
Firmware Drift Outside VMware Support Matrix
- Firmware drift: Updating controller firmware beyond what VMware HCL certifies.
- Mistake: Running “latest” firmware without checking VMware’s support matrix.
- Impact: Driver incompatibility, datastore instability, or rebuild failures.
Recovering VMware Data from a Failed SAS RAID Controller
Even with certified hardware, RAID controllers can fail. When the controller dies but disks remain intact, recovery requires careful handling to avoid permanent VMFS corruption.
What Happens When the Controller Dies but Disks Survive
- Controller dependency: Hardware RAID arrays rely on controller metadata to define disk order, stripe size, and parity.
- Failure scenario: If the controller fails, ESXi may no longer recognize the datastore, even though the disks themselves are healthy.
- Risk: Without the original controller, arrays cannot be mounted directly, leaving VMFS volumes inaccessible.
Why Rebuilding on a New Controller Can Destroy VMFS
- Wrong disk order: A new controller may misinterpret disk sequence, overwriting valid data during rebuild.
- Metadata overwrite: Reinitializing arrays writes new metadata, erasing the original RAID configuration.
- VMware impact: VMFS structures are fragile — once overwritten, recovery becomes far more complex or impossible.
Professional RAID Recovery Options
- Specialized software: Tools can reconstruct logical RAID parameters without forcing a destructive rebuild.
- Filesystem awareness: Recovery solutions must support VMFS to extract virtual machine files safely.
- Best practice: Stop automatic rebuilds, clone disks if possible, and use professional recovery before attempting repairs.
Example: DiskInternals RAID Recovery for VMware Datastores
- Logical RAID reconstruction: DiskInternals free RAID Recovery tool identifies stripe size, parity order, and disk sequence to rebuild arrays virtually.
- VMware support: The software can read VMFS volumes, recovering VMDK files even from damaged or partially corrupted arrays.
- Safe approach: No destructive writes — recovery is performed logically, preserving original disk data.
Ready to get your data back?
To recover data from a RAID disk (documents, databases, images, videos, and other files from your RAID 0, RAID 1, 0+1, 1+0, 1E, RAID 4, RAID 5, 50, 5EE, 5R, RAID 6, RAID 60, RAIDZ, RAIDZ2, and JBOD), press the FREE DOWNLOAD button to get the latest version of DiskInternals RAID Recovery® and begin the step-by-step recovery process. You can preview all recovered files absolutely for free. To check the current prices, please press the Get Prices button. If you need any assistance, please feel free to contact Technical Support. The team is here to help you get your data back!
SAS RAID Controller Comparison Table for VMware
| Controller | Cache / BBU | RAID Levels | VMware HCL | Best For |
|---|---|---|---|---|
| Broadcom MegaRAID | Cache + BBU | 0/1/5/6/10 | Yes | Enterprise hosts |
| HPE Smart Array | Cache + SuperCap | 1/5/6/10 | Yes | HPE servers |
| Dell PERC | Cache + BBU | 1/5/6/10 | Yes | Dell PowerEdge |
Recommendation: Choosing the Right SAS RAID Controller for VMware
Selecting the right SAS RAID controller for VMware ESXi is about balancing certification, performance, and recovery planning. The wrong choice can lead to downtime, unstable datastores, or unsupported drivers.
Decision Checklist for ESXi Administrators
- VMware HCL certification: Confirm the controller is listed for your ESXi version.
- Driver stability: Use controllers with in‑kernel or vendor‑maintained drivers aligned with VMware releases.
- Cache protection: Ensure BBU or SuperCap/FBWC is present to safeguard write‑back cache.
- Queue depth: Match controller queue depth to VM density and workload type.
- Recovery plan: Document how arrays will be rebuilt or recovered if the controller fails.
When Paying More Saves You from Downtime
- Enterprise controllers: Broadcom MegaRAID, Dell PERC, and HPE Smart Array deliver predictable rebuilds and certified stability.
- Downtime cost vs. hardware cost: A higher‑end controller may cost more upfront, but prevents VM outages, datastore corruption, and costly recovery efforts.
- Verdict: For production VMware hosts, investing in a certified enterprise SAS RAID controller is cheaper than risking downtime with budget or uncertified hardware.
FAQ
What is a SAS RAID controller?
A SAS RAID controller is a hardware device that manages multiple SAS (Serial Attached SCSI) drives as a single logical array. It handles tasks like striping, mirroring, and parity calculations to improve performance and provide redundancy. Unlike software RAID, it offloads these operations from the host CPU, ensuring consistent I/O for demanding workloads. Enterprise SAS RAID controllers also include cache with battery or flash‑backed protection to safeguard data during power loss. In VMware ESXi and other virtualization platforms, certified SAS RAID controllers are essential for stable, predictable storage performance.
What are the different types of controllers in VMware?
In VMware, virtual machines use different types of storage controllers to connect disks and devices. The main controller types are:
- SCSI controllers: The default and most common option, supporting multiple virtual disks and optimized for VMFS datastores.
- SATA controllers: Typically used for CD/DVD devices or lightweight workloads; not ideal for high‑performance VM storage.
- NVMe controllers: Designed for modern high‑performance workloads, offering lower latency and higher IOPS compared to SCSI.
- IDE controllers: Legacy option, mainly for backward compatibility with older operating systems.
Each controller type has its own limitations and compatibility considerations, and VMware recommends using SCSI or NVMe for most production workloads.
Can VMware do software RAID?
VMware ESXi does not support software RAID for boot devices or VMFS datastores. Software RAID depends on operating system–specific drivers, which ESXi does not provide. Some HBAs may offer BIOS‑level “fake RAID,” but VMware does not recognize these arrays as supported storage. To run RAID with ESXi, you need a certified hardware RAID controller listed on the VMware Hardware Compatibility List (HCL). For software‑defined redundancy, VMware instead offers vSAN, which uses HBAs to expose raw disks and handles RAID‑like protection at the cluster level.