When setting up a VMware ESXi host for SharePoint development, storage configuration becomes critical. Your Dell R610 with 6x146GB SAS drives presents two viable options:

• RAID 10: Mirroring + Striping (3 usable drives, ~438GB capacity)
• RAID 5: Striping with Parity (5 usable drives, ~730GB capacity)

SharePoint development VMs have unique I/O patterns:

// Sample PowerShell to monitor disk latency (run on ESXi host)
Get-Stat -Entity (Get-VM) -Stat "disk.totalLatency.average" -Realtime -MaxSamples 10 |
Sort-Object -Property Value -Descending |
Format-Table -AutoSize

Key observations from our testing:

For your specific workload mix (AD, SQL, build server), consider these factors:

# Recommended H700 settings for RAID 10
MegaCli -CfgLdAdd -r10 [252:0,252:1,252:2,252:3,252:4,252:5] WT NORA Direct -strpsz256 -a0

# Alternative for RAID 5
MegaCli -CfgLdAdd -r5 [252:0,252:1,252:2,252:3,252:4] WT NORA Direct -strpsz256 -a0

Your 6-drive configuration presents interesting scenarios:

• RAID 10: Better for sustained write operations (VM snapshots, SQL transactions)
• RAID 5: More space for test farms but potential "write hole" risk during power failures

Example VM storage allocation:

// Sample ESXi CLI commands for VMFS allocation
esxcfg-scsidevs -m
esxcli storage filesystem list

With 146GB SAS drives, consider these rebuild durations:

• RAID 10: ~2 hours per failed drive (minimal performance impact)
• RAID 5: ~6 hours per failed drive (significant performance degradation)

For your development environment where VMs can be easily redeployed, RAID 5 might be acceptable. However, for production-like testing scenarios, RAID 10 provides more consistent performance.

Implement these checks regardless of RAID choice:

# Nagios check for RAID health
define command {
    command_name check_raid
    command_line /usr/lib64/nagios/plugins/check_megaraid_sas -b /opt/MegaRAID/MegaCli/MegaCli64 -D 0 -C 0 -v
}

When provisioning storage for multiple VMs running on VMware ESXi, the RAID configuration becomes critical for both performance and reliability. Let's analyze the key metrics:

// Benchmark comparison pseudocode
raid5_write = (single_disk_write * (n-1)) / n
raid10_write = single_disk_write * (n/2) 
raid5_read = single_disk_read * (n-1)
raid10_read = single_disk_read * n

• Usable capacity: (6-1)*146GB = 730GB
• Write penalty: 4 I/O operations per write (read parity, read data, write parity, write data)
• Rebuild time: High risk with large drives (10k RPM helps)

• Usable capacity: (6/2)*146GB = 438GB
• No write penalty (straight mirroring)
• Faster rebuilds (only need to copy mirror pair)

For your SharePoint VM workload:

# Sample ESXi storage configuration
esxcli storage nmp device list # Check current paths
esxcli storage core device vaai status get # Verify acceleration
esxcli storage nmp satp rule add -s VMW_SATP_LOCAL -d DELL -c tpgs_on

For your Dell H700 controller:

# RAID controller monitoring example
/opt/dell/srvadmin/bin/omreport storage pdisk controller=0
/opt/dell/srvadmin/bin/omreport storage vdisk controller=0

The H700's 512MB cache helps RAID5 performance, but can't eliminate write penalties during heavy VM operations.

Consider splitting disks:

• 2 disks RAID1 for OS/VMs
• 4 disks RAID10 for active VM storage
• Use SSD caching if possible