Recommended settings for RAID0 logical volumes on certain LSI based RAID vSAN

Products

VMware vSAN

Issue/Introduction

This article provides recommended settings for creating RAID0 logical volumes on vSAN compatible controllers.

VMware recommends specific settings when creating RAID0 logical volumes on these controllers:

MegaRAID SAS 9271-8i
UCS-RAID9271CV-8I
MegaRAID SAS 9265-8i
MegaRAID SAS 9266-8i
MegaRAID SAS 9267-8i
MegaRAID SAS 9266-4i
MegaRAID SAS 9270-8i
MegaRAID SAS 9272-8i
MegaRAID SAS 9270CV-8i
MegaRAID SAS 9271-8iCC
MegaRAID SAS 9271-4i
SMC2208
PERC H710 mini Adapter
PERC H710P Adapter
PERC H710 Adapter
Fujitsu SAS RAID HDD Module (D2816C )
Fujitsu SAS RAID HDD Module (D2816)
Fujitsu RAID Ctrl SAS 6G 1GB (D3116C)
Fujitsu RAID Ctrl SAS 6G 1GB (D3116)
IBM ServeRAID M5115 SAS/SATA Controller for IBM Flex System (90Y4390)
IBM ServeRAID M5110e SAS/SATA Controller for IBM System x (90Y4343CTO)
IBM ServeRAID M5110 SAS/SATA Controller for IBM System x (81Y4481)

Note: This list is not exhaustive. This article is linked from the relevant controllers' VCG listings.

Environment

VMware vSAN 5.5.x
VMware vSAN 6.1.x
VMware vSAN 6.0.x
VMware vSAN 6.2.x

Resolution

Recommended RAID0 configuration settings:

VMware recommends the following settings:

If you have CEIP (Customer Experience Improvement Program) enabled and controller vendor utility(storcli/perccli) is available, a vSAN online health check RAID controller configuration validates the recommended controller settings and displays the health status.

Follow the steps below to configure the settings on your controller:

Place the host into maintenance mode with Full Data Migration.
1. If there are not enough hosts to accommodate Full Data Migration, use Ensure Accessibility (with the various standard warnings about exposure to data loss, etc.)
2. Delete the disk groups.
3. Reboot the host.
4. Reconfigure the storage controller according to the manufacturer’s documentation. Please follow these guidelines below to configure the settings:
  1. No cache (or read cache only if cache cannot be disabled)
  2. Pass-through/JBOD mode
    1. Do not use a mixed RAID+ pass-through personality. Only use the pass-through personality if applicable.
5. After boot, re-create the disk groups.
6. Exit maintenance mode.
7. Repeat as needed on the remaining hosts.

Note: To configure the recommended settings, you can use a CLI command from the corresponding controller vendor, such as storcli for LSI controllers and perccli for the Dell H710 controller.

Creating the RAID0 volumes and configuring the recommended settings:

Follow the steps below to create the RAID0 volumes with the recommended configuration settings. These instructions on how to apply the recommended settings using the LSI storcli CLI utilities. For Dell controllers, use perccli.

Note: The working directory for all the commands listed in this procedure is /opt/lsi/storcli. Therefore, execute the command cd /opt/lsi/storcli on the ESX host before performing step 1.

Determine the adapter number of the HBA on which the logical drive will be created. (For this step, assume that we are using a LSI 9271 HBA):
1. Run this command:
  
  ./storcli /call show
2. In the output, look for similar entries as shown in this sample output:
  
  Generating detailed summary of the adapter, it may take a while to complete.
  
  Controller = 0 <---------------This gives us the controller number
  Status = Success
  Description = None
  
  Product Name = LSI MegaRAID SAS 9271-8i
  Serial Number = SV40318544
  SAS Address = 500605b0075888c0
  PCI Address = 00:03:00:00
  System Time = 03/18/2015 21:30:06
  
  The sample output shows that the LSI 9271 is controller 0. This is the value being referred to when using the controller_number in the following commands.
Determine the enclosure ID and the slot number of the disks where the HBA is connected:
1. Run this command:
  
  ./storcli /call show
2. In the output, look for the PD List table as shown in this sample:
  
  PD LIST :
  =======
  ----------------------------------------------------------------------------
  EID:Slt DID State DG Size Intf Med SED PI SeSz Model Sp
  ----------------------------------------------------------------------------
  252:0 9 UGood - 465.25 GB SAS HDD N N 512B ST9500620SS D
  252:1 18 Onln 1 185.781 GB SATA SSD N N 512B INTEL SSDSC2BA200G3 U
  252:2 15 Onln 2 418.656 GB SAS HDD N Y 512B HUC106045CSS600 U

The first column lists the physical drives in the format enclosure_id:slot_number . This provides the enclosure_id and slot_number you will use in the next steps.

Decide the physical drive to use to create the logical volume:
1. Refer to the PD List table generated in Step 2.
  
  Any drive with a status of UGood is available for creating a logical volume. Drives with a status of Onln are already in use by logical volumes and unavailable.
2. Select the drive with status of UGood to use. The slot_number of that drive is what you will use in step 4.
Create the logical volume with the appropriate configuration:
- If it is a HDD, run this command:
  
  ./storcli /ccontroller_number add vd type=raid0 name=virtual_drive_name drive=enclosure_id:slot_number ra wt direct strip=256
  
  Example:
  
  Command using the drive in the PD LIST output:
  
  ./storcli /c0 add vd type=raid0 name=HDD-S0 drive=252:0 ra wt direct strip=256
- If it is an SSD, run this command:
  
  ./storcli /ccontroller_number add vd type=raid0 name=virtual_drive_name drive=enclosure_id:slot_number nora wt direct strip=256
  
  Example:
  
  Command using the drive in our PD LIST output: (assuming it is a SSD):
  
  ./storcli /c0 add vd type=raid0 name=HDD-S0 drive=252:0 nora wt direct strip=256
Verify that the virtual drive is created correctly:
1. Run this command
  
  ./storcli /ccontroller_number show
  
  Example:
  
  ./storcli /c0 show
2. Refer to the VD LIST table in this output:

VD LIST :
=======
------------------------------------------------------------------
DG/VD TYPE State Access Consist Cache sCC Size Name
------------------------------------------------------------------
1/1 RAID0 Optl RW Yes NRWTD - 185.781 GB SSD-S2
2/2 RAID0 Optl RW Yes RWTD - 418.656 GB HDD-S3
3/3 RAID0 Optl RW Yes RWTD - 418.656 GB HDD-S4
4/4 RAID0 Optl RW Yes RWTD - 418.656 GB HDD-S5
5/5 RAID0 Optl RW Yes RWTD - 418.656 GB HDD-S6
----------------------------------------------------------------

Check the line with the virtual disk (VD) name that you created:
1. If the physical drive was an HDD, then it should have the following parameters:
  
  RAID0 Optl RW Yes RWTD
  
  RWTD indicates the cache policy and stands for Read ahead, Write Through, Direct.
2. If the physical drive was an SSD, then it should have these parameters:
  
  RAID0 Optl RW Yes NRWTD
  
  NRWTD indicates the cache policy and stands for No Read ahead, Write Through, Direct.

Additional Information

For more information about supported devices, see the VMware Virtual SAN Compatibility Guide.

For the latest version of StorCLI, do a keyword search for "megaraid storcli" at Avago Support Documents and Downloads.

Disclaimer: VMware is not responsible for the reliability of any data, opinions, advice, or statements made on third-party websites. Inclusion of such links does not imply that VMware endorses, recommends, or accepts any responsibility for the content of such sites