This post is intended to cover the VCAP-DCA objective around applying VMware storage best practices. As this is quite a broad topic, I’ll just be covering some general points here. As such, this will likely be a page that I add to from time to time as I work through the VCAP-DCA objectives. For further information there are many VMware documents on the subject and some great VMworld sessions like this one. As always, it is recommended to check any best practices issued by the vendor of your storage devices.
Best Practices when using FC
- There should be a one to one mapping between VMFS datastore and LUN
- Only make the LUNs containing VMFS datastores available to the relevant ESXi hosts and backup devices.
- Ensure you have multiple paths to your storage devices, using a redundant fabric.
- HBA and Storage devices should be on the VMware HCL and the firmware should be supported and up to date.
- Enable Storage I/O Control (SIOC)
- Ensure your solution is designed to meet both capacity and performance needs. Appropriate RAID levels should be used to help achieve this.
- Queue depths should be matched end-to-end.
Best Practices when using iSCSI
- As with FC, only one VMFS volume per LUN
- Bind multiple adaptors to the iSCSI Software Initiator to give redundancy
- Queue depths should be matched end-to-end.
- Give iSCSI traffic it’s own VLAN or network segment (if not using VLANs).
- Enable SIOC
- Ensure hosts have appropriate number of physical nics to handle the iSCSI traffic.
Best Practices when using NFS
- Give NFS traffic it’s own VLAN or network segment (if not using VLANs).
- Enable SIOC
- Ensure the NFS exports are mounted consistently across all hosts.