Creating vLUNs on a Gridstore array
Gridstore provides an alternative to traditional enterprise storage. Basic facts about Gridstore storage technology include:
- It provides storage nodes implemented as 1 RU servers that function collectively as a single storage array.
- Connectivity between the nodes and the storage consumers/compute nodes occurs over one or more 1 or 10 Gbps Ethernet connections.
- NIC teaming can be setup on the Gridstore nodes to provide additional bandwidth and fault tolerance
- It utilizes a virtual controller to present storage to Windows servers
The following is an overview of available vLUN options and features. The lab used consists of 6x Gridstore “H” storage nodes. Gridstore storage nodes are of 2 types: H-nodes and C-nodes. C-nodes are capacity nodes and typically include 4x 3TB 7200 RPM SAS disks. H-nodes are hybrid nodes that include a 550 GB PCIe Flash card. Each node has:
- CPU: 1x Xeon E5-2403 processor at 1.8 GHz with 4 cores (no hyper-threading) and 10 MB L3 cache
- Memory: 32 GB DDR3 1333 MHz DIMM
- Disks (not counting boot/system disk(s)): 4x 3TB 7200 RPM SAS disks and a 550 GB PCIe Flash card
To create a vLUN, in GridControl snap-in, vPools=>(vPool_Name)=>right-click on vLUNs and click Create vLUN
GridProtect level 0: This setting provides no protection against any disk loss in any storage node, or against any node loss in the grid. This option is strongly discouraged.
In this view note:
- vLUN protect level 0 is created on 4 storage nodes listed in the Hostname column (this is the node’s NetBios name)
- The “Disk” and “slot” columns show the actual disks on which this vLUN resides. The following view shows the same information under Storage Nodes:
- vLUNs are thick-provisioned. vLUN entire space is dedicated/reserved on the disks.
To unmask the newly created vLUN and present it to a compute node, right-click on it, and click Add vLUN to Server:
Pick the desired server from the drop down list and select the desired QoS level (Bronze/Gold/Platinum):
The unmasked vLUN becomes visible to the selected compute node:
As any regular disk, we now can bring it online, initialize it (GPT recommended), and format it. I recommend using 32k block size not the 4k default. and naming the volume the same name as the vLUN for consistency. 32k block size enhances IOPS at the expense of potentially wasting disk space if the average file size is under 32k. Given that most workloads will have files larger than 32KB, I recommend using 32KB clock size.