VAAI is a set of APIs and SCSI commands that offload certain I/O-intensive functions from the ESXi host to the storage platform for more efficient performance.
VAAI was introduces in vSphere 4.1 to enable offload of these features:
◎ Full Copy
What is it? Hardware-accelerated copying of data by performing all duplication and migration operations on the array.
Benefits? Faster data movement via Storage vMotion; faster vm creation and deployment from templates; faster vm cloning. Reduces server CPU cycles, memory, IP and SAN network bandwidth, and storage front-end controller I/O.
◎ Block Zero
What is it? Hardware-accelerated zero initialization.
Benefits? Greatly reduces common input/output tasks, such as creating new vm’s. Especially beneficial when creating FT enabled VMs or when performing routine app-level Block Zeroing
◎ Hardware-assisted locking
What is it? Improved locking controls on VMFS.
Benefits? more VMs per datastore. Shortened simultaneous block vm boot times. Faster VM migration.
What is new in 5.0?
Enhancements for environments that use array-based thin provisioning. Specifically:
◎ Dead Space Reclamation
What is it? The ability to reclaim blocks on a thin-provisioned LUN on the array when a virtual disk is deleted or migrated to a different datastore. Historically the blocks used prior to the migration where still reported as “in use” by the array.
Benefits? More accurate reporting of disk space consumption and reclamation of the unused blocks on the thin LUN.
◎ Out-of-space conditions
What is it? If a thin-provisioned datastore reaches 100 percent, only the virtual machines that require extra blocks of storage are temporarily paused, allowing admins to allocate additional space to the datastore. Virtual machines on the datastore that don’t need additional space continue to run.
Benefits? Prevents some catastrophic scenarios encountered with storage oversubscription in thin-provisioned environments.
Configuring / Verifying VAAI Full Copy/Block Zero
In the vSphere client, Host and Clusters > Configuration Tab > (Software) Advanced Settings > DataMover
This one can be a little bit involved. There are various resources addressing this topic, all are referenced at the end of this post.
(In a nutshell)
Step 1 – Verify Hardware Acceleration (VAAI) is supported
Host and Clusters > Configuration tab > (Hardware) Storage > Select Datastore, review details (Not supported in my dinky home lab).
Step 2 – Get the NAA id of the device backing the datastore:
~ # esxcli storage vmfs extent list
Step 3 – Get VAAI status:
esxcli storage core device list -d naa.60a98000572d54724a346a6170627a52
# esxcli storage core device list –d naa.60a98000572d54724a346a6170627a52 naa.60a98000572d54724a346a6170627a52 Display Name: NETAPP Fibre Channel Disk (naa.60a98000572d54724a346a6170627a52) Has Settable Display Name: true Size: 51200 Device Type: Direct-Access Multipath Plugin: NMP Devfs Path: /vmfs/devices/disks/naa.60a98000572d54724a346a6170627a52 Vendor: NETAPP Model: LUN Revision: 8020 SCSI Level: 4 Is Pseudo: false Status: on Is RDM Capable: true Is Local: false Is Removable: false Is SSD: false Is Offline: false Is Perennially Reserved: false Thin Provisioning Status: yes Attached Filters: VAAI_FILTER VAAI Status: supported Other UIDs: vml.020033000060a98000572d54724a346a6170627a524c554e202020
Step 4 – Check if the array supports the UNMAP primitive for dead space reclamation
esxcli storage core device vaai status get -d naa.60a98000572d54724a346a6170627a52
Step 5 – Run the UNMAP primitive command
Caution – We expect customers to use this primitive during their maintenance window, since running it on a datastore that is in-use by a VM can adversely affect I/O for the VM. I/O can take longer to complete, resulting in lower I/O throughput and higher I/O latency.
A point I would like to emphasize is that the whole UNMAP performance is totally driven by the storage array. Even the recommendation that vmkfstools -y be issued in a maintenance window is mostly based on the effect of UNMAP commands on the array’s handling of other commands.
There is no way of knowing how long an UNMAP operation will take to complete. It can be anywhere from few minutes to couple of hours depending on the size of the datastore, the amount of content that needs to be reclaimed and how well the storage array can handle the UNMAP operation.
To run the command, you should change directory to the root of the VMFS volume that you wish reclaim space from. The command is run as:
vmkfstools –y <% of free space to unmap>
Step 6 – Verify
Verify using esxtop > u > f > o > p ; Review the DELETE, DELETE_F and MBDEL/s columns.
For this one I recommend reviewing the article put together by Pauldie O’Riordan, the last referenced below at the end of this post.
Out of Space Conditions / Thin Provisioning Stun
I can’t find a setting for this so I am assuming if VAAI is supported by the array, then the OOS/TPS behavior will apply. I will keep digging on this one. This snippet out of a VMware Community blog clarifies the feature to satisfaction (at least we know what to expect):