Hi There,
We are managing four VMware ESXi 5.0 hosts and VSC 4.2 installed in our vCenter Server. 2TB Thin LUN provisioned from NetApp FAS3240 controller. We used around 1.5TB of data earlier and now we have deleted 1TB of data in VMFS datastore. current usage is only 500GB. But, still NetApp end LUN occupied size is 1.5TB
I understand that, VAAI primitive will be used to reclaim dead space by run the below commands,
#vmkfstools -y <x> where x=percentage of dead space to be reclaimed.
I verified "hardware acceleration" tab of the subjected datastore and it shows "supported"
Please let me know, is there any downtime required to run the above said command to reclaim the space.
Regards,
Dhakshinamoorthy Balasubramanian
You should no need to schedule downtime. However, make sure you are well aware of the command's impact, and also consider carefully what percentage you are going to specify. Unless already done, read http://cormachogan.com/2013/11/27/vsphere-5-5-storage-enhancements-part-4-unmap/ and the linked articles in this blog post.
André
Hi,
I would recommend that you do this reclamation after business hours when there is minimum downtime, also note that you are using ESXi 5.0 and there were some issues that were reported w.r.t. to Storage performance hence this option was disabled in the next ESXi update.
I have the UNMAP primitive for ESXi 5.5 and I see that the controller CPU spikes to 20-40% depending on the existing load, hence suggest that you test this only when there is minimum load on the controller and ensure that you have backups.
I have written a blog specifically about ESXi 5.5 and UNMAP primitive with NetApp, you may refer to this blog post or other vStorage API blog posts for more information.
Data Center on Cloud: Using VAAI UNMAP on vSphere 5.5 & NetApp Storage
-Arun
http://highoncloud.blogspot.in/
About VMware Virtualization on NetApp
On the NetApp side are you doing volume snapshots? The problem is that if you have volume snapshots after you delete that data the space is not immediately freed up until those snapshots have been automatically deleted through retention policy or manually deleted.
GMCON wrote:
On the NetApp side are you doing volume snapshots? The problem is that if you have volume snapshots after you delete that data the space is not immediately freed up until those snapshots have been automatically deleted through retention policy or manually deleted.
This is a good point but Netapp snapshots are at the volume level and dont affect the LUN used space which is what the OP is concerned with.
My recommendation would be to create a new 2TB Thin provisioned LUN, present it as a new datastore and Storage vMotion your VM's to it (assuming you have the relevant licensing).
This will only consume an additional 500GB of space on your SAN and you can then destroy the original LUN that is consuming the 1.5TB of space. However, GMCON's point about snapshots now comes into play - if you have snapshots of the LUN then you will only recover the space to the aggregate once all the snapshots in that volume/or the volume itself are deleted.
This will probably only take 2 hours to complete and as you are using the VSC its just a few clicks
Hello GMCON and Gav0,
There is no volume level snapshot configured.
We have configured SMVI snapshots on VM level.
Please note that, we have enabled SIS (deduplication) on respective volume.