VMware Space Reclamation not working

[Enrico Guarino]

Dear Quadstor,

we have testing and loving your virtual storage (version 3.0.65.10) with VMware ESXI 5.1.0.

Our setting are the following:

• Phisical Storage: D L U C WC
  
• Configurates Pools: D L
  
• Configured VDisks: D V V E

We are expiriencing some issue related the reclaim space on Phisical Storage when we deleted files on the related datastore:

1) We have followed this VMware KB in order to reclaim space manually and this command works fine.

 

Using vmkfstools to reclaim VMFS deleted blocks on thin-provisioned LUNs (2014849)

~ # esxcli storage core device vaai status get -d naa.6e97944b99c67339e748743447e8968b

naa.6e97944b99c67339e748743447e8968b

   VAAI Plugin Name:

   ATS Status: supported

   Clone Status: supported

   Zero Status: supported

   Delete Status: supported

/vmfs/volumes/53946cbe-b3e658d4-02a1-001999edb908 #  vmkfstools -y 99

Attempting to reclaim 99% of free capacity 1022.8 GB (1012.6 GB) on VMFS-5 file system 'QUADSTORE' with max file size 64 TB.

Creating file .vmfsBalloonjqjKuO of size 1012.6 GB to reclaim free blocks.

Done.

2) Then we try to set the UNMAP option to 1 on the ESXi host in order to reclaim space automatically but it didn't work.

 

Disabling VAAI Thin Provisioning Block Space Reclamation (UNMAP) in ESXi 5.0 (2007427)

~ # esxcli system settings advanced set --int-value 1 --option /VMFS3/EnableBlockDelete

~ # esxcli system settings advanced list --option /VMFS3/EnableBlockDelete

   Path: /VMFS3/EnableBlockDelete

   Type: integer

   Int Value: 1

   Default Int Value: 0

   Min Value: 0

   Max Value: 1

   String Value:

   Default String Value:

   Valid Characters:

   Description: Enable VMFS block delete

~ # esxcli storage core device vaai status get -d naa.6e97944b99c67339e748743447e8968b

naa.6e97944b99c67339e748743447e8968b

   VAAI Plugin Name:

   ATS Status: supported

   Clone Status: supported

   Zero Status: supported

   Delete Status: supported

On new esxi release this option should be enabled for default in t10 complaiant storage.

Should we do something else?

We hope this can help other users to better understand the quadstor on vmware infrastructure.

Thank You in advance

Enrico

[QUADStor Support]

This behavior is documented under
http://www.quadstor.com/support/120-creating-and-managing-virtual-disks.html#esxi_unmap

In ESX 5.1 and above, UNMAP(s) are not issued by default and space
reclaimation has to be initiated manually. This can be inconvenient
but unfortunately it isn't under our control.

Also there is an issue with overprovisioned disk if the unmap isn't
run in time. For example
Physical storage space = 100GB
VDisk size = 200 GB
Physical space used = 100GB
Real Usage = 10GB (Actual space used by datastore). In other words
after a vmkfstools -y 99 would bring the usage to 10 GB or so.

However since reached the limit in terms of free physical space,
running vmkfstools will still not reclaim as additional space since
more writes are needed first.

> --
> You received this message because you are subscribed to the Google Groups
> "QUADStor Storage Virtualization" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to quadstor-vir...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.