VCP550 · Question #47
A Windows Server 2008 virtual machine with the following configuration receives an Out of Space condition indicator: - The virtual machine uses thin provisioned virtual disks. - The VMFS5 datastore on
The correct answer is A. Delete unneeded files on the VMFS datastore and then use vmkfstools to reclaim deleted blocks.. When unused VMFS datastore content causes a thin-provisioned LUN to run out of space, the administrator must delete the files from the datastore and then use vmkfstools to issue the VAAI UNMAP primitive, signaling the array to reclaim the freed blocks.
Question
A Windows Server 2008 virtual machine with the following configuration receives an Out of Space condition indicator:
- The virtual machine uses thin provisioned virtual disks.
- The VMFS5 datastore on which the virtual machine's files reside is on
a thin provisioned LUN.
- 30% of the VMFS datastore content is no longer needed.
- The storage array supports VASA and VAAI.
Which action should the vSphere administrator take to resolve the Out of Space problem without requesting additional space on the storage array?
Options
- ADelete unneeded files on the VMFS datastore and then use vmkfstools to reclaim deleted blocks.
- BDelete unneeded files on the VMFS datastore, right-click the datastore in vSphere Web Client, and then
- CDelete some files from the guest OS and defrag its file system.
- DDelete some files from the guest OS and use vmkfstools to reclaim the deleted blocks.
How the community answered
(55 responses)- A67% (37)
- B9% (5)
- C4% (2)
- D20% (11)
Why each option
When unused VMFS datastore content causes a thin-provisioned LUN to run out of space, the administrator must delete the files from the datastore and then use vmkfstools to issue the VAAI UNMAP primitive, signaling the array to reclaim the freed blocks.
Deleting unneeded files from the VMFS datastore frees those blocks at the file system level, but the storage array does not automatically know those blocks are available because of how thin provisioning works at the LUN layer. Running vmkfstools with the reclaim option issues the VAAI UNMAP primitive to the array, which instructs the thin-provisioned LUN to release those blocks back to the storage pool - resolving the Out of Space condition without requesting additional raw capacity.
Right-clicking a datastore in the vSphere Web Client does not trigger a VAAI UNMAP operation to reclaim blocks from a thin-provisioned LUN; vmkfstools must be used from the command line to issue the UNMAP primitive.
Deleting guest OS files and defragmenting only reorganizes data within the virtual disk's allocated blocks and does not release unused blocks back to the VMFS datastore or signal the storage array to recover capacity from the thin-provisioned LUN.
Deleting guest OS files addresses space within the virtual machine's thin virtual disk but does not target the 30% of unneeded VMFS datastore-level content, and UNMAP must be issued at the VMFS-to-LUN layer rather than triggered from inside the guest.
Concept tested: VAAI UNMAP primitive for thin-provisioned LUN space reclamation
Topics
Community Discussion
No community discussion yet for this question.