How to remove vSAN from a cluster of ESXi hosts

[juchestyle]

Hey Everyone,

We are running a POC that is about to turn into Production.  We have 6.5 running vSAN and we were performing testing before our production environment is installed. 

How do we remove vSAN from all 12 hosts so that the disks are usable afterwards?  I ask because I turned off HA and set it to manual and proceeded to remove disk groups.  This worked until I had only 3 hosts left and 3 disk groups.  Now it won't let me continue with the following error:  "A general system error occurred: Failed to evacuate data for disk uuid ##### with error: Out of resouces to complete the operation."

Help?!

[TheBobkin · Accepted solution]

Hello juchestyle​

Did you  put hosts into Maintenance Mode (which option?) before starting to remove Disk-Groups (also which option: Full Data Migration, Ensure Accessibility, No Action)?

Is there only Performance-data currently stored on this cluster and no data that is needed any longer?

Is there a particular reason why you are removing the Disk-Groups and not leaving them configured?

If you are positive that there is no data that is required on the cluster then to fully delete the data and disk-groups, put all hosts into Maintenance-Mode with 'No Action', then delete each Disk-Group with 'No Action' option.

Remove Disk Groups or Devices from vSAN

"A general system error occurred: Failed to evacuate data for disk uuid ##### with error: Out of resouces to complete the operation."

It's throwing this because it is trying to migrate data onto another node but there is no-where to move it to.

Bob

[juchestyle]

Hey Bob,

We are rolling out a VVD (Vmware Validated Design) for NSX, VSAN, and auto provisioning.  Before VMware gets onsite to roll this out we wanted to kick some tires.  Boots on the ground day is getting closer so we need to get the ESX hosts to a fresh starting point after base lining performance.

[TheBobkin]

Hello juchestyle​,

"Before VMware gets onsite to roll this out we wanted to kick some tires."

Of course, good call - knowing what your cluster is capable of is important.

Probably left over Performance VM Objects and/or stuff like stats-db Object.

If you want it completely blanky-blank then remove as I outlined above - hosts in MM with NA, remove DGs with NA.

If for any reason disk-partitions remain stubborn, these can be removed via the Web Client via: Host > Configure > Storage Devices > Select Device > All Actions > Erase Partitions

Bob

[juchestyle]

Completed like a champ!  Should I go into vSAN and uncheck "Turn on vSAN" as the last step?

[TheBobkin]

Hello juchestyle​,

"Completed like a champ! "

Happy to hear that resolved your issue - it stands to reason, as from the error message it was clearly trying to move components (e.g. to ensure accessibility)

"Should I go into vSAN and uncheck "Turn on vSAN" as the last step?"

Yes this might be a good idea as this will mean when it is enabled again it will create a new cluster - otherwise potentially there would be leftover stale cmmds-references etc. (which shouldn't cause issues but cleaner without them).

Bob