VMDK Reconfiguring status in VSAN

[tekhie999]

hi all

I am running VSAN6.2 and have VM's provisioned with an FTT=1.

If i increase the size of a VMDK in a VM, i see the "Compliance status" of the VMDK change to "non-compliant", and the "Component state" change to "reconfiguring".  If i had a failure in the cluster whilst this operation was taking place, would access to the VMDK be impacted ?

so for example, i have a VMDK object on ESXA and on ESXB

I increase the disk size of VMDK from 200GB to 500GB. The single component will now be in a state of "reconfiguring" as a single component size can be a maximum of 255GB. So the VMDK that was a single component is now "reconfiguring" to become 2 components.

If ESXA failed before the "reconfiguring" operation was completed - would the VMDK become inaccessible ?  because the "compliance status" is "non-compliant" whilst its reconfiguring ?

Any info would be very useful

Thanks!

[TheBobkin]

Hello tekhie,

There can be more than one reason for having 'Reconfiguring' components. If it is an availability-related rebuild (e.g. rebuilding a failed component from a healthy replica) then if you lose the host/disk that has the remaining healthy data component, then yes this vmdk Object will become unavailable.

However under normal conditions where everything is healthy, when it starts a reconfigure process for instance to split a vmdk into more sets of components due to size or striping, it basically just creates these components anew and does not discard the original components until this process has successfully completed. So while these also will show as 'non-compliant' it is for a different reason.

RVC is your friend here as always and will give much better insight into the component availability and layout (you can however also see this via the Web Client).

Bob

-o- If you found this comment useful please click the 'Helpful' button and/or select as 'Answer' if you consider it so, please ask follow-up questions if you have any -o-

[tekhie999]

Hi bob - thanks for the info

So when reconfiguring due to a size increase, does it reconfigure both replicas simultaneously ? or does it only do 1, and then once complete create a 2nd replica for the 2 components it has created during the reconfiguration process?

With the above in mind  if during the reconfigure process, I had an ESX Host fail, and on that Hosts disks it was running a replica of the  VMDK component being reconfigured – would the VMDK become inaccessible and the reconfiguration job fail ? Or, would I/O carry on being serviced from the 2nd replica that is running on an ESX Host that didn’t fail and reconfigure itself accordingly to 2 components ?

Then once complete, either wait 60 minutes and create 2 new components to restore the FTT=1, or if the Host came back online either resync or carry on reconfiguring until the FTT=1 is restored

Is that correct or does it work a different way ?

Thanks