Nimble VVol VM Storage Policy with replication Requested operation is not supported

[jchen522]

When creating VM storage policy in vCenter, if I select replication partner, my VVol datastores show not compatible.  If I don't select replication partner and select just snapshot locally, my VVol datastores shows compatible.  Why is that?  VVol datastores was created using vCenter plugin.  VMFS replication works fine between replication partners.

It gives the following reason.

Datastore does not match current VM policy.  Requested operation is not supported.

Any help would would be much appreciated!!!

[mamatadesaiNim]

Making sure I understand correctly:
* there are 2 nimbles, an upstream-A and a downstream-B
* the upstream has vCenter-A registered, vcPlugin from upstream-A is installed on vCenter-A
* there is a vvol datastore-A from upstream-A mounted on at least one ESXi host in vCenter-A
* creating a VM Storage Profile on vCenter-A, in the create screens, if a replication partner is selected in the drop-down, datastore-A shows in the "incompatible" list

Questions:
* "Nimble storage rules" has 3 tabs: Placement, Replication, Tags.  No changes made in "Replication" and "Tags" tab => can you please confirm?
* Can you share a screenshot of the "Placement" screen please?

[jchen522]

1. Correct. Two nimbles, upstream A and downstream B

2. There is only one vCenter.  Both upstream and downstream Nimbles are registered to the same vCenter.  Both shows up in the drop down list when creating VM Storage policy.

3. Yes, There is a vvol datastore A from upstream A mounted on at least one ESXi host in vCenter

4. Correct. Creating VM Storage policy in vCenter with replication partner selected in the drop down, datastore A shows in the incompatible list.

No changes made in the Replicaiton and Tags tab.  Replication has custom NimbleStorgae.Replication with VMware managed replication set to Yes.  If you don't chooose a replication partner, then datastore A shows up as compatible.

Is two vCenters required in order for this to work?

[mamatadesaiNim]

> No changes made in the Replicaiton and Tags tab.  Replication has custom NimbleStorgae.Replication with VMware managed replication set to Yes. 

So - this is an anomaly.  Can you create a new VM Storage Profile? The "replication" tab should list "disabled" by default.  

If you select NimbleStorage.Replication and set it to VMware Managed replication as "Yes", you are choosing to use what is known as "VASA3 replication" for the vvol VMs and recovery for that needs SRM (or some PowerCLI).

Let's set the intent here as well, while we are at it: what is your intention of recovery of these vvol VMs, that is, how do you intend to recover these VMs in case of a disaster and you lose primary array A?

[mamatadesaiNim]

Another thing to fix (reviewed your screenshot again): replication frequency is set to 0.  Click the little "i" there, this needs to be non-zero for replication to happen.

[jchen522]

Thanks for the quick assistance.  The replicaiton tab is disabled by default.  I thought you meant the only changes.

I have been creating new vm storage profile.

I tested it again.  I must have remembered it wrong.  When Vmware managed replication is either yes or no and not disabled, it will always be imcompatible. 

If I leave Replication disabled instead of choosing yes or no for VMwware managed replication, choosing a partner will be imcompatible, not choosing one will show compatible.

That is the part I am trying to find out how it work, be able to recover the VM similiar to what I am doing for the VMFS recovery but simpler

In case we lose primary Array A, we should be able to present the replicated VVol on Array B to the ESXs?

[mamatadesaiNim]

> That is the part I am trying to find out how it work, be able to recover the VM similiar to what I am doing for the VMFS recovery but simpler

Can you please explain - how do you perform VMFS disaster recovery?

In VM Storage Profile --> Replication tab --> if you change this, it means you intend to use VASA3 vvol recovery, using SRM or PowerCLI for your VVol VMs.

[jchen522]

If not using Veeam or SRM, assuming primary is down, I would promote the replicated volume and present to ESX and resignature the datastore to mount, manually add each of the VMs.  If using Veeam or SRM, all these would steps would have be automated.  However, I have only used VMFS but not VVol in the past.

If we leave Replication tab disabled, how would all this work?

[mamatadesaiNim]

For SRM with vvols, you would need that replication tab selected, vmware managed replication set to yes.  I have a series of blogs on how to setup vvol DR with SRM, they are a bit old but all this stuff has been working for a few years with Nimble vvols + SRM now  I shall share the links once we get your original problem resolved, make sure your datastore shows up as "compatible".

If you leave the replication setting unset, there is some support built into vcenter plugin to recover your deleted VMs from primary array on the secondary array.  vcenter plugin UI will show those vvol VMs deleted and give you an option to recover.  best to try this out to see if it works.

coming back to the original problem: let's go ahead and create a new vm storage profile on vcenter-A, select "replicate to" target array from drop-down as array-B, make sure "replicate every" is set to 1 or higher number.  vvol datastore-A should show up as compatible.  don't make any changes on the "replication" tab of the storage profile yet.  LMK how it goes please.

[jchen522]

It shows up compatible now after I changed both replication frequency and snapshots to retain on replica to more than 0.  I thought I tried that combination before.

So the next thing I tried to do is recover locallyy first.  First by revert to eariler snapshot and also try dismount vvol database locally and added the datastore back in by creating it and adding the vmx files back to inventory.  Is that the correct way to do it?

However, when I handover the snapshot collection to the DR site, the replicated VVol volume does not show up when I try to create datastore? What is the correct way of doing this?  I also don't see it show up vvol VMs deleted like you mentioned.  Usually I manually remove the VM from inventory, dismount the datastore and handover, maybe I shouldn't have done that and should test it like a real disaster scenario?

Really appreciate your assistance!!!

[mamatadesaiNim · Accepted solution]

>  It shows up compatible now after I changed both replication frequency and snapshots to retain on replica to more than 0.  I thought I tried that combination before.
Good, thanks for confirming that.   I guess the original problem reported on this post is resolved then!

With this config of vvol, there is some amount of DR built into vCenter plugin. e.g. delete a VM on the upstream.  it can be undeleted for next 72 hrs.  or you can recover on the downstream.  no manual steps like unmount datastore etc. needed. 

To do real DR, you can configure SRM with vvols.  this is a 3-part blog series I had written - hope you find them useful. Try it when you get a chance, and maybe post a new topic when you have questions about this.

• https://community.hpe.com/t5/hpe-storage-tech-insiders/vsphere-7-0-and-hpe-nimble-storage/ba-p/7087675
• https://community.hpe.com/t5/Around-the-Storage-Block/Setup-for-SRM-8-3-and-HPE-Nimble-Storage-vVols/ba-p/7092272
• https://community.hpe.com/t5/Around-the-Storage-Block/Run-SRM-recovery-plans-for-vVol-VMs/ba-p/7092994

[jchen522]

Thank you for the info!!!

I tested deleting a VM and try to recover it.  Using vCenter plugin, I can recover the VM locally, but I get an error message if I try to claim the replicated VM downstream.  It says "Failed to claim VM XXXXXXX.  Unable to find a valid host system."  What is the cause of this error message?

Also, I found couple PowerCLI article and script to test failover.  When I execute "Get-SpbmReplicationGroup -vm XXXXXXX", it returns nothing every though I have VM storage policy with replication configured assigned to the VVol volume that the VM resides on.  What could be the cause of this?

[mamatadesaiNim]

Can you please start a new discussion topic?  I think your original topic is resolved, and this can become a confusing discussion thread.