HPE VM Essentials Manager Redundancy

[dya]

I can't find anything about manager redundancy for HPE VM Essentials, even after checking the official documentation.

What should I do?

*This is a machine translation. Sorry if it's hard to understand.

[DiegoDelgado]

Hello,

VM Essentials Manager supports HA so if the server it's executing in fails, the cluster will power it on in a surviving node after a couple of minutes if setup right. Is this what you're looking for?

[YA1007]

I tried to set it up by looking at the following document.

https://hpevm-docs.morpheusdata.com/en/latest/vme_getting_started/vme_getting_started.html#initial-manager-setup

I prepared a separate storage and connected it via iscsi, and was able to create a virtual machine,
live migration, HA, etc., and check the operation.

However, I understand that if I set it up according to the document,

the VME manager will be installed on one of the hosts. If that happens, I think that if the host on which it is installed breaks down, the VME manager will also become unusable.

To prevent this from happening, I would like to ensure redundancy for the VME manager as well.

[YA1007]

The previous post was made by the questioner (dya). I was logged in with a different account, so I thought it might be confusing, so I'll explain it here.

[DiegoDelgado · Accepted solution]

Got it.

If you have external storage, you can migrate the VM Essentials manager VM to it and set it up for HA so the VM turns itself on again on host failure. Just follow the following steps:

• Check the amount of resources assigned to the VM Essentials Manager VM, if those resources align with a plan in VME, you're fine, else, go to Administration > Plans and create a plan with the needed CPUs and Memory (Pro tip: if you used the Small deployment, you're fine as it uses 2CPUs and 12G of RAM)
• Go to Compute > VMs and find your manager VM, then convert it to managed (Use the right plan and set the user you setup on deployment for agent installation)
• When the Manager shows as an Instance, go to it and select Reconfigure from the Actions menu. Select the external datastore and the VM will get moved to shared storage

One last step you need for HA to work is editing your datastore and set it for Heartbeats so the VM actually powers on after node failure

Hope this helps!

[YA1007]

• Go to Compute > VMs and find your manager VM, then convert it to managed (Use the right plan and set the user you setup on deployment for agent installation)
  →Is this possible even while the VM is running? Does this mean that storage migration, as it is called in VMware, is possible?
  
  
• One last step you need for HA to work is editing your datastore and set it for Heartbeats so the VM actually powers on after node failure
  → I tried to select "Edit Datastore", but the only settings I could find were for group access. Where do I set these?

[Moto_PK]

To make the management appliance highly available, we need to reconfigure “management appliance” from “local storage to a  “Shared DataStore”.

(Click on Cluster node where the Management appliance is running >  Click on Discovered VM > Go to “Actions”  > Click on “Reconfigure” to modify the “Datastore” to Shared volume)

Once reconfigure is completed, you may Convert the “management appliance” from Discovered to “Managed”

( Infrastructure > Compute > Discovered VMs > select vme manager and from the ACTIONS menu select “Convert to Managed”)

[cbunge]

"Convert to Managed" is essentially assigning a VM ownership inside of the platform.  I.E - the group assignment.  At that time you can also assign a set of credentials and optionally install the agent for added benefits of OS level health checks and automation enhancements.  This can be while running safely.

As for the missing "Heartbeat" option only appears on shared Filesystem type datastore.  I.E - CephFS, GFS, or NFS.  If you are utilizing Ceph, an example would be the default "morpheus-images" location.

[dya]

I reconfigured the detected VME Manager and changed it to a GFS2 datastore, and although I tried multiple times there were no errors, but nothing changed.

The disk path does not change when viewed with virsh dumpxml.
"/var/morpheus/kvm/config/morpheus-vm.xml" does not change either.

Even if I try to investigate, it is difficult at the moment because I don't know where to look to find out what.

[PSPrakash]

Can you first perfrom "Convert to Managed"  and then Reconfigure  the Manager to use the Shared DataStore. !

[dya]

It says "Resizing" but the status doesn't change.

[dya]

It was done.
The reconfiguration operation must be done from [Provisioning] > [Instances]. I did it from [Infrastructure] > [Compute].

Thank you.