Drive Setup - Migrating from VmWare

[Tim Poulter]

So I am in the process of testing the HPE VM Essentials to get more familiar with the install and had a question regarding the drive setup.

Currently we have 2 drives on a raid card, 1 drive is a mirrored 147GB for the OS (previously ESXI) and a 2nd drive in a Raid5 or Raid6 for the Datastore where the VMs live.

While following this https://support.hpe.com/hpesc/public/docDisplay?docId=sd00007332en_us&page=GUID-28F18596-4902-4CD1-83F3-1411430C5534.html, it says to leave the storeage as is but obviously its not how we want it to be configured and I just wanted a second pair of eyes to verify my setup

I selected my 136GB Drive and said Use As Boot Drive

I selected my 2TB drive and mounted it as / and formatted it as BTRFS (or would EXT4 be better)

[CalvinZito]

I have a notebook set up with the HPE Morpheus VM Essentials Software Documentation v9.0.0. I HIGHLY recommend that you use the documentation. I haven't used the AI that is on our Support Center but if you go to the document, there's a search and AI search bar at the top. There's also a Deployment Guide that consoldates the release notes, solution overview, qualification matrix, the Software Documentation and more. Here's what my notebook gave me in response to your question:

Your disk selection looks reasonable for a lab/test deployment.

The main thing I'd change is to use the smaller mirrored RAID1 volume as the boot/OS device and avoid mounting the larger datastore volume as / unless you specifically want the entire OS and VM storage to reside on that filesystem.

For the filesystem choice, I would generally recommend BTRFS, as that is what the installer defaults to and what most HVM deployment examples use. I am not aware of any documented advantage to choosing EXT4 for a standard VM Essentials deployment.

Is this the Manager Appliance or are you building an HVM host that will be part of the cluster? If this is an HVM host, I'd personally keep the OS on the 136 GB mirror and reserve the larger RAID volume for datastore use rather than consuming it as the root filesystem. That more closely matches how most virtualization hosts are typically laid out. 

[Arnout_Verbeken]

As Calvin said:
- mount your 136GB drive to /
- mount your 2TB as /var

Why /var ?  Because that is the default location for the "local datastore" that is automatically created when creating an HVM cluster.  By doing this, you never risk that your root filesystem will be out of space since virtual images, backups, vms (on local datastore), image targets (caching)... are all stored by default under /var

Alternatively, you can mount your 2TB to something else like /DATA (and leave /var on your 136GB).
But this means you will have to create yourself a "local datastore" in the cluster.  You will also have to create a "File Share" (can be that same /DATA) for Virtual Images and Backups
You can also select that new "local datastore" as Image Target (caching).

The link you provided assumes a single disk.

I personally still use ext4, which seems to work fine.

[Tim Poulter]

After testing it some more and getting further into the setup this is what I have been trying out, which is pretty much what was suggested by both Calvin and yourself

- mount your 136GB drive to / 
- mount your 2TB as /var

When I had maded the post it did incorrectly say to mount the 2tb as / which after a 2nd install I had realized was wrong. So I had mounted it as /datastore2 and then once installed I did mount it as a local datastore 

I can test it as /var if it automatically mounts that during install as the "local datastore" for vms and such

A second question is when creating the boot drive it only selects about half of the space. Would it be a good idea to 

1 - expand it to use all of that space  

or 

2 - mount the remaining space or some how allocate it to the hpemananger so I can store my files / iso files(the Library - Virtual Images) on there? or would it be best to store them on the main 2TB "local datastore /  /var" location

Just coming from Vmware the way we did it was the 136GB was the boot os / iso storage and the 2nd drive was vm storeage. just trying to see what is best with the new way that HPE VME does it

[Arnout_Verbeken]

Also here it depends on the amount of disks you have available.

If you have only 1 disk, you can mount (but not should) the available free space on /var

If you have 2 disks, you should expand the root disk to use the full size and you can mount the 2nd disk to /var (as explained above).
Or you mount it on a separate mountpount (e.g. /datastore01).  In this case, you will have to add it as a local datastore from within the cluster-storage tab.

If you have only 1 disk, I tend to just expand the rootdisk to max size and not mount /var separately.

For ISO's, Images,...: we do not put them on a datastore, but we add them to a "Virtual Image Repo" from within the manager.  In VME/Morpheus, it can never be added to a datastore.
This "Virtual Image Repo" can be a "local disk inside the manager OS", an NFS or CIFS Share.

My preferred way is to:
- make the manager a "managed VM"
- Reconfigure - add a disk to the VM
- partition + format + automount the disk
- set ownership to morpheus-app
- add as a File Share to be used as default Virtual Image location in Storage tab

[Tim Poulter]

See that is what I am trying to figure out / understand

I have 2 disks 

1 - 137gb mirrored drive, which during the install I select this as the boot drive / host os install. uses about half of it ~63GB

2 - a 2tb drive which is mounted as /datastore2 and I go in the HPE VME Manager - Cluster - Clustername - Storage and added it as Directory Datastore (Local), and set it as image lmage Target. This is where I want to store the VMs

So I am trying to understand how to add the free ~63GB that is left over from the install to the manager so i could use that for the iso images

Also what does make the manager a "managed VM" do?. As this is a single host deployment. Would that make a difference? could not really find anything in the docs, at leas the search did not

[Arnout_Verbeken]

1. By default, your boot OS will indeed not consume the full available space.  You should change this during setup to use the full disk.
See https://support.hpe.com/hpesc/public/docDisplay?docId=sd00007332en_us&page=GUID-28F18596-4902-4CD1-83F3-1411430C5534.html item 6-b and 6-c

You can also expand the lv later on with the following (copied from AI - so doublecheck) procedure:

To expand a Logical Volume (LV) in Ubuntu when unallocated space is already available in your Volume Group, run sudo lvextend -l +100%FREE /dev/mapper/ubuntu--vg-ubuntu--lv followed by sudo resize2fs /dev/mapper/ubuntu--vg-ubuntu--lv. 

Steps to Expand LVM Space

• Check free space: Run sudo vgdisplay to confirm your Volume Group has available free space. (If you just expanded your virtual disk or partition size at the hardware/hypervisor level, run sudo growpart /dev/sda 3 and sudo pvresize /dev/sda3 first to make the space visible to LVM). 

• Extend the Logical Volume: Run the following command to use all remaining free space for your root logical volume (adjust the path if your volume names differ):
  sudo lvextend -l +100%FREE /dev/mapper/ubuntu--vg-ubuntu--lv

• Resize the Filesystem: Expand the filesystem to match the new logical volume size:
  • For ext4 filesystems: sudo resize2fs /dev/mapper/ubuntu--vg-ubuntu--lv
  • For XFS filesystems: sudo xfs_growfs / 

• Verify the expansion: Run df -h to ensure the new storage size is active.
  
  
  2. Image Target is not equal to "datastore for VM's".
  Please read 

  https://community.hpe.com/t5/hpe-morpheus-vm-essentials/disk-space-filling-up-on-host-where-vmessentails-console-is/m-p/7252607#M513 and let us know if still not clear
  
  3. Even with those 63GB extra on your HVM host, you will not be able to use it for ISO images.  The location for those ISO images lives "inside the manager VM", not "inside the HVM host"
  
  4. By default, the manager VM will be a "Discoverd VM".  the manager sees "a VM" but can't do anything with it (as in: no Day2 operations: start/stop/change network/add VM hardware/...).
  If you make the VM a "Managed VM", you can do all of those things directly from the manager.  So the "manager application (VME web UI)" will be able to manage the "manager VM/OS".
  So once it's managed, you can add a 2nd disk to the VM and configure it as your "Virtual Image Rep" to host your ISO's
  It is unrelated to a single-node or multi-node cluster.
  
  
  
  

[Tim Poulter]

I think I have an understading of how to configure my 2nd disk / datastore2 / where I store my VM machines, this is just mounted as a /something and then add it in the cluster view vs the infrastructure -> storage view.

That one seems pretty simple. 

I guess my confusion is coming from what the hosts sees and what the manager sees. 

My question is now this, would it be better to add all the free space to the host at setup like you linked in sections 6-b and 6-c

or would it be better to leave it at what it automatically does(in this case half or less if the host side does not need much) and then add the other half to the manager as local storage? so that I could have room more isos and such (not that we would have alot mind you) 

or mount it on the vm side and add it for other use in the manager under the cluster - clustername - storage

In my scenario we are trying to keep it simple, as they are single host deployments, replacing vmware in small shops, and we are trying to keep it as similar to how our vmware installs are. I do understand that they are different models. So I may be trying to force a vmware view into something that is not designed that way.