Managing Datastores

Managing Datastores

Datastores are logical containers used by the HX Data Platform to manage your storage usage and storage resources. Datastores are where the host places virtual disk files and other VM files. Datastores hide the specifics of physical storage devices and provide a uniform model for storing VM files.

You can add, refresh the list, edit name and size, delete, mount and unmount datastores from either the HX Connect UI or the HX Data Platform Plug-in UI. You can only rename an unpaired datastore that is unmounted. Renaming the HX Datastore from the vCenter administration interface is not supported, and should not be done.


Important

Do not rename an HX datastore from vCenter. The datastore names shown in HX Connect or Intersight and in the ESXi host datastore (that appears in vCenter) must be identical including case sensitive. If they are not identical, some operations such as expansion, mount/unmount of a datastore will be impacted.


Note

Enabling encryption on your cluster is only possible during the datastore creation procedure. Encryption cannot be disabled for a datastore once enabled.


Important

  • For best start-up and upgrade performance, use the fewest number of datastores as possible.

    The Cisco HyperFlex best practice recommendation is to not exceed 15 number of datastores.

    • The impact of using more than 15 datastores per cluster include:

      • Excessive start-up delay when you perform maintenance work (updates, upgrades and reboots). The start-up delay of each host is linear to the number of datastores created. Each host experiences a 30-second additive delay per datastore created.

      • Timeouts on upgrade.

      • Datastores fail to mount.

  • Keep the number of datastores to as few as possible to avoid start-up delay and to keep clone savings high.

  • HX Native Snapshots are not supported with multiple datastores.

  • If using an M4 node, never use either the HyperFlex NFS or local Springpath datastore for ESXi logging or coredump partition. If using an M5/M6 node, you can use any left over space in the HyperFlex NFS or local Springpath datastore for these purposes.

  • When VMs have flat vmdk files, one with thin provisioned and one with thick provisioned, the total combined storage usage of all flat VMDK files as reported by the vCenter/ESXi and HX Connect could be higher than the Datastore usage itself reported by vCenter & HX Connect. This could be due to ESXi and vCenter space reporting for each VM files ignoring the "uniqueBytes" attributes sent by underlying NFS storage in Extended stats and attributes via VAAI APIs.

  • For Vmware ESXi environments, ensure Storage I/O is disabled for all HyperFlex datastores in the vCenter. This setting is on a per datastore setting, and enabling this can cause unexpected performance impacts.

Procedure

Step 1

Choose an interface.

  • From the vSphere Web Client Navigator, select vCenter Inventory Lists > Cisco HyperFlex Systems > Cisco HX Data Platform > cluster > Manage > Datastores.

  • From HX Connect, select Datastores.

Step 2

Create a new or select an existing datastore, to view options.

  • Create a new datastore
  • Refresh the datastore list
  • Edit the datastore name and size
  • Delete the datastore
  • Mount the datastore on the host
  • Unmount the datastore from the host

Adding Datastores

Datastores are logical containers, similar to file systems, that hide specifics of physical storage and provide a uniform model for storing VM files. You can also use datastores to store ISO images and VM templates.

Procedure

Step 1

Choose an interface.

  • From the vSphere Web Client Navigator, select vCenter Inventory Lists > Cisco HyperFlex Systems > Cisco HX Data Platform > cluster > Manage > Datastores.

  • From HX Connect, select Datastores.

Step 2

Select the create datastore.

Step 3

Enter a name for the datastore. vSphere Web Client enforces a 42 character limit for the datastore name. Assign each datastore a unique name.

Step 4

Specify the datastore size. Choose GB or TB from the drop-down list.

Step 5

Specify the data blocksize. From HX Connect, choose 8K or 4K. Default is 8K. In the HX Data Platform Plug-in, the default is assumed. For VDI workloads, default is 4k.

Step 6

To encrypt your datastore, click the Software Encryption check box.

For more information on enabling Software Encryption on your cluster, see Enabling HyperFlex Software Encryption Workflow.

Step 7

Click OK to accept your changes or Cancel to cancel all changes.

Step 8

Verify the datastore. Click the Refresh icon if needed to display your new datastore.

From HX Data Platform Plug-in, Click the Manage > Datastores > Hosts tab to see the mount status of the new datastore.

If you check the datastore through the vSphere Client application, host > Configuration> Datastores, the Drive Type is listed as Unknown. This is expected vSphere behavior, to list NFS datastores as Unknown.

Editing Datastores

A HX Data Platform datastore can be modified using the edit (pencil) option. Edit options are: 1. Change the datastore name, or 2. Change the datastore storage allocation. That is, the size of the datastore.

Note

Do not rename datastores with controller VMs.

Procedure

Step 1

Choose an interface.

  • From the vSphere Web Client Navigator, select vCenter Inventory Lists > Cisco HyperFlex Systems > Cisco HX Data Platform > cluster > Manage > Datastores.

  • From HX Connect, select Datastores.

Step 2

Select a datastore.

Step 3

Unmount the datastore.

If you are only resizing the datastore, you do not need to unmount the datastore. Skip this step.
Step 4

Click the Edit (pencil icon) datastore.

Step 5

Change the datastore name and/or size, as needed. Click OK.

Step 6

Remount the datastore, if you previously unmounted it.

Mounting Datastores

Prepare to mount a datastore.

  • No VM, template, snapshot, or CD/DVD image resides on the datastore. This is the most common error while unmounting.

  • Storage I/O Control is disabled for the datastore.

  • The datastore is not used for vSphere HA heartbeat.

  • The datastore is not used to host RDM metadata files. RDM is not supported.

  • The datastore is not used as a scratch location.


    Note

    You cannot select an NFS datastore as a destination for the persistent scratch location on ESXi. If you select the HX datastore for the persistent scratch location, it will be removed after the ESXi host reloads.

    For all M5 servers, M.2 boot SSD is automatically selected for use as scratch. This is configured out of the box on any new install.

    For HX240M4 (non-SED), Intel SSD is used for persistent logs/scratch (same applies on 220M5/240M5, but on a different local SSD).

    For HX220M4 and HX240M4 (SED), there is no location to store the scratch partition. So, the only option is to use syslog for persistent logging over the network.

Mount a datastore.

Procedure

Step 1

Choose an interface.

  • From the vSphere Web Client Navigator, select vCenter Inventory Lists > Cisco HyperFlex Systems > Cisco HX Data Platform > cluster > Manage > Datastores.

  • From HX Connect, select Datastores.

Step 2

Select a datastore.

Step 3

Click the Mount.

Step 4

Confirm to mount the datastore, click OK.

Unmounting Datastores

Prepare to unmount a datastore.

  • No VM, template, snapshot, or CD/DVD image resides on the datastore. This is the most common error while unmounting.

  • Storage I/O Control is disabled for the datastore.

  • The datastore is not used for vSphere HA heartbeat.

  • The datastore is not used to host RDM metadata files. RDM is not supported.

  • The datastore is not used as a scratch location.

Unmount a datastore.

Procedure

Step 1

Choose an interface.

  • From the vSphere Web Client Navigator, select vCenter Inventory Lists > Cisco HyperFlex Systems > Cisco HX Data Platform > cluster > Manage > Datastores.

  • From HX Connect, select Datastores.

Step 2

Select a datastore.

Step 3

Click the Unmount.

Step 4

Confirm to unmount the datastore, click OK.

Step 5

If needed, recover from partial unmounts.

  1. Go through the above checklist and unmount or delete through one of the UIs or CLI again.

  2. Use the UI or CLI to re-mount the datastore.

For additional information on recovering from partial unmounts, see Recovering from Partially Unmounted Datastores.

Deleting Datastores

Prepare to delete the datastores.

  • Power off all VMs.

  • Close all open shells on the datastore mount point.

  • Disable HA on the datastore.

  • Close all applications that use the datastore.

Delete datastores.

Procedure

Step 1

Choose an interface.

  • From the vSphere Web Client Navigator, select vCenter Inventory Lists > Cisco HyperFlex Systems > Cisco HX Data Platform > cluster > Manage > Datastores.

  • From HX Connect, select Datastores.

Step 2

Select a datastore.

Step 3

Click Delete.

Step 4

Confirm to delete the datastore, click OK.

Recovering from Partially Unmounted Datastores

When mounting, unmounting, or deleting datastores, sometimes a datastore can become partially unmounted. If this occurs, complete the following as needed.

Procedure

Step 1

Depending upon the task you are attempting, complete the items in Prepare to mount a datastore, Prepare to unmount a datastore, or Prepare to delete the datastores.

Step 2

Retry to mount, unmount, or delete the datastore through the HX Connect or HX Data Platform Plug-in UI or CLI again.

Step 3

If the datastore is not in the desire mount, unmount, or deleted state, complete the following.

  1. Ensure VMs are not running on the datastore.

  2. From ESX host, check to see if the HX Data Platform datastore is being used by VMware service, storageRM.

    # ls -ltra /vmfs/volumes/stfs-ds1/ | grep -i iorm

    Sample response 
    
-rwxr-xr-x 1 root root 16511 Jan 20 20:05 .iormstats.sf 
drwxr-xr-x 1 root root 1125 Jan 20 20:06 .iorm.sf

  3. Check the storagerm status.

    # /etc/init.d/storageRM status

    Sample response

    storageRM is running

  4. Stop the storagerm service.

    # /etc/init.d/storageRM stop

    Sample response 
    
watchdog-storageRM: Terminating watchdog process with PID 34096 
storageRM stopped

  5. Try to mount, unmount, or delete the datastore again.

  6. This is one possible solution, if this doesn't resolve the issue, contact Technical Assistance Center (TAC).