Application Persistent Storage Tab

The Application Persistent Storage tab of the Systems Settings screen (see The System Settings Screen) enables Platform Administrators to connect to an external storage pool that will be used to store crucial container folders and enable migrating containers between hosts. You may create, expand, and shrink this external storage resource just like any other storage resource.

This topic applies to HPE Ezmeral Container Platform deployments that implement EPIC only.

All of the following criteria must be met in order to migrate containers from host to host:

• You must configure an external storage resource for use.
• The configured external storage resource must be mapped on this tab for use as a persistent storage pool.
• One or more flavors must be created that include a specified amount of persistent storage, in GB. See Creating a New Flavor.
• The selected external persistent storage resource must have enough free space to support the flavor-defined per-container persistent storage allocation times the number of containers using that flavor. For example, if Flavor_A specifies 20GB of persistent storage and there are 25 containers using that flavor while Flavor_B specifies 30GB of persistent storage and there are 15 containers using that flavor, then you must specify a persistent storage pool that has at least 950GB of free space, which is (20*25)+(30*15)=500+450=950.
• Containers must be created using a flavor with an allocated persistent storage amount. See Creating a New Cluster, Creating a New Training Cluster, Creating a New Notebook Cluster, or Creating a New Deployment Cluster.
• The deployment must have enough resources to support the migration, including all placement constraints.

See the following for more information about migrating containers:

• General information: See Stateful Container Migration.
• To migrate individual containers: See Viewing and Migrating Virtual Nodes.
• To migrate all of the containers from a specified host: See The EPIC Hosts Installation Screen.

To map an external storage pool for use as persistent storage:

1. Use the Select Type pull-down menu to select the filesystem used by the external resource, and then enter the appropriate parameters based on your selection. The available options are:
   • None: Select this option to disable persistent storage.
   • Local MapR: This option is only available for on-premises deployments. See Local MapR.
   • CEPH RBD: This option is only available for on-premises deployments. See CEPH RBD.
   • NFS: This option is only available for on-premises deployments. See NFS.
   • ScaleIO: This option is only available for on-premises deployments. See ScaleIO.
   CAUTION:
   Changing the persistent storage pool settings will cause containers that are currently using persistent storage to become ineligible for migration.
2. Click Submit to save your changes.

Local MapR

If you selected Local MapR using the Select Type pull-down menu, then you may either accept the defaults or edit as needed:

• Mount Path: /opt/bluedata/mnt
• Volume Path: /hcp/pers_stor
• Num replicas: 1

CEPH RBD

If you selected CEPH RBD using the Select Type pull-down menu, then enter the following information:

• Monitors: Comma-separated list of the IP addresses or hostnames of the CEPH RBD monitors.
• Pool Name: Name of the CEPH RBD pool that was created for use by HPE Ezmeral Container Platform.
• Client Name: Client name, if CEPHX authentication has been enabled on the persistent storage pool.
• Client Key: Client key associated with the client name, if CEPHX authentication has been enabled on the persistent storage pool.

See Working with CEPH Pools and Users for additional information about basic CEPH operations.

NFS

If you selected NFS using the Select Type pull-down menu, then enter the following information:

• Server: IP address or hostname of the NFS server.
• Share: NFS share.

ScaleIO

Before selecting Scale IO using the Select Type pull-down menu, verify that the ScaleIO client (SDC) is properly installed on all HPE Ezmeral Container Platform hosts, as described in the EMC Scale IO Installation Guide (link opens an external website in a new browser tab/window).

If any host is running RHEL/CentOS 7.5, then apply the following workaround on the affected hosts:

1. Edit /etc/default/grub by adding nokaslr at the end of the options line in the GRUB_CMDLINE_LINUX entry.
2. Execute the command grub-2-mkconfig -o /boot/grub2/grub.cfg on each affected host.
3. Reboot the affected hosts.

Once you have verified proper configuration, enter the following information in the Application Persistent Storage tab:

• Gateway: IP address or hostname of the ScaleIO gateway.
• Port: Port number on the ScaleIO gateway.
• Protection Domain: ScaleIO protection domain.
• Storage Pool: Name of the storage pool.
• Storage Mode: Use the pull-down menu to select either Thin Provisioned or Thick Provisioned, as appropriate.
• MDM Username: ScaleIO MDM username.
• MDM Password: ScaleIO MDM password.