NFS storage pools
Each NFS storage pool contains one or more mount points that are associated with specific NFS storage components. Each NFS storage pool includes all of the storage that’s accessed using the NFS mount points included in the pool.
To enable HCP to access the storage that’s represented by an NFS storage pool, when you create that component, you specify the following information:
- The storage pool name.
- Optionally, a description of the pool.
- Whether you want HCP to compress object data that’s stored on the storage that’s accessed using the NFS mount points in the storage pool.
- Whether you want HCP to encrypt object data that’s stored on the storage that’s accessed using the NFS mount points in the storage pool. If encryption is disabled for the system, this option is not visible.
- For each NFS mount point you want to include in the storage pool:
- The name of the NFS storage component that represents the physical storage device that’s accessed using the NFS mount point
- The full path name of the directory that you want to access using the NFS mount point
An NFS storage pool is compliant only if all of the NFS mount points in the pool are associated with compliant NFS storage components.
When you add an NFS mount point to a new or existing NFS storage pool, HCP creates that mount point and mounts the applicable storage volume (called an NFS volume or an external storage volume) on a node in the HCP system. HCP then adds that NFS volume to the NFS storage pool.
HCP uses a round-robin algorithm to determine which node to associate with each new NFS volume that’s added to an NFS storage pool. This method of assigning NFS volumes to the nodes in the HCP system ensures that the volumes are distributed evenly among the nodes.
If the node with which an NFS storage volume is associated becomes unavailable, that volume also becomes unavailable. HCP does not reassign the volume to a different node. When the node returns to service, the volume becomes available again.
In the HCP System Management Console, you can use the Hardware and Storage Node pages to view information about the NFS storage volumes (called external storage volumes) that are associated with each node in the HCP system.
Considerations for using NFS volumes
These considerations apply to using NFS volumes with HCP:
- HCP can use multiple NFS shares from a single device that’s represented by an NFS storage component. Keep in mind, however, that the larger the number of shares HCP uses, the greater the I/O load on the device.
- Typically, you specify export options for a share according to the standards for your site. However, if HCP is unable to mount the extended storage volume that you created as an NFS share, you may need to change the export options. After changing the export options, you need to export the NFS share again.
- For each NFS mount point that’s associated with an NFS storage component, you can specify more mount options than the required ones. You might do this, for example, to set the network block size for read or write requests to the optimal size for the storage device that’s represented by the NFS storage component. However, if HCP is unable to mount the extended storage volume that you created as an NFS share, you may need to change the additional mount options that you specified.
- If the share for an NFS volume becomes unavailable (for example, because the extended storage device that’s hosting the share is inaccessible), HCP tries periodically to remount the volume. If, after the share is available again, the remount fails, you can try to manually remount the NFS volume.