Create the IBM Storage Scale file system and pools

After you create NSDs, the IBM Storage Scale file system can be created.

To use FPO, a single file system is recommended. The following example creates a file system with mount point /mnt/gpfs that is set to auto mount. This mount point is used in Hadoop configuration later. The replication for both data and metadata is set to 3 replicas. Quotas are not activated on this file system. An inode size of 4096 is recommended for typical MapReduce data sizes \-S and -E settings help improve performance for mtime and atime updates. The mmcrfs command also creates IBM Storage Scale storage pools based on the disk file %pools setting.

# mmcrfs gpfs-fpo-fs -F diskfile -T /mnt/gpfs -n 32 -m 3 -M 3 -r 3 -R 3 -i 4096 -A yes -Q no -S relatime -E no [-v no]

For more information on the pool configuration, see Create IBM Storage Scale Network Shared Disks (NSD).

Mount the file system on all nodes:

 mmmount all -a

Use the mmlsfs command to display the file system configuration:

 mmlsfs all

Use the mmlsdisk command to display the status of the NSDs:

 mmlsdisk gpfs-fpo-fs –L

Use the mmdf command to view the disk usage for the file system:

 mmdf gpfs-fpo-fs

Use the mmlspool command to view the storage pools:

 mmlspool gpfs-fpo-fs all –L