Overview of Resource Allocation

StorNext’s File System Expansion feature enables you to dynamically add LUNs to a selected file system without interrupting that file system’s operation.

The only disruption that occurs during File System Expansion is a short pause of new metadata requests as StorNext updates its internal system and clients to be aware of the new overall capacity and physical disk resources that are used. The short pause is a result of taking the file system offline to use the command cvupdatefs to add a new stripe group.

Beginning with StorNext 6, the command sgadd allows the file system to remain online.

Note: While the file system remains online, clients are notified to re-connect and there may be a slight pause in the order of a small fraction of a second.

File System Expansion is often done in conjunction with the Stripe Group Movement feature. That is, you might want to add new stripe groups knowing you will want to use those stripe groups for Stripe Group Movement.

Note: The command sgadd only adds data stripe groups. You can use the StorNext GUI, the command cvupdatefs, or the command cvmkfs –r (-e) to add a meta data stripe group.

Stripe Group Movement moves data files off one or more data stripe groups onto the remaining data stripe groups in a file system, which frees data LUNS so they can be decommissioned or reused. In a similar way, the metadata on a single LUN can be moved to a new LUN. StorNext provides a Movement Wizard to simplify these processes, which is launched when you select Migrate Data from the Tools menu.

Note: Stripe Group Management utilities currently run on Linux only clients, not just the MDCs.

During data stripe-group movement, you indicate one or more source stripe groups from which to move data. StorNext automatically moves all data from the source stripe groups to the remaining stripe groups in the file system. All other data stripe groups are targets, allowing an even distribution of data across remaining disk resources. During movement, the file system is online and read/write operations occur normally, but the source data stripe group(s) are in read-only mode (write disabled).

Beginning with StorNext 6, the source data stripe group(s) are no longer in read-only mode when moving data off of a stripe group. There is a concept of alloc true and alloc false. When alloc is set to true, new data extents are created on a stripe group, which is the normal case. When preparing for offload, alloc is to false. This prevents new extents from being placed on the stripe group or for existing extents on that stripe group to be expanded. Since the source data stripe group(s) are not in read-only mode, write-in-place applications continue to run.

Beginning with StorNext 6, disks (and disk labels) can be completely removed from the configuration. There is a concept of a vacant stripe group. When a stripe group is vacated, it has no disks and serves as a placeholder so that existing stripe group ordinals do not get changed. The next time a stripe group is added, the system looks for for a vacant stripe group and uses it first.

Beginning with StorNext 6, you can add a stripe group to all file systems, including the HA shared file system using the stripe group management utilities. With the stripe group management utilities, the file system remains on-line throughout the procedure.

For additional information, see the following topics:

• Perform File System Actions
• Tools > File Systems > Stripe Group Actions

Note: Beginning with StorNext 6, it is no longer necessary to put an HA system into config mode to enact the change to the HA shared file system.

Below are high-level steps required for expanding a file system and moving stripe groups.

1. Check the file system before you begin. See Check File System.
2. Expand the file system. See File Systems OverviewFile Systems Overview.
3. Move data stripe groups or metadata/journal stripe groups. See Tools > File Systems > Migrate Data.
4. Mark source stripe groups as read-only.
5. Reboot all clients after the expansion.