What replication is

Replication is one of two mechanisms used by StorageGRID Webscale to store object data. When StorageGRID Webscale matches objects to an ILM rule that is configured to create replicated copies, the system creates exact copies of object data and stores the copies on Storage Nodes or Archive Nodes.

When you configure an ILM rule to create replicated copies, you specify how many copies should be created, where those copies should be placed, and how long the copies should be stored at each location.

In the following example, the ILM rule specifies that two replicated copies of each object be placed in a storage pool that contains three Storage Nodes.


Make 2 Copies rule

When StorageGRID Webscale matches objects to this rule, it creates two copies of the object, placing each copy on a different Storage Node in the storage pool. The two copies might be placed on any two of the three available Storage Nodes. In this case, the rule placed object copies on Storage Nodes 2 and 3. Because there are two copies, the object can be retrieved if any of the nodes in the storage pool fails.

While placing replicated copies on different Storage Nodes in a single storage pool works well for deployments that have only one site, using a single storage pool does not provide site-loss protection in deployments that have multiple sites. Because the system uses any available nodes in the storage pool when it places the replicated copies, it might place all copies of some objects within only one of the sites.

The following example illustrates what can happen if a single storage pool contains Storage Nodes from two sites. When placing replicated copies, the system stores two copies of object AAA on Storage Nodes at Site 1, and two copies of object CCC on Storage Nodes at Site 2. Only object BBB is protected if one of the sites fails or becomes inaccessible.


Make 2 Copies rule at two sites but only one storage pool

Using storage pools for cross-site replication

If your StorageGRID Webscale deployment includes more than one site, you can enable site-loss protection by distributing copies of object data to storage pools that are at different sites. For example, if you configure an ILM rule to make two copies and select storage pools at two sites, one copy of each object will be placed at each site. If you configure a rule to make two copies and select three storage pools, the copies are distributed to keep disk usage among the pools balanced, while ensuring that the two copies are stored at different sites.

The following example illustrates how objects are stored when you use multiple storage pools. In the example, the ILM rule specifies that two replicated copies of each object be created, and that the copies be distributed to two storage pools. Each storage pool contains all Storage Nodes at one site. Because a copy of each object is stored at each site, object data is protected from site failure or inaccessibility.


Make 2 Copies rule at two sites and two storage pools

When using multiple storage pools, keep the following rules in mind:
  • If you are creating n copies, you must add n or more pools. For example, if a rule is configured to make three copies, you must specify three or more storage pools.
  • If the number of copies equals the number of storage pools, one copy of the object is stored in each storage pool.
  • If the number of copies is less than the number of storage pools, the system distributes the copies to keep disk usage among the pools balanced and to ensure that two or more copies are not stored in the same storage pool.
  • If the storage pools overlap (contain the same Storage Nodes), all copies of the object might be saved at only one site. You must ensure that the selected storage pools do not contain the same Storage Nodes.