D-ECS-OE-23 Sample Questions & Answers

ECS provides distinct quota controls for both storage capacity (size) and object count. To limit the total number of objects within a namespace, the administrator must configure the 'Quota on Number of Objects' setting. The 'Block Size Quota' or 'Notification Quota' relates to storage capacity.

Integrating ECS with Active Directory involves two steps: authentication and authorization. While the user can authenticate successfully, their AD group membership must be explicitly mapped to an ECS management role (like Namespace Admin or System Admin) within the ECS portal. Without this mapping, the authenticated user has no rights within ECS, resulting in 'permission denied' errors.

The 'Data Under Replication' metric indicates the amount of data that has been ingested but has not yet been fully protected by erasure coding or replication across the required number of nodes. A consistently high value points to a slowdown in the background process that distributes and protects data chunks. This process relies heavily on the high-speed inter-node private network. A bottleneck, high latency, or failures on this network would directly impact the system's ability to protect data promptly.

When creating or editing an object user in the ECS Management Portal, there is a specific and distinct option labeled 'Set OpenStack Swift Password'. The administrator must check this box and then enter the desired password. This is separate from the S3 secret key generation and is the required method for enabling Swift API authentication for that user.

Data Domain's Retention Lock feature, when used with an ECS target, ensures that tiered data is immutable for a specified period. When files are moved to the ECS Cloud Tier from a Retention Lock-enabled MTree, the immutability settings are preserved on the objects within ECS. This combination provides an end-to-end WORM solution that is essential for regulatory compliance and proving data integrity.

The core trade-off is between storage efficiency and computational cost. 3x replication is simple but has a high storage overhead (200%). 12+4 erasure coding is much more space-efficient (33% overhead) but requires significant CPU cycles to calculate the parity chunks during writes and to reconstruct data during reads or recovery. Therefore, replication uses more disk space, while erasure coding uses more CPU.

When an object is written to ECS, it is first divided into smaller, 128 MB segments. Each segment is then further divided into 2 MB chunks. This chunking process is fundamental to how ECS protects and distributes data. The chunks are then processed by the storage engine to be erasure-coded or replicated and spread across multiple nodes and disks in the VDC.

The S3 REST API provides a standardized, programmatic way to interact with ECS. The GET Service operation (listing buckets) is the standard S3 API call for this purpose. When the call is authenticated with credentials tied to a specific namespace, it will return the list of buckets within that namespace. The Management Portal is a GUI, and while the ECS CLI could be used, the REST API is the most direct and universally compatible interface for application integration.

This solution meets all requirements. 1. Geo-replication via the replication group protects against a single site failure. 2. Native replication ensures data is available in London for local reads. 3. Erasure coding is ideal for large files. 4. Using 12+4 erasure coding at both sites provides excellent storage efficiency compared to replication.

ARM policies can be applied directly to individual users.

ARM policies can be applied to groups of users.

ARM policies are defined at the bucket level to control access.