1Z0-1072-25 Question 256
Single answerYou have an HPC cluster deployed across two private subnets in the same VCN. The cluster� applications need to continuously read and write large volumes of shared data from Oracle Cloud Infrastructure File Storage. You also want to minimize administrative overhead and the number of mount targets required. Which approach should you take to enable all nodes in both subnets to access the File Storage service?
- A
Create a single File Storage file system and deploy one mount target in a subnet accessible to both HPC subnets, and configure security rules to allow NFS traffic from each subnet.
- B
Set up separate File Storage file systems, each with its own mount target in each subnet, so each subnet� instances have their own dedicated storage.
- C
Use block volumes attached to each node in each subnet and periodically replicate them to keep data synchronized between subnets.
- D
Host the data in an Object Storage bucket and mount it directly to each instance using a third-party FUSE driver.
Show answer and explanation
Correct answer: A
Explanation
When multiple compute instances in different subnets need to share file-based storage, the recommended approach is to create a single File Storage file system with one mount target in a subnet accessible by all instances, then configure security rules to allow NFS traffic from the appropriate subnets. This setup simplifies management while providing high availability, performance, and data consistency. Refer to the Oracle Cloud Infrastructure File Storage documentation for detailed best practices on creating and configuring file systems, mount targets, and network security.
- A. Correct.
Correct: You can create a single File Storage file system and one mount target in a subnet that is routable from both HPC subnets. By properly configuring security lists or network security groups and opening the appropriate NFS ports, all instances in both subnets can mount the shared file system. This approach minimizes administrative overhead and the number of mount targets.
- B. Incorrect.
Incorrect: Creating separate file systems for each subnet increases maintenance complexity and overhead. It is not necessary to maintain multiple file systems if the subnets are in the same VCN and can be configured to access a single mount target within the same AD.
- C. Incorrect.
Incorrect: Block volumes do not provide a managed NFS-like shared storage experience. Replicating block volumes for each instance would significantly increase operational complexity and is not the recommended approach for shared file data among multiple instances.
- D. Incorrect.
Incorrect: While Object Storage can be used for data storage, it is not a direct replacement for shared file systems, particularly for HPC workloads that require standard NFS semantics, high I/O performance, and low-latency access. Mounting an Object Storage bucket with a third-party FUSE driver is also not officially supported and introduces potential reliability and performance issues.