| 1: | What are the two kinds of cluster storage designs? |
| A1: | Answer: Shared everything and shared nothing |
| 2: | Why is locking more important for a cluster file system than a traditional file system? |
| A2: | Answer: The cluster file system can have multiple systems trying to access the same data. |
| 3: | True or false: You typically do not have to worry about data in host buffers with a cluster file system. |
| A3: | Answer: False. You have to account for all cached data. |
| 4: | Why are SANs valuable for cluster file systems? |
| A4: | Answer: They are the easiest way to connect all systems to all storage. |
| 5: | True or false: NAS clusters are used primarily to increase storage capacity. |
| A5: | Answer: False. They are used primarily for high availability. |
| 6: | What is the main difference between a cluster file system and a DFS farm? |
| A6: | Answer: There is no failover in a DFSfarm. |
| 7: | What are two ways to implement a lock manager for a distributed file system? |
| A7: | Answer: Distributed or centralized |
| 8: | What is the purpose of direct I/O? |
| A8: | Answer: To prevent host systems from caching data for database applications |
| 9: | Database logs are an example of what kind of redundancy technique? |
| A9: | Answer: Delta or difference redundancy |