The Design and Implementation of a Log-Structured File System
-
Why the paper says the disk operation will be dominated by writes?
-
What are pros and cons of a large file buffer?
-
For a small file focused workload, what is the performance characteristics?
-
Compare the inode discoery mechanisms of BFS and LFS
-
given we have dir1/file1, and dir2/file2 next to each other on our log, what will the log look like?
6.describes segment cleaning up process
7.How does LFS decide if a block is live?
8.Describe the segment usage table
-
How does LFS recover when the checkpointing op itself failed?
-
what are the benefits of having inode block AFTER data blocks?
-
How does LFS recover from the case where inode and directory info are not consistent?
-
Why LFS does not support fuzzy checkpoints, unlike ARIES and its WAL?
-
Why LFS wants a bimodal distribution of segment utilization?
-
explain the cost-benefit policy and why it helps achieve the bimodal distribution of utilization?