Skip to main content

Section 5.6 Log-structures file systems

We now look at a considerably different file system approach. Read the intro to Log-structures File Systems chapter 43 for an explanation of the motivation behind the system.
Read sections 43.1, 43.2 and 43.3 about the basic ideas behind a file system with only sequential writes and the technique of write-buffering.

Practice 5.6.1.

    True or False: write buffering is the process of pooling a number of file system updates in memory then writing them all at once.
  • True.

  • False.

Practice 5.6.2.

A segment is:
  • The chunk of changes that are collectively written to the disk at once.
  • The small part of the changes that are to be written to the disk, that corresponds to a single change like a new file.
  • A particular range in memory where the changes to be written are kept.
Read sections 43.4, 43.5, 43.6 about how inodes are managed in such a system.

Practice 5.6.3.

    True or False: The inode actual location on disk never has to change in a log-structured system. Inode number 5 is always found in the same place every time.
  • True.

  • False.

Practice 5.6.4.

    True or False: The inode map associates inode numbers to their most recent actual physical location on disk.
  • True.

  • False.

Read sections 43.7, 43.8 about what is involved to perform a file read from the LFS system.
Read sections 43.9, 43.10, 43.11 about the problem of recovering out of date portions of the file system effectively and its solution.

Practice 5.6.5.

    True or False: When recovering no-longer-usable parts, the LFS cleaner does so in whole segments, re-writing in a new segment any data that is still needed.
  • True.

  • False.

Read section 43.12 about how crash recovery works in LFS, and section 43.13 for a brief summary.
You have attempted of activities on this page.