Why does LFS update an inode by writing a new copy to the log rather than overwriting the inode in place (as FFS does)?

If two containers are both started from the same 300 MB image, the second container requires roughly another 300 MB of disk space because each container gets its own full copy of the image.

On a hard disk, reading two sectors that are adjacent on the same track is typically much faster than reading two sectors that are on different tracks and far apart on the disk.

Which of the following structures is NOT stored per cylinder group in FFS?

Consider the following commands sent to a UNIX-based system (assume we start with an empty directory): Line 1: echo hello >> file1 Line 2: ln file1 file2 Line 3: cat file2 Line 4: rm file1 Line 5: cat file2 Line 6: echo more >> file1 Line 7: cat file1 Ignore any problems with whitespace or capitalization. Assume “;” shows text separated across multiple lines in the shown output.

A process calls open(“/foo/bar”) on a freshly mounted file system (nothing is cached). Counting only disk reads (not writes), how many disk reads does the file system perform to complete the open?

FSCK restores a file system to the logical state it would have had if the crash had not happened.

A hard disk spins at 7,200 RPM (8.3ms/rotation). What is the approximate worst-case rotational delay for a single read request (the time to wait for the desired sector to rotate under the head)?

Polling is always worse than using interrupts for waiting on an I/O device to complete.

The time to read a page from an SSD depends strongly on the physical page number and on the location of the previous request.