In some cases, a cluster does not compress well. For instance, a cluster with only one block of data will not compress into less than one block. (At the time of writing, `fragments' are not supported in the ext2 filesystem, so the minimum allocation grain size is a block.) Sometimes compressed data occupies more bytes than if uncompressed. In such cases we just leave the cluster as it was. For instance if cluster #1 does not compress well, the compressed file would look like this:
block# bytes ------ -----  compressed data <- cluster #0  missing block  missing block  missing block  data <- cluster #1  data (does not compress well)  data  data  compressed data <- cluster #2  compressed data  missing block
Go to the first, previous, next, last section, table of contents.