From Linux Raid Wiki
Revision as of 17:47, 1 July 2006 by DavidGreaves (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Historically when the kernel boots it used a mechanism called 'autodetect' to identify partitions (marked as partition type 'fd') which are used in RAID arrays. It then attempts to automatically assemble and start these arrays.

This approach can cause problems in several situations (see various mailing list archives for the debates) and kernel autodetect is deprecated.

The recommended approach now is to use 'initramfs'. This approach allows a great deal of flexibility in preparing the kernel for booting. Essentially the normal kernal boot image is extended to contain an initram filesystem containing a variety of data and scripts. These scripts can do almost anything. A typical usage is to include kernel modules and userspace configuration tools (such as mdadm). This is of interest to us because mdadm could, for example, examine the partitions and only assemble arrays with a defined UUID.

Personal tools