ANNOUNCE-3.0-devel2

   1 Subject:  ANNOUNCE: mdadm 3.0-devel2 - A tool for managing Soft RAID under Linux
   2
   3 I am pleased to announce the availability of
   4    mdadm version 3.0-devel2
   5
   6 It is available at the usual places:
   7    countrycode=xx.
   8    http://www.${countrycode}kernel.org/pub/linux/utils/raid/mdadm/
   9 and via git at
  10    git://neil.brown.name/mdadm
  11    http://neil.brown.name/git?p=mdadm
  12
  13 Note that this is a "devel" release.  It should be used with
  14 caution, though it is believed to be close to release-candidate stage.
  15
  16 The significant change which justifies the new major version number is
  17 that mdadm can now handle metadata updates entirely in userspace.
  18 This allows mdadm to support metadata formats that the kernel knows
  19 nothing about.
  20
  21 Currently two such metadata formats are supported:
  22   - DDF  - The SNIA standard format
  23   - Intel Matrix - The metadata used by recent Intel ICH controlers.
  24
  25 Also the approach to device names has changed significantly.
  26
  27 If udev is installed on the system, mdadm will not create any devices
  28 in /dev.  Rather it allows udev to manage those devices.  For this to work
  29 as expected, the included udev rules file should be installed.
  30
  31 If udev is not install, mdadm will still create devices and symlinks
  32 as required, and will also remove them when the array is stopped.
  33
  34 mdadm now requires all devices which do not have a standard name (mdX
  35 or md_dX) to live in the directory /dev/md/.  Names in this directory
  36 will always be created as symlinks back to the standard name in /dev.
  37
  38 The man pages contain some information about the new externally managed
  39 metadata.  However see below for a more condensed overview.
  40
  41 Externally managed metadata introduces the concept of a 'container'.
  42 A container is a collection of (normally) physical devices which have
  43 a common set of metadata.  A container is assembled as an md array, but
  44 is left 'inactive'.
  45
  46 A container can contain one or more data arrays.  These are composed from
  47 slices (partitions?) of various devices in the container.
  48
  49 For example, a 5 devices DDF set can container a RAID1 using the first
  50 half of two devices, a RAID0 using the first half of the remain 3 devices,
  51 and a RAID5 over thte second half of all 5 devices.
  52
  53 A container can be created with
  54
  55    mdadm --create /dev/md0 -e ddf -n5 /dev/sd[abcde]
  56
  57 or "-e imsm" to use the Intel Matrix Storage Manager.
  58
  59 An array can be created within a container either by giving the
  60 container name and the only member:
  61
  62    mdadm -C /dev/md1 --level raid1 -n 2 /dev/md0
  63
  64 or by listing the component devices
  65
  66    mdadm -C /dev/md2 --level raid0 -n 3 /dev/sd[cde]
  67
  68 To assemble a container, it is easiest just to pass each device in turn to
  69 mdadm -I
  70
  71   for i in /dev/sd[abcde]
  72   do mdadm -I $i
  73   done
  74
  75 This will assemble the container and the components.
  76
  77 Alternately the container can be assembled explicitly
  78
  79    mdadm -A /dev/md0 /dev/sd[abcde]
  80
  81 Then the components can all be assembled with
  82
  83    mdadm -I /dev/md0
  84
  85 For each container, mdadm will start a program called "mdmon" which will
  86 monitor the array and effect any metadata updates needed.  The array is
  87 initially assembled readonly. It is up to "mdmon" to mark the metadata
  88 as 'dirty' and which the array to 'read-write'.
  89
  90 The version 0.90 and 1.x metadata formats supported by previous
  91 versions for mdadm are still supported and the kernel still performs
  92 the same updates it use to.  The new 'mdmon' approach is only used for
  93 newly introduced metadata types.
  94
  95 Any testing and feedback will be greatly appreciated.
  96
  97 NeilBrown  5th November 2008
  98