TODO

   1 2004-june-02
   2   * Don't print 'errors' flag, it is meaningless. DONE
   3   * Handle new superblock format
   4   * create device file on demand, particularly partitionable devices. DONE
   5       BUT figure a way to create the partition devices.
   6               auto=partN
   7   * Use Event: interface to listen for events. DONE, untested
   8   * Make sure mdadm -As can assemble multi-level RAIDs ok.
   9   * --build to build raid1 or multipath arrays
  10        clean or not ???
  11
  12 ----------------------------------------------------------------------------
  13 * mdadm --monitor to monitor failed multipath paths and re-instate them.
  14
  15 * Maybe make "--help" fit in 80x24 and have a --long-help with more info. DONE
  16
  17
  18 * maybe "missing" instead of <bold>missing</> in doco DONE
  19 * possibly wait for resync to start, or even finish while assembling.- NO
  20
  21 * -Db should have a devices= entry if possible. - DONE
  22 * when assembling multipath arrays, ignore any error indicators. - DONE
  23 * rationalise --monitor usage:
  24      mdadm --monitor
  25   doesn't do as expected. DONE
  26
  27 * --assemble could have a --update option. - DONE
  28   following word can be:
  29         sparc2.2
  30         super-minor
  31
  32 * mdadm /dev/md11, where md11 is raid0 can segfault, particularly when looking in the
  33    [UU_UUU] string ... which doesn't exist !
  34 It should be more sensible.  DONE
  35
  36 Example:
  37
  38 from  Raimund Sacherer <raimund.sacherer@ngit.at>
  39
  40 mke2fs -m0 -q /dev/ram1 300
  41 mount -n -t ext2 /dev/ram1 /tmp
  42 echo DEVICE /dev/[sh]* >> /tmp/mdadm.conf
  43 mdadm -Esb /dev/[sh]* 2>/dev/null >> /tmp/mdadm.conf
  44 mdadm -ARsc /tmp/mdadm.conf
  45 umount /tmp
  46
  47
  48 ?? Allow -S /dev/md? - current complains subsequent not a/d/r - DONE
  49
  50 * new "Query" mode to subsume --detail and --examine.
  51    --query or -Q, takes a device and tells if it is an MD device,
  52    and also tells in a raid superblock is found.
  53  DONE
  54
  55 * write mdstat.c to parse /proc/mdstat file
  56    Build list of arrays:  name, rebuild-percent
  57   DONE
  58
  59 * parse /proc/partitions and map major/minor into /dev/* names,
  60   and use that for default DEVICE list ????
  61
  62 * --detail --scan to read /proc/mdstat, and then iterate over these,
  63     but assume --brief.  --verbose can override
  64     check each subdevice to see if it is in conf_get_devs.
  65     Warn if not.
  66   DONE, but don't warn yet...
  67
  68 * Support multipath ... maybe...
  69   maybe DONE
  70
  71 * --follow to syslog
  72
  73 * --follow to move spares around DONE
  74
  75 * --follow to notice other events: DONE
  76      rebuild started
  77      spare activated
  78      spare removed
  79      spare added
  80
  81 ------------------------------------
  82 - --examine --scan scans all drives and build an mdadm.conf file DONE
  83
  84 - check superblock checksum in examine DONE
  85 - report "chunk" or "rounding" depending on raid level DONE
  86 - report "linear" instead of "-1" for raid level DONE
  87 - decode ayout depending on raid level DONE
  88 - --verbose and --force flags. DONE
  89
  90 - set md_minor, *_disks for Create  - DONE
  91 - for create raid5, how to choose between
  92    all working, but not insync
  93    one missing, one spare, insync  DONE (--force)
  94 - and for raid1 - some failed drives...  (missing)
  95
  96 - when RUN_ARRAY, make sure *_disks counts are right
  97
  98 - get --detail to extract extra stuff from superblock,
  99    like uuid  DONE
 100 - --detail --brief to give a config file line DONE
 101 - parse config file. DONE
 102 - test...
 103
 104 - when --assemble --scan, if an underlying device is an md device,
 105   then try to assemble that device first.
 106
 107
 108 - mdadm -S /dev/md0 /dev/md1 gives internal error FIXED
 109
 110 - mdadm --detail --scan print summary of what it can find? DONE
 111
 112
 113 ---------
 114 Assemble doesn't add spares. - DONE
 115 Create to allow "missing" name for devices.
 116 Create to accept "--force" for do exactly what is requested
 117 - get Assemble to upgrade devices if force flag.
 118 ARRAY lines in config file to have super_minor=n
 119 ARRAY lines in config file to have device=pattern, and only accept
 120    those devices
 121    If UUID given, insist on that
 122    If not, but super_minor given, require all found with that minor
 123     to have same uuid
 124    If only device given, all valid supers on those devices must have
 125     same uuid
 126 allow /dev/mdX as first argument before any options
 127 Possible --dry-run option for create and assemble--force
 128
 129 Assemble to check that all devices mentioned in superblock
 130   are present.
 131
 132 New mode: --Monitor (or --Follow)
 133   Periodically check status of all arrays (listed in config file).
 134   Log every event and apparent cause - or differences
 135   Email and alert - or run a program - for important events
 136   Move spares around if necessary.
 137
 138   An Array line can have a spare-group= field that indicates that
 139    the array shares spares with other arrays with the same
 140    spare-group name.
 141    If an array has a failed and no spares, then check all other
 142      arrays in the spare group.  If one has no failures and a spare,
 143      then consider that spare.
 144     Choose the smallest considered spare that is large enough.
 145     If there is one, then hot-remove it from it's home, and
 146     hot-add it to the array in question.
 147
 148   --mail-to address
 149   --alert-handler program
 150
 151   Will also extract information from /proc/mdstat if present,
 152   and consider 20% marks in rebuild as events.
 153
 154   Events are:
 155      drive fails  - causes mail to be sent
 156      rebuild started
 157      spare activated
 158      spare removed
 159      spare added