TODO

   1 2005-dec-20
   2   Want an incremental assembly mode to work nicely with udev.
   3   Core usage would be something like
   4        mdadm --incr-assemble /dev/newdevice
   5   This would
   6      - examine the device to determine  uuid etc.
   7      - look for a match in /etc/mdadm.conf, abort if not found
   8      - find that device and collect current contents
   9      - perform an 'assemble' analysis to make sure we have the best set of devices.
  10      - remove or add devices as appropriate
  11      - possibly start the array if it was complete
  12
  13    Other usages could involve
  14      - specify which array to auto-add to.
  15        This requires an existing array for uuid matching... is there any point?
  16
  17      -
  18
  19
  20 2004-june-02
  21   * Don't print 'errors' flag, it is meaningless. DONE
  22   * Handle new superblock format
  23   * create device file on demand, particularly partitionable devices. DONE
  24       BUT figure a way to create the partition devices.
  25               auto=partN
  26   * Use Event: interface to listen for events. DONE, untested
  27   * Make sure mdadm -As can assemble multi-level RAIDs ok.
  28   * --build to build raid1 or multipath arrays
  29        clean or not ???
  30
  31 ----------------------------------------------------------------------------
  32 * mdadm --monitor to monitor failed multipath paths and re-instate them.
  33
  34 * Maybe make "--help" fit in 80x24 and have a --long-help with more info. DONE
  35
  36
  37 * maybe "missing" instead of <bold>missing</> in doco DONE
  38 * possibly wait for resync to start, or even finish while assembling.- NO
  39
  40 * -Db should have a devices= entry if possible. - DONE
  41 * when assembling multipath arrays, ignore any error indicators. - DONE
  42 * rationalise --monitor usage:
  43      mdadm --monitor
  44   doesn't do as expected. DONE
  45
  46 * --assemble could have a --update option. - DONE
  47   following word can be:
  48         sparc2.2
  49         super-minor
  50
  51 * mdadm /dev/md11, where md11 is raid0 can segfault, particularly when looking in the
  52    [UU_UUU] string ... which doesn't exist !
  53 It should be more sensible.  DONE
  54
  55 Example:
  56
  57 from  Raimund Sacherer <raimund.sacherer@ngit.at>
  58
  59 mke2fs -m0 -q /dev/ram1 300
  60 mount -n -t ext2 /dev/ram1 /tmp
  61 echo DEVICE /dev/[sh]* >> /tmp/mdadm.conf
  62 mdadm -Esb /dev/[sh]* 2>/dev/null >> /tmp/mdadm.conf
  63 mdadm -ARsc /tmp/mdadm.conf
  64 umount /tmp
  65
  66
  67 ?? Allow -S /dev/md? - current complains subsequent not a/d/r - DONE
  68
  69 * new "Query" mode to subsume --detail and --examine.
  70    --query or -Q, takes a device and tells if it is an MD device,
  71    and also tells in a raid superblock is found.
  72  DONE
  73
  74 * write mdstat.c to parse /proc/mdstat file
  75    Build list of arrays:  name, rebuild-percent
  76   DONE
  77
  78 * parse /proc/partitions and map major/minor into /dev/* names,
  79   and use that for default DEVICE list ????
  80
  81 * --detail --scan to read /proc/mdstat, and then iterate over these,
  82     but assume --brief.  --verbose can override
  83     check each subdevice to see if it is in conf_get_devs.
  84     Warn if not.
  85   DONE, but don't warn yet...
  86
  87 * Support multipath ... maybe...
  88   maybe DONE
  89
  90 * --follow to syslog
  91
  92 * --follow to move spares around DONE
  93
  94 * --follow to notice other events: DONE
  95      rebuild started
  96      spare activated
  97      spare removed
  98      spare added
  99
 100 ------------------------------------
 101 - --examine --scan scans all drives and build an mdadm.conf file DONE
 102
 103 - check superblock checksum in examine DONE
 104 - report "chunk" or "rounding" depending on raid level DONE
 105 - report "linear" instead of "-1" for raid level DONE
 106 - decode ayout depending on raid level DONE
 107 - --verbose and --force flags. DONE
 108
 109 - set md_minor, *_disks for Create  - DONE
 110 - for create raid5, how to choose between
 111    all working, but not insync
 112    one missing, one spare, insync  DONE (--force)
 113 - and for raid1 - some failed drives...  (missing)
 114
 115 - when RUN_ARRAY, make sure *_disks counts are right
 116
 117 - get --detail to extract extra stuff from superblock,
 118    like uuid  DONE
 119 - --detail --brief to give a config file line DONE
 120 - parse config file. DONE
 121 - test...
 122
 123 - when --assemble --scan, if an underlying device is an md device,
 124   then try to assemble that device first.
 125
 126
 127 - mdadm -S /dev/md0 /dev/md1 gives internal error FIXED
 128
 129 - mdadm --detail --scan print summary of what it can find? DONE
 130
 131
 132 ---------
 133 Assemble doesn't add spares. - DONE
 134 Create to allow "missing" name for devices.
 135 Create to accept "--force" for do exactly what is requested
 136 - get Assemble to upgrade devices if force flag.
 137 ARRAY lines in config file to have super_minor=n
 138 ARRAY lines in config file to have device=pattern, and only accept
 139    those devices
 140    If UUID given, insist on that
 141    If not, but super_minor given, require all found with that minor
 142     to have same uuid
 143    If only device given, all valid supers on those devices must have
 144     same uuid
 145 allow /dev/mdX as first argument before any options
 146 Possible --dry-run option for create and assemble--force
 147
 148 Assemble to check that all devices mentioned in superblock
 149   are present.
 150
 151 New mode: --Monitor (or --Follow)
 152   Periodically check status of all arrays (listed in config file).
 153   Log every event and apparent cause - or differences
 154   Email and alert - or run a program - for important events
 155   Move spares around if necessary.
 156
 157   An Array line can have a spare-group= field that indicates that
 158    the array shares spares with other arrays with the same
 159    spare-group name.
 160    If an array has a failed and no spares, then check all other
 161      arrays in the spare group.  If one has no failures and a spare,
 162      then consider that spare.
 163     Choose the smallest considered spare that is large enough.
 164     If there is one, then hot-remove it from it's home, and
 165     hot-add it to the array in question.
 166
 167   --mail-to address
 168   --alert-handler program
 169
 170   Will also extract information from /proc/mdstat if present,
 171   and consider 20% marks in rebuild as events.
 172
 173   Events are:
 174      drive fails  - causes mail to be sent
 175      rebuild started
 176      spare activated
 177      spare removed
 178      spare added