Re: Mdadm, udev and fakeraid?

[Seblu <seblu@seblu.net>]

On Sat, Apr 23, 2011 at 4:37 AM, Dan Williams <dan.j.williams@gmail.com> wrote:
> On Sun, Apr 17, 2011 at 5:38 PM, NeilBrown <neilb@suse.de> wrote:
>> On Fri, 15 Apr 2011 16:15:50 +0200 Seblu <seblu@seblu.net> wrote:
>>
>>> On Tue, Apr 5, 2011 at 8:20 AM, NeilBrown <neilb@suse.de> wrote:
>>> > On Sun, 3 Apr 2011 18:03:50 +0200 Seblu <seblu@seblu.net> wrote:
>>> >
>>> >> Hello,
>>> >>
>>> >> In the following commit, udev rules load isw_raid (fakeraid). From my
>>> >> test, this doesnt work. I have to call dmraid to have something
>>> >> working.
>>> >> http://neil.brown.name/git?p=mdadm;a=commit;h=475a01b8bce8575dd1b2ab6495e65e854702ac0e
>>> >>
>>> >> isw_raid is only fakeraid devices? mdadm is able to mount fakeraid partition?
>>> >>
>>> >
>>> > I'm sorry but I cannot parse those questions successfully so I'm not sure
>>> > what you are asking.
>>>
>>> Hello Neil,
>>>
>>> in my previous mail, i used word fakeraid about raid created with
>>> dmraid and i used softraid about raid created with mdadm. it was not
>>> clear.
>>>
>>> So my question was about compatibily. Raids created by dmraid can be
>>> assembled with mdadm and vice versa?
>>>
>>> > Both dmraid and mdadm can manage some 'fakeraid' arrays.  dmraid supports a
>>> > wider variety.  mdadm supports raid1 and raid5 more completely than dmraid
>>> > does.
>>> mdadm -> create soft raid for linux  (now there is new format: ddf and imsm) ?
>>> dmraid -> create soft raid from industry raid card format  ?
>>
>> No, it isn't that simple.
>>
>> dmraid uses the 'dm' kernel module.  mdadm uses the 'md' kernel module.
>>
>> As such dmraid doesn't support RAID5 (yet) and doesn't support RAID1 very
>> well.
>> mdadm supports both of these well, but doesn't support the same range of
>> "industry raid card formats".
>>
>> There is a growing amount of overlap.
>>
>>>
>>> > Both should support isw to some degree.
>>> > Intel are currently working with mdadm to make it provide full support for
>>> > "IMSM" (Intel Matrix Storage Manager).  I don't know the exact relationship
>>> > between 'isw' and 'IMSM' - maybe they are different names for the same thing.
>>> ok
>>>
>>> > If mdadm doesn't work for your isw arrays, and you want it to, then I suggest
>>> > you report details about what is, or is not, happening.
>>> My purpose is to improve archlinux startup detection of fakeraids
>>> (mdadm + dmraid).
>>>
>>> With mdadm everything works correctly without call to "mdadm -As"
>>> With dmraid, no raid is created by udev rules, so we need to run
>>> "dmraid -i -ay" at startup.
>>>
>>> To test this kind of raid, i created a dmraid array in a vm. This
>>> created me a /dev/mapper/isw_bfbjdbadhb_testF device.
>>> call blkid on a disk member of this raid tell me this:
>>> /dev/sde: TYPE="isw_raid_member"
>>> and on "mdadm" created raid:
>>> /dev/sdd: UUID="a974b525-993a-1481-f860-6471f3f120e1"
>>> UUID_SUB="eb22aee2-b2ee-e56d-1008-44d52c63564d" LABEL="archipel:0"
>>> TYPE="linux_raid_member"
>>>
>>> This misled me because mdadm udev rules uses the output of blkid to
>>> mount raids which have type "isw_raid_member".
>>> What disturbs me is that mdadm cannot mount raid created by dmraid
>>> with type isw_raid_member.
>>>
>>> About outputs:
>>> mdadm -I --verbose /dev/sde
>>> mdadm: no RAID superblock on /dev/sde.
>
> Seblu can you verify that:
>    export IMSM_NO_PLATFORM=1
>    mdadm -E /dev/sde

archipel ~ 0 # export IMSM_NO_PLATFORM=1
archipel ~ 1 # mdadm -E /dev/sde
/dev/sde:
          Magic : Intel Raid ISM Cfg Sig.
        Version : 1.1.00
    Orig Family : 5a8ed623
         Family : 5a8ed623
     Generation : 00000000
           UUID : ae2e9cd8:7fa43248:47c694a1:24990cbc
       Checksum : c23b6c88 correct
    MPB Sectors : 1
          Disks : 2
   RAID Devices : 1

  Disk00 Serial : 66faec8-9f5b237d
          State : active
             Id : 00040000
    Usable Size : 1019486 (497.88 MiB 521.98 MB)

[testF]:
           UUID : 6640a4cc:5faa1ce3:c1bff2b3:1093ca7d
     RAID Level : 1
        Members : 2
          Slots : [UU]
    Failed disk : none
      This Slot : 0
     Array Size : 1014446 (495.42 MiB 519.40 MB)
   Per Dev Size : 1014792 (495.59 MiB 519.57 MB)
  Sector Offset : 0
    Num Stripes : 3963
     Chunk Size : 64 KiB
       Reserved : 0
  Migrate State : idle
      Map State : normal
    Dirty State : clean

  Disk01 Serial : 0b540c6-4e527908
          State : active
             Id : 00050000
    Usable Size : 1019486 (497.88 MiB 521.98 MB)

>
> finds no superblock? It may be that dmraid has laid down something incompatible.
>
>> As has been mentioned elsewhere, mdadm only recognised IMSM arrays on
>> machines with IMSM hardware.  I'm not entirely happy about this and may well
>> change it.
>
> I have trouble answering the "least surprise" question in this area.
>
> Is it more surprising to go into your BIOS, explicitly turn off raid
> support and still see raid devices showing up?
>
> Or is it more surprising to take a raid array from a raid enabled
> system to raid disabled system and wonder why things won't assemble?
>
> For safety I think it is better if mdadm not perform operations that
> might be incompatible with the platform option-rom.  But if you need
> to recover to a usb attached drive, or some other
> platform-incompatible configuration, you can use the environment
> variable in a pinch.
>
> --
> Dan
>



-- 
Sébastien Luttringer
www.seblu.net
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html