RAID starting at md127 instead of md0

[annagel]

I have recently created a RAID 5 array under Ubuntu 11.04 and am having some trouble getting it to assemble correctly on system boot. So far after every boot the array assembles correctly but in read-only mode at /dev/md127. Stopping this array and running:

mdadm --assemble --scan

will correctly mount the array at /dev/md0. My mdadm.conf file:

DEVICE partitions
ARRAY /dev/md/0 metadata=1.2 name=nagelnas:0 UUID=e4665ceb:15f8e4b6:b186d497:7d365254

detail of the array:

sudo mdadm --detail /dev/md0

/dev/md0:
        Version : 1.2
  Creation Time : Sun May 15 22:39:38 2011
     Raid Level : raid5
     Array Size : 5860544512 (5589.05 GiB 6001.20 GB)
  Used Dev Size : 1465136128 (1397.26 GiB 1500.30 GB)
   Raid Devices : 5
  Total Devices : 5
    Persistence : Superblock is persistent

    Update Time : Sun May 22 09:52:09 2011
          State : clean
 Active Devices : 5
Working Devices : 5
 Failed Devices : 0
  Spare Devices : 0

         Layout : left-symmetric
     Chunk Size : 512K

           Name : nagelnas:0  (local to host nagelnas)
           UUID : e4665ceb:15f8e4b6:b186d497:7d365254
         Events : 18

    Number   Major   Minor   RaidDevice State
       0       8        0        0      active sync   /dev/sda
       1       8       16        1      active sync   /dev/sdb
       2       8       32        2      active sync   /dev/sdc
       3       8       48        3      active sync   /dev/sdd
       5       8       64        4      active sync   /dev/sde


cat /proc/mdstat

Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10] 
md0 : active raid5 sda[0] sde[5] sdd[3] sdc[2] sdb[1]
      5860544512 blocks super 1.2 level 5, 512k chunk, algorithm 2 [5/5] [UUUUU]
      
unused devices: <none>

examine of one of the disks in the array:

sudo mdadm --examine /dev/sda

/dev/sda:
          Magic : a92b4efc
        Version : 1.2
    Feature Map : 0x0
     Array UUID : e4665ceb:15f8e4b6:b186d497:7d365254
           Name : nagelnas:0  (local to host nagelnas)
  Creation Time : Sun May 15 22:39:38 2011
     Raid Level : raid5
   Raid Devices : 5

 Avail Dev Size : 2930273007 (1397.26 GiB 1500.30 GB)
     Array Size : 11721089024 (5589.05 GiB 6001.20 GB)
  Used Dev Size : 2930272256 (1397.26 GiB 1500.30 GB)
    Data Offset : 2048 sectors
   Super Offset : 8 sectors
          State : clean
    Device UUID : 922f3fc5:b9078b42:86230bc3:b8abfc7e

    Update Time : Sun May 22 09:51:29 2011
       Checksum : 5274f4ec - correct
         Events : 18

         Layout : left-symmetric
     Chunk Size : 512K

   Device Role : Active device 0
   Array State : AAAAA ('A' == active, '.' == missing)

Any idea what I am doing wrong here? Not the biggest deal in the world to assemble the array by hand but still something I want to have happen automatically.

Thanks,
Andrew

[annagel]

Some additional information, after boot the status of the array is:

cat /proc/mdstat 

Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10] 
md127 : active (auto-read-only) raid5 sdc[2] sdd[3] sde[5] sdb[1] sda[0]
      5860544512 blocks super 1.2 level 5, 512k chunk, algorithm 2 [5/5] [UUUUU]
      
unused devices: <none>

[gregmcc]

I'm having exactly the same problems as you. I was running 10.10 with a RAID5 set as md0. Been working fine for ages. I upgraded to 11.04 and my raid is now appearing at md127.

I have to stop the raid, run assemble which puts it back to md0 and I can then mount it.

[technopagan]

Thought I would post here to say that I had this same problem with ubuntu 11.04 -generic

With the -server kernel installed however, (11.04) works fine, for anyone who finds this thread.

[justifiedandancient]

I have exactly the same problem on Fedora 13:
uname -a
*** 2.6.34.8-68.fc13.x86_64 #1 SMP ***
Curiously, I tried changing /etc/fstab to mount /dev/md127 rather than what I originally had /dev/md0 and I still got the boot message:
Mounting local filesystems: mount: special device /dev/md0 does not exist

[YesWeCan]

Hi all.
This seems to be a "feature" of the mdadm in the newest kernel. There was another thread with similar symptoms recently.
For reasons I do not yet understand, mdadm seems to invent an array device name out of thin air when there is a problem. This confuses everybody. md127 seems to be a common name it now chooses.

Two things seem to be important:

1) The ARRAY statement in mdadm.conf needs to be correct. I am not convinced the direct output of mdadm -Es is reliable anymore. Especially the name directive. Mine broke because I had a space in the name!
So I would try minimizing the specifiers in your ARRAY statements: all you normally need is device name and UUID:

ARRAY /dev/md0 UUID=e4665ceb:15f8e4b6:b186d497:7d365254

And there seems to be some change to do with /dev/md/* and /dev/md* which I don't yet understand.


2) You need to update initramfs so it contains your mdadm.conf settings during boot.

sudo update-initramfs -u

Try those and report back.

[Steve Evans]

Thank you! This was exactly the problem I was having and it's now all working wonderfully!

Steve

[nasha]

Yep, that's correct.

The new mdadm/kernel setup creates the default mdadm.conf file with --name=NAS:0 or equivalent.

When a --name parameter is set, a random (seems to always be 127) md device is created that actually symlinks to /dev/md/NAS:0 or equivalent.

Removing the name= from the mdadm.conf sets it back to normal

[annagel]

Thanks YesWeCan that worked for me as well...good thing I came back to this thread to check it out for some reason the auto-notify feature had stopped notifying me.

[jonathanwalter]

Yup. Worked for me too. Thanks!