F10 install - RAID - nightmare (Solved)

[Robin Laing]

Robin Laing wrote:
> Hello,
> 
> The system is at home and so are all my notes.
> 
> Since I first started using RAID arrays, this is the first time I have 
> had problems with an install.  I have been fighting this for over a 
> week.  The machine was running F7 with RAID arrays.
> 
> I first tried to install F10 using a DVD that was checked by both 
> sha1sum and disk check on install including the RAID array.
> 
> The install is working without the RAID array.
> 
> After installing on the non-RAID drive, I started going through the 
> install to get the RAID working.
> 
> After much reading I found out that due to the problem install, I had to 
> zero the Superblocks.  I did this and ensured that there was no 
> superblock data with mdadm --examine {partitions}.
> 
> Recreated the multiple RAID partitions.
> 
> I am using a 1.5T drive partitions into 8 usable partitions.
> 
> I created the 8 partitions using mdadm.
> 
> I created /etc/mdadm/mdadm.conf with mdadm --examine --scan as per the 
> man page.
> 
I am providing this in a hope that it will help someone either today or 
in the future.  Someone else success helped me.  After 10 days I can say 
I have a working F10 installation.  Hey, 10 for 10.  :)

To solve the issue I did a full re-install without the RAID array.  I 
have read reports about anaconda having issues with RAID arrays.  After 
making sure that the install was working well I started playing with the 
RAID.

With no /etc/mdadm.conf, the system scanned and created inactive arrays.

md_d9 : inactive sdc9[0](S)
       615723136 blocks

md_d8 : inactive sdc8[0](S)
       104864192 blocks

md_d7 : inactive sdc7[0](S)
       73408896 blocks

md_d6 : inactive sdc6[0](S)
       73408896 blocks

md_d5 : inactive sdc5[0](S)
       73408896 blocks

md_d3 : inactive sdc3[0](S)
       209728448 blocks

md_d2 : inactive sdc2[0](S)
       209728448 blocks

md_d1 : inactive sdc1[0](S)
       104864192 blocks

I created a new /etc/mdadm.conf file with the two drives in it like this.
   DEVICE /dev/sdb* /dev/sdc*

I then scanned the drives by using
   mdadm --examine --scan
ARRAY /dev/md1 level=raid1 num-devices=2 \ 
UUID=512ebb9b:05c4c817:22ba247c:074b5b12
ARRAY /dev/md2 level=raid1 num-devices=2 \ 
UUID=bdd5f629:8788d740:b569c872:71bb0d9f
ARRAY /dev/md3 level=raid1 num-devices=2 \ 
UUID=649f208e:07a19b6b:119481b7:34c39216
ARRAY /dev/md5 level=raid1 num-devices=2 \ 
UUID=1a428b1f:5b8a7214:e195441f:012ae200
ARRAY /dev/md6 level=raid1 num-devices=2 \ 
UUID=f222563b:a73aba50:e34cb61b:312f8680
ARRAY /dev/md7 level=raid1 num-devices=2 \ 
UUID=dc04f2ee:11b76d67:77b1b096:0fea140a
ARRAY /dev/md8 level=raid1 num-devices=2 \ 
UUID=82bbc5d9:f612fb5b:15177e5c:b51a48df
ARRAY /dev/md9 level=raid1 num-devices=2 \ 
UUID=62c32558:310c027c:fdacac45:9b3ade78

I then ran
   mdadm --examine --scan >> /etc/mdadm.conf
as suggested in the mdadm man page.  This added the drives to mdadm.conf

I then ran
   mdadm -As
which found and activated one of the two drives as shown with
   cat /proc/mdstat

[root at eagle2 etc]# cat /proc/mdstat
Personalities : [raid1]
md9 : active raid1 sdb9[1]
       615723136 blocks [2/1] [_U]

md8 : active raid1 sdb8[1]
       104864192 blocks [2/1] [_U]

md7 : active raid1 sdb7[1]
       73408896 blocks [2/1] [_U]

md6 : active raid1 sdb6[1]
       73408896 blocks [2/1] [_U]

md5 : active raid1 sdb5[1]
       73408896 blocks [2/1] [_U]

md3 : active raid1 sdb3[1]
       209728448 blocks [2/1] [_U]

md2 : active raid1 sdb2[1]
       209728448 blocks [2/1] [_U]

md1 : active raid1 sdb1[1]
       104864192 blocks [2/1] [_U]

md_d9 : inactive sdc9[0](S)
       615723136 blocks

md_d8 : inactive sdc8[0](S)
       104864192 blocks

md_d7 : inactive sdc7[0](S)
       73408896 blocks

md_d6 : inactive sdc6[0](S)
       73408896 blocks

md_d5 : inactive sdc5[0](S)
       73408896 blocks

md_d3 : inactive sdc3[0](S)
       209728448 blocks

md_d2 : inactive sdc2[0](S)
       209728448 blocks

unused devices: <none>

I then ran
   mdadm --stop /dev/md_d{x}

to stop all the inactive RAID devices as shown in the /proc/mdstat file.

I tried a reboot and only one of the two drives were starting.  More 
reading of bug reports and came across a discussion on adding
   auto=md
to each line of the mdadm.conf file for each raid array.

Old

ARRAY /dev/md1 level=raid1 num-devices=2 \ 
UUID=512ebb9b:05c4c817:22ba247c:074b5b12

New

ARRAY /dev/md1 level=raid1 auto=md num-devices=2 \ 
UUID=512ebb9b:05c4c817:22ba247c:074b5b12

Now running
   mdadm -As
gives this nice message.

mdadm: /dev/md1 has been started with 2 drives.
mdadm: /dev/md2 has been started with 2 drives.
mdadm: /dev/md3 has been started with 2 drives.
mdadm: /dev/md5 has been started with 2 drives.
mdadm: /dev/md6 has been started with 2 drives.
mdadm: /dev/md7 has been started with 2 drives.
mdadm: /dev/md8 has been started with 2 drives.
mdadm: /dev/md9 has been started with 2 drives.

Confirmed by

[root at eagle2 etc]# cat /proc/mdstat
Personalities : [raid1]
md9 : active raid1 sdc9[0] sdb9[1]
       615723136 blocks [2/2] [UU]

md8 : active raid1 sdc8[0] sdb8[1]
       104864192 blocks [2/2] [UU]

md7 : active raid1 sdc7[0] sdb7[1]
       73408896 blocks [2/2] [UU]

md6 : active raid1 sdc6[0] sdb6[1]
       73408896 blocks [2/2] [UU]

md5 : active raid1 sdc5[0] sdb5[1]
       73408896 blocks [2/2] [UU]

md3 : active raid1 sdc3[0] sdb3[1]
       209728448 blocks [2/2] [UU]

md2 : active raid1 sdc2[0] sdb2[1]
       209728448 blocks [2/2] [UU]

md1 : active raid1 sdc1[0] sdb1[1]
       104864192 blocks [2/2] [UU]

unused devices: <none>

And it works with a reboot.

-- 
Robin Laing