/dev/md/hostname:somename ???

[masuch]

Hi,

I have created software raid 0 arrays.
when I boot up NON-RAID system I have got:

once upon a time it is mapped like
/dev/md/somename
and after further reboot/s it is like:
/dev/md/hostname:somename

what is the rules about it ?

---my /etc/mdadm/mdadm.conf is:
...
#HOMEHOST <system>
...
ARRAY /dev/md/raid0ocz metadata=1.2 UUID=152fb8b4:95265477:57e0dde8:91307b7b name=raid0ocz
ARRAY /dev/md/raid0ext41 metadata=1.2 UUID=8248abdc:c6cb1226:1f60ded4:3b5dbffa name=raid0ext41
...

--- my /etc/fstab:
...
# --- RAID 0 first try:
/dev/md/raid0ocz /media/raid0ocz ext4 rw,suid,dev,exec,auto,nouser,async,acl,user_xattr, nobootwait,errors=remount-ro,noatime,discard 0 0
/dev/md/raid0ext41 /media/raid0ext41 ext4 rw,suid,dev,exec,auto,nouser,async,acl,user_xattr, nobootwait,errors=remount-ro 0 0

# --- RAID 0 if previous failed try following:
/dev/md/PCEUBU1:raid0ocz /media/raid0ocz ext4 rw,suid,dev,exec,auto,nouser,async,acl,user_xattr, nobootwait,errors=remount-ro,noatime,discard 0 0
/dev/md/PCEUBU1:raid0ext41 /media/raid0ext41 ext4 rw,suid,dev,exec,auto,nouser,async,acl,user_xattr, nobootwait,errors=remount-ro 0 0
...

--- it is happening as well when I am in busybox and/or grub prompt.

Does have any influence on creating /dev/md/somename LABEL/s of the raid partitions (to make name like /dev/md/partition-label) ?
I am asking about this because I do not know where else to look how OS figured out the mapping like /dev/md/hostname:somename(partition label) ? I have found that only partitions LABEL contain hostname string.


thank you for explanation or hint to better understand.
kind regards,
Martin

[mattheworiordan]

I have this exact same problem, on a server on EC2.

Did you ever find a solution to this?

[mattheworiordan]

BTW. I've just found a solution I don't undersand, but it works.

After updating your mdadm file, possibly with something like:
mdadm --detail --scan >> /etc/mdadm/mdadm.conf

Simply run the following command
update-initramfs -u

Then reboot, and everything will work as expected.

[masuch]

BTW. I've just found a solution I don't undersand, but it works.

After updating your mdadm file, possibly with something like:
mdadm --detail --scan >> /etc/mdadm/mdadm.conf

Simply run the following command
update-initramfs -u

Then reboot, and everything will work as expected.

Hi,
I did not find a solution but I am using the same workaround as you mentioned. I run from crontab 4 times daily update-initramfs -u -k all
But I do not understand it anyway and did not figured out why.
In /etc/fstab I am using label name with all my raid0 arrays instead of /dev/something.