software raid + LVM layout/design question

[pblinux2]

From the Software-RAID howto:
http://unthought.net/Software-RAID.HOWT ... TO-11.html

Code: Select all

RAID devices cannot be partitioned, like ordinary disks can. This can be a real annoyance on systems where one wants to run, for example, two disks in a RAID-1, but divide the system onto multiple different filesystems. A horror example could look like:

# df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/md2              3.8G  640M  3.0G  18% /
/dev/md1               97M   11M   81M  12% /boot
/dev/md5              3.8G  1.1G  2.5G  30% /usr
/dev/md6              9.6G  8.5G  722M  93% /var/www
/dev/md7              3.8G  951M  2.7G  26% /var/lib
/dev/md8              3.8G   38M  3.6G   1% /var/spool
/dev/md9              1.9G  231M  1.5G  13% /tmp
/dev/md10             8.7G  329M  7.9G   4% /var/www/html

I'm building a box with a raid1 layout similar to this, but I'm not seeing why this is called a "horror example". Is something going to bite me later on down the road if I forgo using LVM here?

[jtaylor72]

I think they're just pointing out that while this would be a simple task to do in non-raid, it would be a horror to implement in a raid setting. You can't just setup one raid device and then break it into smaller partitions, you have to set your partitions each as a RAID device.

[RaceTM]

Forgive my ignorance but I am pretty sure that a raid array is treated like a regular device. as in, you can fdisk it and subdivide it. I set up a software raid 5 array on my gentoo server a few months ago using 5 250 gig drives. I can fdisk /dev/md/1 and it appears as a 1 tb device. I dont know if you can actually use fdisk to partition it, but mdadm should be able to do it. in which case it would be a matter of creating a new node to represent your new partition, after creating the partition on each physical drive. Hardly anything to keep you from using a "complex" partition scheme.

Hope this helps

[neysx]

From the Software-RAID howto:
http://unthought.net/Software-RAID.HOWT ... TO-11.html
I'm building a box with a raid1 layout similar to this, but I'm not seeing why this is called a "horror example". Is something going to bite me later on down the road if I forgo using LVM here?

Advantages of an LVM-based config are explained in the link you posted.
http://www.gentoo.org/doc/en/gentoo-x86 ... nstall.xml might hep you setting up your system.

Hth

[jtaylor72]

Thanks RaceTM. Looking again at the link he posted, it is from 2004, so this may have been true back then. I was referring to the paragraph right after the one he quoted. I guess we both could have read a little better.

If a RAID device could be partitioned, the administrator could simply have created one single /dev/md0 device device, partitioned it as he usually would, and put the filesystems there. Instead, with today's Software RAID, he must create a RAID-1 device for every single filesystem, even though there are only two disks in the system.

[sobaka]

I am myself an advocate of using LVM2 due to the flexibility of such a solution (in particular if using Xen), but if you don't want to go that way, 2.6-branch kernels can handle partitionable md RAID arrays.

Code: Select all

man mdadm

And search for partitionable.

For the lazy:

Code: Select all

       -a, --auto{=no,yes,md,mdp,part,p}{NN}
              Instruct mdadm to create the device file if needed, possibly
              allocating an unused minor number.  "md" causes a non-partition-
              able array to be used.  "mdp", "part" or "p" causes a partition-
              able array (2.6 and later) to be used.  "yes" requires the named
              md device to have a 'standard' format, and the type and minor
              number will be determined from this.  See DEVICE NAMES below.

              The argument can also come immediately after "-a".  e.g. "-ap".

              If --scan is also given, then any auto= entries in the config
              file will over-ride the --auto instruction given on the command
              line.

              For partitionable arrays, mdadm will create the device file for
              the whole array and for the first 4 partitions.  A different
              number of partitions can be specified at the end of this option
              (e.g.  --auto=p7).  If the device name ends with a digit, the
              partition names add a 'p', and a number, e.g. "/dev/home1p3".
              If there is no trailing digit, then the partition names just
              have a number added, e.g. "/dev/scratch3".

/sobaka