[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Please explain status/background of /dev/mdX versus /dev/md/X



also sprach Frans Pop <elendil@planet.nl> [2008.04.22.1820 +0200]:
> What's responsible for creating them: udev or mdadm itself?

mdadm.

> We do. The superblock is version-0, but mdcfg does use /dev/md/X (and has 
> done since its original upload in 2004).
> 
> What we do for new RAID devices is:
> mdadm --create /dev/md/$MD_NUM --auto=yes --force -R -l <type> <dev spec>

That's the problem. You should be able to change that to md/$MD_NUM
and then just use mdX everywhere without much trouble.

You can control the creation mode with mdadm.conf like this:

  CREATE owner=root group=disk mode=0660

> > But in any case, mdadm should not create /dev/md/X files if you tell
> > it to create /dev/mdX instead (passing --auto=yes).
> 
> Sure, but it does when the above call is used, or when you have
>    ARRAY /dev/md/X ...
> in the config file.

Yes. That's expected, although probably not desirable or sensible or
whatever...

> It seems that maybe we should currently not be using /dev/md/X at all in the 
> installer. Especially since the installed system will probably also _not_ 
> have them since it creates a new config file based on output of
> 'mdadm --detail --scan'...
> 
> What do you think?

Yes, I think /dev/mdX for now should do. In the long-run, /dev/md/*
might be cleaner. However, to get there, it has to be default, and
when it is, I have to provide a migration path. Yay!

-- 
 .''`.   martin f. krafft <madduck@debian.org>
: :'  :  proud Debian developer, author, administrator, and user
`. `'`   http://people.debian.org/~madduck - http://debiansystem.info
  `-  Debian - when you have better things to do than fixing systems
 
"the scientific paper in its orthodox form does embody a totally
 mistaken conception, even a travesty, of the nature of scientific
 thought."
                                                -- sir peter medawar

Attachment: digital_signature_gpg.asc
Description: Digital signature (see http://martin-krafft.net/gpg/)


Reply to: