Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 62749 - Doesn't have RAID device nodes required by mdadm
Summary: Doesn't have RAID device nodes required by mdadm
Status: RESOLVED DUPLICATE of bug 80702
Alias: None
Product: Gentoo Release Media
Classification: Unclassified
Component: Everything (show other bugs)
Hardware: All All
: High normal (vote)
Assignee: Gentoo's Team for Core System packages
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2004-09-03 13:30 UTC by Lasse Kärkkäinen
Modified: 2005-07-17 13:06 UTC (History)
1 user (show)

See Also:
Package list:
Runtime testing required: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Lasse Kärkkäinen 2004-09-03 13:30:48 UTC
mdadm --create won't work like it should without proper device nodes (and devfs doesn't automatically create these because there aren't any arrays at that point!).

Workaround: (this should be done in LiveCD startup scripts)
for i in 0 1 2 3 4 5 6 7 8 9 10; do mknod /dev/md$i b 9 $i; done

Real fix: make mdadm create the nodes, make devfs create few spare ones (in addition to active arrays) or some other hack. I don't really see any really good solution for this :/
Comment 1 Caleb Tennis (RETIRED) gentoo-dev 2004-09-03 13:32:52 UTC
/dev/md* gets created automatically whenever you load the raid modules, no?

it does here, anyway.
Comment 2 Lasse Kärkkäinen 2004-09-07 18:01:54 UTC
Devices are only auto-created for RAID arrays that actually exist. When installing the system, there are none. Creating an array with mdadm requires the device node _before_ the array is even created, and thus _before_ it is active and running (which is when devfs would create one).
Comment 3 Chris Gianelloni (RETIRED) gentoo-dev 2004-09-07 19:05:51 UTC
I know a solution to this.... don't use friggin devfs!  *grin*

You could always use udev.  In fact, udev works perfectly in this situation.  Gentoo is heading in the udev direction.  In fact, the next LiveCD will be udev-based entirely.  This does not mean that you won't be able to use devfs on your installed system, just that when you boot the LiveCD, it'll be using udev.

Since we won't be releasing a LiveCD with a fixed devfs (or any hacks to work around the inabilities of devfs), I am going to mark this as RESOLVED-WONTFIX.

If you think this needs to be discussed further, feel free to REOPEN this bug.

Thank you for pointing this out, though.  I have added it to my list of things to check for on the next LiveCD during QA.
Comment 4 Chris Gianelloni (RETIRED) gentoo-dev 2004-10-15 23:29:28 UTC
Alrighty then... I'm not too sure it'll be technically feasable for us to use udev in the time we have left, so I am going to REOPEN this just in case.
Comment 5 Chris Gianelloni (RETIRED) gentoo-dev 2004-10-21 22:00:09 UTC
Alrighty... apparently udev has the same behavior... I can tell you that this won't get fixed due to time for 2004.3, but I am making sure it defintely gets done for 2005.0...
Comment 6 Chris Gianelloni (RETIRED) gentoo-dev 2004-11-15 08:57:28 UTC
We've documented this for 2004.3...

I am going to reassign this to base-system and see what they want to do with it.  Also, I think sys-apps mdadm should probably be moved to sys-fs/mdadm instead.
Comment 7 J. Ryan Earl 2004-12-27 08:31:58 UTC
Yes, this behavior is definitely there for udev.  I freshly imaged a few servers and found I couldn't setup the raid devices because there were no /dev/mdX entries.  Where may I ask was this documented for 2004.3?  I never saw it, I was about to create a bug when I found this.
Comment 8 J. Ryan Earl 2004-12-28 16:37:29 UTC
BTW, the following is correct, the first comment didn't work:

for i in 0 1 2 3 4 5 6 7 8 9 10; do mknod /dev/md${i} b 9 ${i}; done
Comment 9 Mark Schoonover 2005-01-31 12:10:23 UTC
You can create the /dev/md* devices by using mknod.

.mark
Comment 10 Tom Wolfe 2005-02-03 18:52:00 UTC
Would this possibly be affecting me (see http://forums.gentoo.org/viewtopic.php?p=2053048 ) and this will be in 2005.0 still, right (just double checking, i know it was already said) because it does not list the bug as closed and I imagine 2005.0 will be out soon (correct me if I am wrong)?  Nobody on the forums seems to know.  thanks.
Comment 11 SpanKY gentoo-dev 2005-02-04 05:52:30 UTC

*** This bug has been marked as a duplicate of 80702 ***