[linux-lvm] missing device nodes for vg, but vg appears in scan
Geoff Mishkin
gmishkin at acs.bu.edu
Sun Feb 12 05:51:06 UTC 2006
Hi, and thanks for the response. Unfortunately it is still not working. For
completeness, I have:
device-mapper: 4.4.0-ioctl (2005-01-12) (Kernel)
device-mapper 1.01.03 (Userspace)
On Saturday 11 February 2006 09:19 pm, Zac Slade wrote:
> On Sunday 22 January 2006 13:53, Geoff Mishkin wrote:
> > I am using lvm2 2.01.09 on Gentoo Linux. My kernel version is 2.6.15 and
> > I'm using udev 079.
>
> I'm running 2.6.15 also using lvm
>
> > I am trying to set up an LVM2 VG using a loop device. My problem is that
> > device nodes are never created for my VG. However, my VG shows up in
> > scans (vgscan) and displays (vgdisplay). Even vgscan --mknodes does not
> > make the nodes. Because I never get a node for my VG, I can't figure out
> > how I am supposed to create LV's in it.
>
> First make sure that /etc/lvm/lvm.conf is not filtering out /dev/loop/
> devices. Based on the ouput you gave from pvcreate I guess that it isn't,
> but you should double check anyway. Also read through the options in that
> file one might jump out at you.
>
I checked this and it is fine. No other options jump out at me.
> > I have the modules loop, dm-mod and dm-snapshot loaded. Therefore, I do
> > have /dev/mapper/control. Here is what I have done so far:
>
> dm-mod is the only required module. Well and loop for your set up.
>
Okay well I suppose I'd need dm-snapshot eventually if I ever get this part
working. :)
> > It looks like my VG is there, but it still isn't showing up in /dev. I
> > read something about how if udev isn't configured properly, you will get
> > devices like /dev/dm-# where # is 0-9, instead of named VG nodes. I
> > don't have those either, but I follow the instructions which are to
> > uncomment a line in my udev rules:
> >
> > KERNEL=="dm-[0-9]*", PROGRAM="/sbin/devmap_name %M %m", NAME="%k",
> > SYMLINK+="%c"
>
> I looked at my /etc/udev/rules.d/50-udev.rules and I have the following:
> # dm devices (ignore them)
> KERNEL=="dm-[0-9]*", OPTIONS="ignore_device"
> # create a symlink named after the device map name
> # note devmap_name comes with extras/multipath
> #KERNEL=="dm-[0-9]*", PROGRAM="/sbin/devmap_name %M %m", NAME="%k",
> SYMLINK+="%c"
> KERNEL=="device-mapper", NAME="mapper/control"
>
Now this is interesting. So you say that you have working volume groups on
loop devices? But you still have the "ignore_device" option active. So I
went backed and switched this in my udev rules but it still didn't work.
> I'm just not sure why they aren't getting created. Even without vgscan you
> should get the devices if you are using udev. Maybe you upgraded udev and
> zapped the new rules during and etc-update or dispatchconf.
I'm pretty careful about doing etc-update, and that section of my udev rules
now looks the same as yours. Is there anything else specific you think the
problem could be?
--Geoff Mishkin <gmishkin at bu.edu>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/linux-lvm/attachments/20060212/02e62356/attachment.sig>
More information about the linux-lvm
mailing list