[CentOS] DRBD Problem

Sat Jun 24 13:00:53 UTC 2006
Johnny Hughes <mailing-lists at hughesjr.com>

On Sat, 2006-06-24 at 13:50 +0100, centos at bathnetworks.com wrote:
> > On Sat, 2006-06-24 at 11:34 +0100, centos at bathnetworks.com wrote:
> >> > On Sat, 2006-06-24 at 07:02 +0100, centos at bathnetworks.com wrote:
> >> >> Hi all,
> >> >>
> >> >> I've been wrestling with a problem with drdb and centos. I have
> >> >> successfully created one drbd resource, but when I try the create a
> >> 2nd,
> >> >> I
> >> >> get an error on one of the nodes:
> >> >>
> >> >> Lower device is already mounted.
> >> >> Command 'drbdsetup /dev/drbd1 disk /dev/hdd1 internal -1' terminated
> >> >> with
> >> >> exit code 20
> >> >>
> >> >> The partition is not mounted from fstab etc and is newly created with
> >> >> parted after wiping this disk with dd.
> >> >>
> >> >> On the node with the problem I see:
> >> >>
> >> >> cat /proc/partitions
> >> >> major minor  #blocks  name
> >> >>
> >> >>    3     0   60034968 hda
> >> >>    3     1     514048 hda1
> >> >>    3     2    1534207 hda2
> >> >>    3     3   57986617 hda3
> >> >>   22     0  120060864 hdc
> >> >>   22     1  114688003 hdc1
> >> >>   22    64  117220824 hdd
> >> >>   22    65  114688003 hdd1
> >> >>  253     0  117219800 dm-0
> >> >>  253     1  114688003 dm-1
> >> >>  147     0  114556928 drbd0
> >> >>
> >> >> On the other node the dm-0 and dm-1 do not occur. Being new to Linux,
> >> I
> >> >> am
> >> >> not sure where these come from. I have tried google, but nothing
> >> makes
> >> >> sense.  This maybe a red herring, but dm-1 has the same number of
> >> blocks
> >> >> as hdd1 which I am trying to mount so I'm guessing that this is the
> >> >> cause
> >> >> of the problem.
> >> >>
> >> >> As I have tried changing every thing I can except the mb/processor
> >> and
> >> >> reinstalling, I'm really stuck.
> >> >>
> >> >> HELP Please.
> >> >>
> >> >> Rob
> >> >>
> >> >>
> >> >>
> >> >
> >> > I am not an expert on filesystems ... but dm-0 and dm-1 look like raid
> >> 0
> >> > or LVM partitions as they have 253 as their Major device number.
> >> >
> >> > What does fdisk -l say for that drive (/dev/hdd).
> >> >
> >>
> >> Thanks for the reply.
> >>
> >> As far as I am aware, there is no raid or lvm setup on the machine.
> >>
> >> disk -l for hdd gives:
> >>
> >> Disk /dev/hdd: 120.0 GB, 120034123776 bytes
> >> 255 heads, 63 sectors/track, 14593 cylinders
> >> Units = cylinders of 16065 * 512 = 8225280 bytes
> >>
> >>    Device Boot      Start         End      Blocks   Id  System
> >> /dev/hdd1               1       14278   114688003+  83  Linux
> >>
> >> Rob
> >>
> >
> > If you are not using raid on purpose, try removing the package dmraid
> > like this:
> >
> > rpm -e dmraid
> >
> > then rebooting
> 
> Johnny,
> 
> Once again you are proving to be the man.
> 
> That fixed it. Both resources are up and the 2nd is syncing.
> 
> One question, I guess I don't need dmraid as I don't have a raid array,
> but why would it 'grab' the disk as it had never been set in the 1st
> place?
> 

No idea what caused it to do that ... it is obviously some kind of bug
if you didn't configure it that way.

I have had issues with the dmraid module in the past, and since I saw
that it was loading those modules, I figured removing it might help.

Some things to check are ...

1.  If the machine has RAID settings in the BIOS (and you are not trying
to do RAID) ... turn them off.

2.  Be on the look out for dmraid to be reinstalled ... it is in the
base/core package group in comps.xml (and therefore part of the minimal
install) ... so it is liable to make it's way back in.
--------------------------------------------------------
The purpose of dmraid is to allow the SW RAID controllers that M$
Winders has tricked people into thinking are real hardware RAIDs to be
configured as RAID on Linux too ... see this readme:

http://people.redhat.com/heinzm/sw/dmraid/readme

So ... if you have one of those controllers w/RAID set to initialize, it
will be found and mapped automagically.

Not a good thing ... IMHO ... but I am not as smart as the upstream
provider :)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://lists.centos.org/pipermail/centos/attachments/20060624/4c073adb/attachment-0004.sig>