[DRBD-user] Primary diskless

Philipp Marek philipp.marek at linbit.com
Thu Aug 11 08:28:21 CEST 2016

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.

Hi Jason,

> I did a resize of a resource.  I have done many times before.  This is the
> first time on the  DRBD9 9.0.3 release.
> I performed the resize.  Everything came up fine.  Then the primary changed
> to diskless after starting the resource.  The system is still up and
> running. I tried to migrate to secondary but it fails to move with no
> errors.
> So Primary is showing Diskless.  Secondary is showing UpToDate.
> On Primary I get this when trying to adjust:
> drbdadm adjust vm-105-disk-1
> No valid meta data found   Command 'drbdmeta 100 v09
> /dev/drbdpool/vm-105-disk-1_00 internal apply-al' terminated with exit code
> 255
> Any ideas?  Is primary writing to the secondary?  So to recover, what should
> I do?  Force stop the VM by removing the fence and have the primary
> disregard its data for that resource to fix the diskless state?  Does anyone
> know of another method to resolve this?
This sounds as if you did an offline resize, and you didn't have the old 
size stored in "/var/lib/drbd/drbd-minor-100.lkbd", so now the DRBD 
metadata can't be found anymore.

If you can find out the exact old size (LVM backup files, logfiles, 
monitoring, etc.), you can try to create that file with a content like


and then try a

    drbdadm adjust all

Better use online resize next time.

More information about the drbd-user mailing list