[DRBD-user] Access to the slave node

[Roland Kammerer]

On Thu, Mar 15, 2018 at 10:55:58AM +0000, Ondrej Valousek wrote:
> 
> > That is perfectly expected behavior. Imagine a resync from the primary that, for some time, makes the secondary inconsistent (see what Lars already told you). Would not make sense to mount that
> > one... Error codes are limited, "Wrong medium type" is the one that makes most sense.
> 
> > For DRBD9 it is possible to mount a secondary RO when there is no primary.
> 
> I am using DRBD9 and I am experiencing the same error when I unmount
> the filesystem on the primary node first (i.e. to make sure the
> filesystem is 100% consistent)

If it is not automatically demoted by the unmount (see drbd9
autopromote), it is still DRBD Primary. <- period. In that case it is
expected behavior.

Regards, rck