[DRBD-user] Problem while do drbdmanage node-add

Robert Altnoeder robert.altnoeder at linbit.com
Tue Oct 3 14:42:55 CEST 2017

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


This seems to be the same bug that I hit on one of my test clusters.
It is a race condition that we are currently investigating.

If a somehow outdated resource is started, udev causes a read-only open
attempt on the resource, which will - at least in some cases - attempt
to wait for an established replication connection to attain access to up
to date data, but when the other system is in the Primary role, the
connection fails, because if there is a Primary then there must not be
any Readers on Secondaries, and therefore a state change is aborted.
The typical symptom that you will see is a flapping connection state on
the resource, Connecting -> NetworkFailure -> Unconnected -> Connecting,
or Connecting -> BrokenPipe -> ... if you have quick enough live
monitoring (if your drbd-utils are up to date, you can probably see it
using drbdmon).

We are currently trying to fix or at least work around this problem.

The problem has probably always been there, but was hit rarely due to a
very short time window for it to occur. This time window has grown due
to additional locking that was added recently to prevent another
unrelated race condition.

On 10/02/2017 05:18 PM, Tsirkas Georgios wrote:
> Hello,
> When i do drbdmanage node-add  i get errors. See the attached
> screenshots.
> Can someone help me?
>

br,
-- 
Robert Altnoeder
+43 1 817 82 92 0
robert.altnoeder at linbit.com

LINBIT | Keeping The Digital World Running
DRBD - Corosync - Pacemaker
f /  t /  in /  g+

DRBD® and LINBIT® are registered trademarks of LINBIT, Austria.



More information about the drbd-user mailing list