Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
you need to be subscribed to get your posts through to the list directly. On Thu, Dec 04, 2008 at 02:23:17PM +1100, John Youkhana wrote: > Hi, > > I got a question about drbd. When the drbd connection is interrupted > (from unplugging the lan cable) the nodes go into a standalone status > which is fine, but when i plug the cable back in to resume the > connection the nodes never connect to each other. I read on the drbd > user guide that this connection should be done automatically by drbd. > > Let me talk about the setup so you have some sort of understanding what > i got here. > > I got 2 machines acting as the cluster, the shared resource is a mysql > database and heartbeat managers the cluster. The nodes in the cluster > have 2 network interfaces, one is used to connect to the main network > and other is used to have a direct connection between the nodes. > Everything works how it supposed to, heartbeat takes over when primary > goes down, drbd allows the nodes to connect and synchronize the mysql > database. > > Now when the drbd connection is down (either by unplugging the cable or > shutting down the machine) the nodes go in a standalone status which is > the correct behavior, no it is not. they are supposed to go to WFConnection (wait for connection) the kernel logs should tell you why they go to Standalone instead. maybe you have some automagic deamon that removes the IP address when you unplug the cable? > BUT when the connection is established again (so plugging the cable > back in or turning the machine on) the nodes will never seem to > connect again and the nodes stay in the standalone status. once in standalone, drbd remains there until told to try to reconnect. best way to do so: drbdadm adjust. > It seems > like after the connection is disturbed just once the connection > will never be active again. The only way to have the nodes to connect > again is to do the whole drbd process again, i.e. create the metadata > again, attach the device, connect etc. > > So my question is why does this happen? shouldn't the nodes > automatically connect once there is an active connection between them? normally they do. > How can i fix this? check what is going on, the kernel logs / syslogs should tell you. -- : Lars Ellenberg : LINBIT | Your Way to High Availability : DRBD/HA support and consulting http://www.linbit.com DRBD® and LINBIT® are registered trademarks of LINBIT, Austria. __ please don't Cc me, but send to list -- I'm subscribed