1. Find the active node

You run 

mount

and the Server that is active will be shown with /data mounted. This
is done by heartbeat, so we can assume that this one is the active
node. This is just the easiest way, your mileage to find out may vary.

2. Make sure ldap is working

To connect samba needs root as an account in smbpasswd. For this you
run

smbpasswd -a

while you are logged on as root.

3. Adjust settings in /etc/samba/smb.conf

Make sure that you have the ldap settings and passwd settings correct,
Referenc smb.conf.ldap

4. Move the cluster data on the DRBD device

On this node you run

cp /etc/samba/smb.conf /data

rm /etc/samba/smb.conf

ln -s /data/smb.conf /etc/samba/smb.conf

So we copied the smb.conf to the drbd device, removed the original one
and then created a link from the drbd device, which is mirrored to the
other node, to /etc/samba/.

We do the same with two other essential files, first smbpasswd, for
the root account:

cp /etc/samba/smbpasswd /data

rm /etc/samba/smbpasswd

ln -s /data/smbpasswd /etc/samba/smbpasswd

and then

cp /etc/passwd /data

rm /etc/passwd

ln -s /data/passwd /etc/passwd

the reason to do this also with the passwd file is, that the Machine
accounts are stored in the passwd. Unfortunately ldap is a little bit
tricky with machine accounts on SLES9.

Now we move our Users

cp -ax /home /data

rm -rf /home

ln -s /data/home /home

5. Failover

Run

rcsmb stop

As nagios monitors the smb it will start a failover after a few
seconds.

6. Logon to the second node

Now logon to the second node. The smb.conf here is the old local one.
Samba is running, but it does not matter. We issue

rm /etc/samba/smb.conf

ln -s /data/smb.conf /etc/samba/smb.conf

rm /etc/samba/smbpasswd

ln -s /data/smbpasswd /etc/samba/smbpasswd

rm /etc/passwd

ln -s /data/passwd /etc/passwd

rm -rf /home

ln -s /data/home /home

So from the drbd device which is now active on the second node we make
a link to /etc/samba/. By this we made the configuration originating
from node 1 available on node 2.

6. Failover, again

Run

rcsmb stop

As nagios monitors the smb it will start a failover after a few
seconds. Once node 1 is active again, we are using ONE smb.conf on
both servers. Any modification made on this file will be automatically
transferred to the other node in case of a failover.


7. The LDAP data is transferred. So it should work now.