[DRBD-user] DBRD over sdp not working

[rf at q-leap.de]

Hi,

I'm trying to use DRBD over sdp:

DRBD 8.3.8.1
Kernel 2.6.32.23
OFED 1.5.2 (1.4.x doesn't work on 2.6.32 anymore).
Debian Lenny

When trying to connect, I get the famous:

block drbd0: connect failed, err = -22

kernel error message.

Relevant excerpt from my drbd.conf:

--------------------
resource r0 {
  protocol C;

  startup { wfc-timeout 0; degr-wfc-timeout 120; }
  disk    { on-io-error detach; }
  net     { timeout 60; connect-int 10; ping-int 10;
            max-buffers 2048; max-epoch-size 2048; }
  on beo-121 {
    address sdp 192.168.41.121:7788;
    #address 192.168.41.121:7788;
    disk /dev/md3; device /dev/drbd0; meta-disk "internal";
  }
  on beo-122 {
    address sdp 192.168.41.122:7788;
    #address 192.168.41.122:7788;
    disk /dev/md3; device /dev/drbd0; meta-disk "internal";
  }
}
------------------

When using "address 192.168.41.122:7788;" instead of 
"address sdp 192.168.41.121:7788;" everything works flawlessly.

Trying the trick from
http://lists.linbit.com/pipermail/drbd-user/2010-September/014759.html
also doesn't help: If I do 

drbdsetup 0 net sdp:192.168.41.122:7788 ipv4:192.168.41.121:7788 C --set-defaults --create-device --timeout=60 --connect-int=10 --ping-int=10 --max-buffers=2048 --max-epoch-size=2048

I get
 
[7848.567907] _sdp_exch_state:478 sdp_sock( 5691:2 7788:33747): sdp_close:722: trying to exchange state from unexpected state TCP_FIN_WAIT1 to state TCP_TIME_WAIT. expected states: 0x102

with 

drbdsetup 0 net ipv4:192.168.41.122:7788 sdp:192.168.41.121:7788 C --set-defaults --create-device --timeout=60 --connect-int=10 --ping-int=10 --max-buffers=2048 --max-epoch-size=2048

I get

[ 8005.238845] block drbd0: connect failed, err = -97

How can I get SDP working?

Thanks,

Roland