Description of problem:
After restarting the services (cman, clvmd) the clvmd service startup fails and freezes at "Activating VGs:"
These messages are visible in /var/log/messages:
Apr  1 04:13:55 a2 clogd[26973]: [NqVaQc2F] Failed to open checkpoint: SA_AIS_ERR_LIBRARY 
Apr  1 04:13:55 a2 clogd[26973]: [NqVaQc2F] Failed to import checkpoint from 3 
Apr  1 04:13:55 a2 kernel: clogd(26973): unaligned access to 0x6000000000010bdc, ip=0x4000000000056821
Apr  1 04:14:10 a2 kernel: device-mapper: dm-log-clustered: [NqVaQc2F] Request timed out: [DM_CLOG_RESUME/25] - retrying

The last one keeps repeating with the number (25) increasing.

Version-Release number of selected component (if applicable):
lvm2-cluster-2.02.40-7.el5
cman-2.0.98-1.el5
Linux a3 2.6.18-128.el5 #1 SMP Wed Dec 17 11:44:28 EST 2008 ia64 ia64 ia64 GNU/Linux

How reproducible:
every time

Steps to Reproduce:
1. reboot two nodes (tested with 2nodes cluster)
2. nodes form quorum and operate as expected, vgs up and visible
3. login to first node and do: service clvmd stop; service cman stop; service cman start; service clvmd start
4. the last command freezes as described
  
Actual results:
service clvmd start freezes and log keeps repeating DM_CLOG_RESUME error

Expected results:
clvmd should come up, properly register to cluster and activate all VGS as it does when it reboots

Additional info:
# cat /etc/cluster/cluster.conf
<?xml version="1.0"?>
<cluster name="a-cluster" config_version="8">
                <cman two_node="1">
                </cman>
                <fence_daemon post_join_delay="20" post_fail_delay="20" clean_start="1"/>

                <clusternodes>
<!--                    <clusternode name="a1" votes="1" nodeid="1">
                                <fcdriver>lpfc</fcdriver>
                                <arch>ia64</arch>
                        </clusternode>
-->                     <clusternode name="a2" votes="1" nodeid="2">
                                <fcdriver>lpfc</fcdriver>
                                <arch>ia64</arch>
                        </clusternode>
                        <clusternode name="a3" votes="1" nodeid="3">
                                <fcdriver>lpfc</fcdriver>
                                <arch>ia64</arch>
                        </clusternode>
                </clusternodes>

                <fencedevices>
                        <fencedevice name="apc1" agent="fence_apc" ipaddr="link-apc" login="xxx" passwd="xxx"/>
                </fencedevices>

        </cluster>

All following commands are run from the 'healthy' node which still sees all VGS:
# group_tool 
type             level name     id       state       
fence            0     default  00010002 none        
[2 3]
dlm              1     clvmd    00020002 none        
[2 3]

# cman_tool status
Version: 6.1.0
Config Version: 8
Cluster Name: a-cluster
Cluster Id: 43956
Cluster Member: Yes
Cluster Generation: 420
Membership state: Cluster-Member
Nodes: 2
Expected votes: 2
Total votes: 2
Quorum: 1  
Active subsystems: 8
Flags: 2node Dirty 
Ports Bound: 0 11  
Node name: a3
Node ID: 3
Multicast addresses: 239.192.171.96 
Node addresses: 10.15.89.181 

# vgs
  VG         #PV #LV #SN Attr   VSize   VFree  
  GFSMIRROR    3   1   0 wz--nc   1.45T   1.43T
  VolGroup00   1   2   0 wz--n- 136.62G      0 
  clustervg    1   1   0 wz--n- 495.71G 485.71G

# lvs
  LV       VG         Attr   LSize   Origin Snap%  Move Log         Copy%  Convert
  GFSM-1   GFSMIRROR  mwi-a-  10.00G                    GFSM-1_mlog 100.00        
  LogVol00 VolGroup00 -wi-ao 126.81G                                              
  LogVol01 VolGroup00 -wi-ao   9.81G                                              
  first    clustervg  -wi-a-  10.00G