[Samba] smbfs issue when viewing a share from a MS Cluster

[Westerdale, John]

All,
 
I am digging into an smbfs issue where we have a single server (RHES3 U3
kernel) mounting a
share from a Microsoft Cluster.  
 
It happens once or twice each night around 8:40 and then around 10:40.
 
Doesnt seem to be related to any other services (backup etc).
 
Syslog shows:
 
Jan 28 20:33:30 servername kernel: SMB server not responding
Jan 28 20:33:30 servername kernel: smb_get_length: recv error = 5
Jan 28 20:33:30 servername kernel: smb_trans2_request: result=-5,
setting invalid
Jan 28 20:33:30 servername mount.smbfs[1153]: [2007/01/28 20:33:30, 0]
tdb/tdbutil.c:tdb_log(725)
Jan 28 20:33:30 servername mount.smbfs[1153]:
tdb(/var/cache/samba/gencache.tdb): tdb_lock failed on list 76 ltype=0
(Bad file descriptor)
Jan 28 20:33:50 servername mount.smbfs[1153]: [2007/01/28 20:33:50, 0]
client/smbmount.c:do_connection(237)
Jan 28 20:33:50 servername mount.smbfs[1153]:   1153: tree connect
failed: Call timed out: server did not respond after 20000 milliseconds
Jan 28 20:34:03 servername mount.smbfs[1153]: [2007/01/28 20:34:03, 0]
tdb/tdbutil.c:tdb_log(725)
Jan 28 20:34:03 sosbllxa02 mount.smbfs[1153]:
tdb(/var/cache/samba/gencache.tdb): tdb_lock failed on list 76 ltype=0
(Bad file descriptor)
Jan 28 20:34:23 servername mount.smbfs[1153]: [2007/01/28 20:34:23, 0]
client/smbmount.c:do_connection(192)
Jan 28 20:34:23 servername mount.smbfs[1153]:   1153: protocol
negotiation failed

and then resolves
 
Jan 28 20:37:36 sosbllxa02 mount.smbfs[1153]: [2007/01/28 20:37:36, 0]
tdb/tdbutil.c:tdb_log(725)
Jan 28 20:37:36 servername mount.smbfs[1153]:
tdb(/var/cache/samba/gencache.tdb): tdb_lock failed on list 76 ltype=0
(Bad file descriptor)
Jan 28 20:37:54 servername kernel: smb_retry: successful, new pid=16125,
generation=4
Jan 28 20:37:54 servername kernel: smb_lookup: find //* failed, error=-5
Jan 28 20:37:55 servername kernel: smb_retry: successful, new pid=1153,
generation=10
 
 
I have a cron script that does a "net rap share" to show advertised
shares.
 
When the SMBFS file system is unavailable, the "net rap share" line
evokes: 
 
   "Could not connect to server cluster123"
 
 
The outage lasts between 3 and 6 minutes and disappears.
 
No entries in either of the 3 MS Event-Viewer topics.
 
 
 
Has anyone seen or debugged this issue?
 
Mount made with: mount -t smbfs
 
Server: rpm -qf /sbin/mount.smbfs
samba-client-3.0.7-1.3E.1
Server: 
 
Am hoping this is the proper forum for smbfs question, it involves
samba-client :^)
 
 
Thanks Very Much
 
John Westerdale
Mahwah, NJ