URGENT: request_oplock_break failing ... time out

[Mark E. Drummond]

I started playing ... despite warnings not to I increased the oplock
break wait time to 20 from it's default of 10. I am no longer seeing the
previous messages however my users are still having problems. Bu they
seem to be connecting just fine. Here is some log data for one machine
at log level 3:

[1999/10/21 11:35:25, 1] smbd/service.c:make_connection(521)
  gpnte222a (137.94.128.75) connect to service s0001 as user s0001
(uid=399, gid=34) (pid 8230)
[1999/10/21 11:35:25, 3] lib/doscalls.c:dos_ChDir(336)
  dos_ChDir to /local/local/samba/var/log
[1999/10/21 11:35:25, 3] smbd/reply.c:reply_tcon_and_X(341)
  tconX service=s0001 user=s0001
[1999/10/21 11:35:25, 3] smbd/process.c:process_smb(615)
  Transaction 3 of length 106
[1999/10/21 11:35:25, 3] smbd/process.c:switch_message(448)
  switch message SMBntcreateX (pid 8230)
[1999/10/21 11:35:25, 3] lib/doscalls.c:dos_ChDir(336)
  dos_ChDir to /home/s0001
[1999/10/21 11:35:25, 3] lib/util.c:unix_clean_name(609)
  unix_clean_name [/CLIENTUS/IGEN.INV]
[1999/10/21 11:35:25, 3] lib/util.c:unix_clean_name(609)
  unix_clean_name [clientus/IGEN.INV]

Everything looks OK to me. The connection seem sto have been successful
and it, plus other connections from that box, all have smbd's running
for them. trussing any of those processes shows them as sleeping, even
though the user is trying to browse the shares in question.

I can turn off oplocks but what about performance? Besides which, this
has been working just fine for ages. Well, actually I just upgraded from
2.0.4 to 2.0.5a, _but_ I have had 2.0.5a running on my second file
server for months without issue.

HELP!

-- 
Gang Warily