[Samba] 2.2.6 leaking session ids?

Wed Oct 30 16:08:01 GMT 2002

is this normal? my user opened a database and then smbd closes connections
and shuts down. obviously the user receives an error on the other end
regarding system resources....

[2002/10/30 10:48:36, 2] smbd/open.c:open_file(245)
  DUCHENAN opened file shared/tempdb/temp_db.mdb read=Yes write=Yes
(numopen=3)
[2002/10/30 10:48:36, 2] smbd/open.c:open_file(245)
  DUCHENAN opened file shared/tempdb/temp_db.ldb read=Yes write=Yes
(numopen=4)
[2002/10/30 10:48:36, 2] smbd/close.c:close_normal_file(213)
  duchenan closed file shared/tempdb/temp_db.ldb (numopen=3)
[2002/10/30 10:48:36, 2] smbd/open.c:open_file(245)
  DUCHENAN opened file shared/tempdb/temp_db.ldb read=Yes write=Yes
(numopen=4)
[2002/10/30 10:48:36, 0] lib/util_sock.c:read_data(436)
  read_data: read failure for 4. Error = Connection reset by peer
[2002/10/30 10:48:36, 2] smbd/server.c:exit_server(461)
  Closing connections
[2002/10/30 10:48:36, 1] smbd/service.c:close_cnum(677)
  nsduche (151.103.28.166) closed connection to service nextstep
[2002/10/30 10:48:36, 2] smbd/close.c:close_normal_file(213)
  duchenan closed file shared/tempdb/temp_db.ldb (numopen=3)
[2002/10/30 10:48:36, 2] smbd/close.c:close_normal_file(213)
  duchenan closed file shared/tempdb/temp_db.mdb (numopen=2)
[2002/10/30 10:48:36, 2] smbd/close.c:close_normal_file(213)
  duchenan closed file shared/9xdbase/9xdbase-OLD/system.ldb (numopen=1)
[2002/10/30 10:48:36, 2] smbd/close.c:close_normal_file(213)
  duchenan closed file shared/9xdbase/9xdbase-OLD/system.mda (numopen=0)

then it gets even more strange....

i get this as it attempts to reestablish...

[2002/10/30 10:48:36, 2] lib/interface.c:add_interface(81)
  added interface ip=151.103.16.28 bcast=151.103.31.255
nmask=255.255.240.0
[2002/10/30 10:48:36, 2] lib/access.c:check_access(329)
  Allowed connection from  (151.103.28.166)
[2002/10/30 10:48:36, 2] smbd/reply.c:reply_special(92)
  netbios connect: name1=STAFF            name2=NSDUCHE
[2002/10/30 10:48:36, 2] smbd/reply.c:reply_special(111)
  netbios connect: local=staff remote=nsduche

followed by

[2002/10/30 10:48:37, 2] lib/interface.c:add_interface(81)
  added interface ip=151.103.16.28 bcast=151.103.31.255
nmask=255.255.240.0
[2002/10/30 10:48:37, 2] lib/access.c:check_access(329)
  Allowed connection from  (151.103.28.166)
[2002/10/30 10:48:37, 0] smbd/service.c:make_connection(381)
  make_connection: duchenan logged in as admin user (root privileges)
[2002/10/30 10:48:37, 1] smbd/service.c:make_connection(636)
  nsduche (151.103.28.166) connect to service nextstep as user duchenan
(uid=0,
gid=1) (pid 117036)
[2002/10/30 10:48:37, 1] lib/debug.c:debug_message(258)
  INFO: Debug class all level = 2   (pid 104458 from pid 104458)

followed by

[2002/10/30 10:48:37, 2] lib/interface.c:add_interface(81)
  added interface ip=151.103.16.28 bcast=151.103.31.255
nmask=255.255.240.0
[2002/10/30 10:48:37, 1] lib/debug.c:debug_message(258)
  INFO: Debug class all level = 2   (pid 140448 from pid 140448)
[2002/10/30 10:48:37, 1] [lib/debug.c2002/10/30 10:48:37:,
debug_message2(] 258p
aram/loadparm.c)
:do_section(3055)
  INFO: Debug class all level = 2   (pid 40486 from pid 40486)
  Processing section "[homes]"

which is not docotored....the previous is really that screwed up at
"debug_message"

which is then followed by

[2002/10/30 10:48:37, 2] lib/interface.c:add_interface(81)
  added interface ip=151.103.16.28 bcast=151.103.31.255
nmask=255.255.240.0
[2002/10/30 10:48:37, 2] lib/interface.c:add_interface(81)
  added interface ip=151.103.16.28 bcast=151.103.31.255
nmask=255.255.240.0
[2002/10/30 10:48:38, 1] lib/debug.c:debug_message(258)
  INFO: Debug class all level = 2   (pid 85220 from pid 85220)

followed by *dozens* of sections just like the one above for the next six
seconds of the log....

something is very very wrong here....

Bill

On Wed, 30 Oct 2002, William Jojo wrote:

> 
> 
> i'm seeing stranger stuff today...especially with 98 machines accessing
> Excel spreadsheets and databases.
> 
> we not using any oplocks at all and there are no special options set.
> 
> i'm going to try a log level 10 and then go back to 2.2.5.
> 
> where should i post the log when i have it along with my smb.conf?
> 
> 
> Bill
> 
> 
> On Tue, 29 Oct 2002, William Jojo wrote:
> 
> > 
> > 
> > i checked my error log and it is not recording any core dumps for smbd.
> > 
> > i do not think it is crashing since the smbd seems to keep running and
> > trying to grant access to different shares, but the user ends up with only
> > a couple or none depending on if it can get a sessionid.
> > 
> > i'll reimplement tomorrow and take a stronger log level...
> > 
> > 
> > Bill
> > 
> > 
> > On Wed, 30 Oct 2002, Andrew Bartlett wrote:
> > 
> > > William Jojo wrote:
> > > > 
> > > > moved back to 2.2.5 and the messages regarding "Failed to claim
> > > > session" are gone....
> > > > 
> > > > this one server has the greatest load as 500+ workstations are constantly
> > > > logging on and off every hour. our lighter servers are not exhibiting this
> > > > problem under 2.2.6, so I have not downgraded them.
> > > > 
> > > > we are running AIX 5.1.0-02 and clients are WinXP Pro (no sp1).
> > > 
> > > I don't think the sessionids are your real problem - that code hasn't
> > > changed - but is smbd crashing?  That might be why it never cleans up
> > > the sessionid record.
> > > 
> > > Andrew Bartlett
> > > 
> > > -- 
> > > Andrew Bartlett                                 abartlet at pcug.org.au
> > > Manager, Authentication Subsystems, Samba Team  abartlet at samba.org
> > > Student Network Administrator, Hawker College   abartlet at hawkerc.net
> > > http://samba.org     http://build.samba.org     http://hawkerc.net
> > > 
> > 
> > -- 
> > To unsubscribe from this list go to the following URL and read the
> > instructions:  http://lists.samba.org/mailman/listinfo/samba
> > 
> 
> -- 
> To unsubscribe from this list go to the following URL and read the
> instructions:  http://lists.samba.org/mailman/listinfo/samba
> 