Slow Mappings with NIS and large groups

Lee Cashmore lee.cashmore at
Sun Sep 24 13:11:09 GMT 2000

Does anyone have any experiance of these problems

OS Redhat 6.2
Samba version 2.0.7

Machine runs a slave NIS server and has 2 sets of groups one for staff and
one for students e.g.


Student groups follow the same format but are groups  stud0 to stud9

all the groups for each user are correctly reported by linux and I can find
no other problems and logon speed under linux is fine. The problem we have
is that we have upto 550 pc's mapping to a share which runs off this box
(Details of share below). All the main logons happen within a short period
of time i.e. when the students login in a morning and after breaks, for
some reason it is taking a long time for any windows NT machine to map a
drive to this particular share.

Samba is doing Domain Authentication and after doing various amounts of
testing and debuging I can find no way of improving the speed of the drive
mapping process. We are thus experiancing 2 problems, firstly a lot of the
drive mappings will fail (with a 240 error) when lots of people are logging
on and secondly the box is put under so much load due to the exessive
amounts of time samba spends checking against the groups. If I copy the
password, shadow and group files off the primary nis server and put them on
the local machine the drive mappings are almost immediate. We presently
have 1000 or so students in the groups and this will double if not tripple
by the end of the year, There are also around 300 staff accounts.

This share is a specific one which allows students Read only access and
staff read/write access. the files in the share are owned by the user/group
shares/shares respectivly.

Any help or ideas would be greatly appreciated

Lee Cashmore

P.S. Relavant sections of configs and log files listed below

Share config from smb.conf

     comment = Staff/Student Share
     path = /var/exports/qdrive
     browseable = yes
     writable = yes
     available = yes
     public = no
     admin users = leec
     read list = +staff0, +staff1, +staff2, +staff3, +staff4, +stud0,
     +stud1, +stud2, +stud3, +stud4, +stud5, +stud6, +stud7, +stud8, +stud9
     valid users = +staff0, +staff1, +staff2, +staff3, +staff4, +stud0,
     +stud1, +stud2, +stud3, +stud4, +stud5, +stud6, +stud7, +stud8, +stud9
     write list = +staff0, +staff1, +staff2, +staff3, +staff4
     force user = shares
     force group = shares
     force create mode = 770
     force directory mode = 770

Section of Log file of NT 4 Wks Mapping to the samba server

[2000/09/24 13:10:19.810779, 5] lib/util.c:show_msg(446)
[2000/09/24 13:10:19.810825, 10] lib/util.c:dump_data(2918)
   [000] 00 5C 5C 49 4E 54 52 41  4E 45 54 5C 51 44 52 49  .\\INTRA NET\QDRI
[2000/09/24 13:10:19.810925, 10] lib/util.c:dump_data(2926)
   [010] 56 45 00 41 3A 00                                 VE.A:.
[2000/09/24 13:10:19.811011, 3] smbd/process.c:switch_message(448)
   switch message SMBtconX (pid 15697)
[2000/09/24 13:10:19.811075, 4] smbd/reply.c:reply_tcon_and_X(311)
   Got device type A:		
[2000/09/24 13:10:27.032062, 3] smbd/password.c:authorise_login(759)
   ACCEPTED: validated uid ok as non-guest  <<------- PROBLEM HERE TIME
   DIFF OF 8 seconds ??? [2000/09/24 13:10:28.810178, 3]
smbd/service.c:make_connection(380) Forced user shares
[2000/09/24 13:10:28.810741, 3] smbd/service.c:make_connection(428)
   Forced group shares
[2000/09/24 13:10:28.810813, 3] smbd/service.c:make_connection(441)
   Connect path is /var/exports/qdrive
[2000/09/24 13:10:28.869361, 3] smbd/password.c:setup_groups(192)
   shares is in 1 groups: 1004
[2000/09/24 13:10:28.869495, 5] smbd/connection.c:claim_connection(137)
   trying claim /var/lock/samba STATUS. 100000
[2000/09/24 13:10:28.869607, 8] lib/util.c:fcntl_lock(2672)

Lee Cashmore                              Telephone : +44 (0)1952 642286
Network Administrator (Postmaster)
Telford College of Arts & Technology    Web Site :

More information about the samba mailing list