[OpenIndiana-discuss] samba keeps dumping core on oi_147

[Christopher Chan]

Hi all,

I have a samba installation that is part of a AD domain and uses opends 
for its ldap backend both for winbind and samba.

I keep getting problems with smbd. No problems with winbindd however. On 
a separate note, why is nmbd not started? Anyway, back to smbd.

The logs have entries such as this:

[2010/10/26 09:41:05.325972,  3] smbd/server.c:259()
   smbd/server.c:258 Unclean shutdown of pid 20181

and this:

[2010/10/26 09:41:05.326256,  1] smbd/server.c:267()
   Scheduled cleanup of brl and lock database after unclean shutdown

and stuff like this:

[2010/10/26 09:41:25.327120,  2] lib/messages_local.c:289()
   message to process 19991 failed - No such process
[2010/10/26 09:41:25.327219,  2] lib/messages_local.c:379()
   pid 19991 doesn't exist - deleting messages record
[2010/10/26 09:41:25.327318,  2] lib/messages.c:127()
   pid 19991 doesn't exist - deleting connections -1 []

core dumps can be found for some of these pids

Finally, all smbd processes disappear and I find logs such as these in 
the SMF logs:

[ Oct 26 09:13:10 Leaving maintenance because clear requested. ]
[ Oct 26 09:13:10 Enabled. ]
[ Oct 26 09:13:10 Executing start method ("/usr/sbin/smbd -D"). ]
[ Oct 26 09:13:10 Method "start" exited with status 0. ]
[ Oct 26 09:32:19 Stopping because process dumped core. ]
[ Oct 26 09:32:19 Executing stop method ("/usr/bin/kill `cat 
/var/samba/locks/smbd.pid`"). ]
[ Oct 26 09:32:19 Method "stop" exited with status 0. ]
[ Oct 26 09:32:20 Executing start method ("/usr/sbin/smbd -D"). ]
[ Oct 26 09:32:20 Method "start" exited with status 0. ]
[ Oct 26 09:32:27 Stopping because process dumped core. ]
[ Oct 26 09:32:27 Executing stop method ("/usr/bin/kill `cat 
/var/samba/locks/smbd.pid`"). ]
[ Oct 26 09:32:27 Method "stop" exited with status 0. ]
[ Oct 26 09:32:33 Executing start method ("/usr/sbin/smbd -D"). ]
[ Oct 26 09:32:33 Method "start" exited with status 0. ]
[ Oct 26 09:33:10 Stopping because process dumped core. ]
[ Oct 26 09:33:10 Executing stop method ("/usr/bin/kill `cat 
/var/samba/locks/smbd.pid`"). ]
[ Oct 26 09:33:10 Method "stop" exited with status 0. ]
[ Oct 26 09:33:17 Executing start method ("/usr/sbin/smbd -D"). ]
[ Oct 26 09:33:17 Method "start" exited with status 0. ]
[ Oct 26 09:33:24 Stopping because process dumped core. ]
[ Oct 26 09:33:24 Executing stop method ("/usr/bin/kill `cat 
/var/samba/locks/smbd.pid`"). ]
[ Oct 26 09:33:24 Method "stop" exited with status 0. ]
[ Oct 26 09:33:31 Restarting too quickly, changing state to maintenance. ]

Having to go and run svcadm clear samba is not fun for a production 
system...

Backtrace of sample core dump:
  pstack smbd.19991.bradsuper1
core 'smbd.19991.bradsuper1' of 19991:  /usr/sbin/smbd -D
  feb72297 _lwp_kill (1, 6, 8046ee8, feb19f5e) + 7
  feb19f6a raise    (6, 0, 8046f38, feaf19fa) + 22
  feaf1a1a abort    (86f02b9, 8992a14, 8046f78) + f2
  083a6059 dump_core (13, 8a59dd0, 0) + 211
  083b8e9a smb_panic (8719318, 8992a14, 8046f98, 819b07c) + 12e
  0819b08f set_unix_security_ctx (c399, c357, 13, 8a59dd0) + 47
  0819b15d set_sec_ctx (c399, c357, 13, 8a59dd0, 8a599d0, 64) + b5
  0818b69e change_to_user (8a594f8, 64, 8a53780, 871fc28) + 2ee
  081b1728 make_connection_snum (8a353b0, 15, 8a437e8, 8a5cb78, 1, 
80471f0) + a8c
  081b22ef make_connection (8a353b0, 8a8b17d, 8a5cb78, 1, 8a8b210, 64) + 4cf
  08167867 reply_tcon_and_X (8a8b050, 8992a14, 80475f8, 81adbd8) + 23f
  081adcba switch_message (75, 8a8b050, 5a, 0) + 3ca
  081ade86 construct_reply (0, 5a, 0, 0, 0, 0) + de
  081ae0ad process_smb (8a353b0, 8a8af70, 5a, 0, 0, 0) + 135
  081aeedc smbd_server_connection_read_handler (8a353b0, 871dc40, 0, 
81aeef4) + 9c
  081aef2d smbd_server_connection_handler (8a35338, 8a37e30, 1, 8a353b0) 
+ 45
  083c923c run_events (8a35338, 1, 8047770, 80477f0) + 1e8
  081ad56f smbd_server_connection_loop_once (8a353b0, 8992a14, 8047948, 
81af85e) + 10f
  081af86c smbd_process (0, 0, 40, 10, 2b060002, 1366a8c0) + 6c4
  086dd67d smbd_accept_connection (8a35338, 8a85bc8, 1, 8a85b78) + 209
  083c923c run_events (8a35338, 1, 8047af0, 8047b70) + 1e8
  083c943d s3_event_loop_once (8a35338, 898033c, 8821ffc, 83c9e9a) + 111
  083c9f05 _tevent_loop_once (8a35338, 898033c, 8047c78, 86ddf16) + 79
  086ddf22 smbd_parent_loop (8a856c0) + 82
  086df0ce main     (2, 8047e54, 8047e60, 8133baf) + bea
  08133c0d _start   (2, 8047ef8, 8047f07, 0, 8047f0a, 8047f1e) + 7d

Please let me know what I can do to help identify the problem and get 
this fixed. BTW, this also keeps appearing everytime samba is restarted:

rlimit_max: rlimit_max (256) below minimum Windows limit (16384)