Hi,
I probably found a minor security hole (as it may be IMHO) in my (latest) stable 
release of SAMBA (2.2.9a-1)

Using Win98SE I am logging to Samba server. I am using root preexec and root postexec 
on Home share's to maintain a list of "correctly logged" users.
When a Windows workstation craches, Samba finds this after some 5 long minutes. Until 
that time, although a new user logs into the same workstation, it STILL doesn't make a 
"POSTEXEC" operation on the previous share !!! (Still alive)

Example of current behaviour - in-out log of current Samba release (IMHO bad)
EVID [2003/07/31 15:04:20] hugo BEG z IP:195.113.84.205  (* PREEXEC -> Starting first 
connection, RESET at 15:06 approx.)
EVID [2003/07/31 15:07:55] hugo BEG z IP:195.113.84.205  (* PREEXEC -> Logging again 
to the station)
EVID [2003/07/31 15:11:29] hugo END z IP:0.0.0.0  (* POSTEXEC -> The first connection 
finaly falls)  (See @#$% lower)

I have studied log-files of previous stable versions, where the same problem ALWAYS 
gave "FALL" before "NEW LOG-IN". No idea how he did it or why the current does not.
(Example of in-out log with older samba release : )
EVID [2003/01/14 14:01:27] hugo BEG z IP:194.108.183.205 (First log-in, CRASH)
EVID [2003/01/16 14:55:35] hugo END z IP:0.0.0.0 (Huh, the comp fell down!)
EVID [2003/01/16 14:55:51] hugo BEG z IP:194.108.183.205 (And now I log-in again)
In older releases, there were ALWAYS at least 5 seconds between FALL (0.0.0.0) and 
following log-in.



@#$%   And, the last piece of my message, the part of log-file talking about "Yes, the 
old connection is dead"  @#$%

[2003/07/31 15:11:29, 0] lib/util_sock.c:read_data(436)
  read_data: read failure for 4. Error = Connection reset by peer
[2003/07/31 15:11:29, 0] lib/util_sock.c:get_socket_addr(1012)
  getpeername failed. Error was Transport endpoint is not connected
[2003/07/31 15:11:29, 0] lib/util_sock.c:get_socket_addr(1012)
  getpeername failed. Error was Transport endpoint is not connected
[2003/07/31 15:11:29, 0] lib/util_sock.c:get_socket_addr(1012)
  getpeername failed. Error was Transport endpoint is not connected
[2003/07/31 15:11:29, 0] lib/util_sock.c:get_socket_addr(1012)
  getpeername failed. Error was Transport endpoint is not connected


Please let me know your opinion and (eventualy) solution. No conditions changed during 
the whole year, except of Samba releases.

Best regards, Vladimir Jakubal
IT Gymnasium Jana Keplera, Prague, Czech Republic
--
To unsubscribe from this list go to the following URL and read the
instructions:  http://lists.samba.org/mailman/listinfo/samba

Reply via email to