[Samba] PDC: random problems, especially NETLOGON script not always loading

Koen Linders koen.news at koca.be
Fri Feb 1 11:03:13 GMT 2008


Debian Latest stable.
Linux newton 2.6.18-4-686
Samba 3.0.24
PDC

3Com 3812 Gigabit switch (connection between WinXP client & server)
Others connect through 3Com 3225 100 Mbit -> 3Com 3812 -> Server

Windows XP SP2 fully updated.
F-Secure client security (look down for firewall settings)

In advance, thanks for reading this. Any suggestions are welcome!

I'm having a hard time here. I read a whole lot, browses, searched. I try  
to provide as much info as possible but if you need more, Let me know.
The main problem is the "randomness". Or atleast it looks random for me.

People with mapped shares working for +1 year don't have connection  
problems. So most of the time it seems to work fine.
But i really want to troubleshoot & get all errors out.

I guess the problem i describe next also is the cause of random Roaming  
profile problems, but lets not focus on that part, too many random factors  
in my opinion. And they were only a part of the test. It might work if i  
get this problem solved.

I got a script in the netlogon mapping drives dynamically depending on  
groups.
I don't want them permanent. I don't want a workaround!

Everything works fine, most of the time...
Sometimes it seems the test pc doesn't see the netlogon during boot.
But it's still accesible when browsing to the share and is executable.

Through policy i tried setting:
1)I tried setting detecting slow network (1Gbit though) on.
or
2) Also run script synchronously
No changes.

Firewall:
Even when i change allow all trafic to Samba server in both directions.

Firewall has all necessary port open in both directions:
SMB (TCP)		445	SMB over TCP/IP (TCP)	
SMB (UDP)		445	SMB over TCP/IP (UDP)
Windows Networking (1)	137-138	Both broadcast and multicast	Windows  
network browsing	
Windows Networking (2)	139	Windows file sharing and network printers	
WINS(1)			42	Both broadcast and multicast	WINS / Windows Internet Name  
Service (UDP)	
WINS(2)			42	WINS / Windows Internet Name Service (TCP)	

Windows doesn't show an error in the log, except the autoenrollment one,  
which is normal (no AD).
Samba log file when it goes wrong (a bit lower).

The error when it goes wrong:

1) Error writing 5 bytes to client. -1. (Connection reset by peer):
=> has to do with client going over NETBIOS (139) & 445 and closing one of  
2 connections
Nothing wrong here.
The weird part: For test: i blocked 137-138-139 on firewall and i couldn't  
connect to samba share...

Server is listening on both 139 & 445:

netstat -an | egrep '(137|138|139|445)'
tcp        0      0 0.0.0.0:139             0.0.0.0:*               LISTEN
tcp        0      0 0.0.0.0:445             0.0.0.0:*               LISTEN
tcp        0      0 192.168.1.2:139         192.168.1.59:1075        
ESTABLISHED
tcp        0      0 192.168.1.2:445         192.168.1.82:4409        
ESTABLISHED
tcp        0      0 192.168.1.2:445         192.168.1.4:40578        
ESTABLISHED
tcp        0      0 192.168.1.2:445         192.168.1.44:3465        
ESTABLISHED
tcp        0      0 192.168.1.2:139         192.168.1.109:1209       
ESTABLISHED
udp        0      0 192.168.1.2:137         0.0.0.0:*
udp        0      0 0.0.0.0:137             0.0.0.0:*
udp        0      0 192.168.1.2:138         0.0.0.0:*
udp        0      0 0.0.0.0:138             0.0.0.0:*


2) lib/util_sock.c:write_data(562)
   write_data: write failure in writing to client 192.168.1.98. Error  
Connection reset by peer
=> Could this error point somewhere?

Samba log:

[2008/01/29 13:13:27, 0] lib/util_sock.c:get_peer_addr(1229)
   getpeername failed. Error was Transport endpoint is not connected
[2008/01/29 13:13:27, 0] lib/access.c:check_access(327)
[2008/01/29 13:13:27, 0] lib/util_sock.c:get_peer_addr(1229)
   getpeername failed. Error was Transport endpoint is not connected
   Denied connection from  (0.0.0.0)
[2008/01/29 13:13:27, 1] smbd/process.c:process_smb(1103)
[2008/01/29 13:13:27, 0] lib/util_sock.c:get_peer_addr(1229)
   getpeername failed. Error was Transport endpoint is not connected
   Connection denied from 0.0.0.0
[2008/01/29 13:13:27, 0] lib/util_sock.c:write_data(562)
   write_data: write failure in writing to client 192.168.1.98. Error  
Connection reset by peer
[2008/01/29 13:13:27, 0] lib/util_sock.c:send_smb(769)
   Error writing 5 bytes to client. -1. (Connection reset by peer)
[2008/01/29 13:13:27, 1] smbd/service.c:make_connection_snum(950)
   mpi057 (192.168.1.98) connect to service profiles initially as user  
verah (uid=1003, gid=1001) (pid 12835)
[2008/01/29 13:13:27, 1] smbd/service.c:close_cnum(1150)
   mpi057 (192.168.1.98) closed connection to service profiles
[2008/01/29 13:13:30, 1] smbd/service.c:make_connection_snum(950)


Smb.conf:

#======================= Global Settings =======================

[global]
	netbios name = NEWTON
	workgroup = KOCALM
	domain master = yes
	domain logons = yes
	local master = yes
  	preferred master = yes
	os level = 65

	wins support = yes
	name resolve order = wins lmhosts host bcast

	time server = yes

	security = user
	encrypt passwords = true

	browse list = yes
	browseable = no
	max disk size = 20480

	
	# [globals] User and group related

	idmap gid = 15000-20000
	add group script = /usr/sbin/groupadd %g
	delete group script = /usr/sbin/groupdel %g

	idmap uid = 15000-20000
	add user script = /usr/sbin/useradd -m %u
	delete user script = /usr/sbin/userdel -r %u
	
	add machine script = /usr/sbin/useradd -s /bin/false -d /dev/null -g  
machines %u

	passdb backend = tdbsam
	hosts allow = 192.168.1. 127.0.0.1	

#	logon home = \\NEWTON\%U			
	logon path = \\NEWTON\profiles\%U 		
	logon script = logon.bat	
#	logon drive = H:
	

#	printing = cups
	printcap name = cups

[netlogon]
    	comment = Network Logon Service
    	path = /data/netlogon
    	guest ok = yes
	browseable = No



More information about the samba mailing list