[Samba] Is Samba made for trash ?!?

[Michael Paarmann]

Hi Sambafriends !

I have done a migration from Novell to a Samba-PDC in a medium enterprise 
with nearly 50 workstations eight weeks ago.

When i look at the /var/log/warn listing i still see some errors, and i 
like to ask you, if this messages are "normal" ?

smbd[7809]: [2003/02/11 18:26:45, 0] 
smbd/password.c:domain_client_validate(1517)
smbd[7809]:   domain_client_validate: could not fetch trust account 
password for domain OFFICE

smbd[4902]: [2003/02/11 18:32:32, 0] lib/util_sock.c:read_data(436)
smbd[4902]:   read_data: read failure for 4. Error = Connection reset by peer

smbd[3535]: [2003/02/11 18:33:28, 0] lib/util_sock.c:read_data(436)
smbd[3535]:   read_data: read failure for 4. Error = No route to host

smbd[30829]: [2003/02/11 14:02:15, 0] 
rpc_server/srv_netlog.c:api_net_sam_logon(206)
smbd[30829]:   api_net_sam_logon: Failed to marshall NET_R_SAM_LOGON.

smbd[30829]: [2003/02/11 14:02:15, 0] rpc_server/srv_pipe.c:api_rpcTNP(1200)
smbd[30829]:   api_rpcTNP: api_netlog_rpc: NET_SAMLOGON failed.

smbd[3777]: [2002/12/23 12:03:43, 0] lib/util_sock.c:write_socket_data(500)
smbd[3777]:   write_socket_data: write failure. Error = Broken pipe

The workstations run with Win98, Win 2000 and Win XP Pro. I have diabled 
all oplocks (because i had a lot of oplock-failures). The workstation now 
hang more often than with the Novell-Server. I fear, that the read_date and 
the write_socket_data errors cause this hangs.
And after i have disabled the oplocks and the socket options Samba is very 
slow.
The same hardware has worked under Novell very well and more stable in 
inaction with the workstations.

I'm frustrated - please help.

Thanx.

Michael