Samba - windows client log off problem [RESENO]

[zveckoo]

Imam problem sledece vrste:

Linux serveru (suse 11.0 / samba) pristupaju razni klijenti bazirani na WINXP 32bita, WINXP 64 bita, WIN 2003 server.. itd....
Pristup sambi je organizovan preko username/passworda.
I sve mi savrseno radi, svako ima svoj nivo pristupa i zna se ko moze sta da radi na serveru ( samba serovanim folderima).

Problem nastaje kada WINDOWS pukne na nekim od klijent racunara... Ne desava se cesto... ali windows je to, i normalno desava se da neko uspe da zaglavi i krahira svoj racunar.

E tada nastaje sou posto ako je taj klijent drzo neki fajl na serveru otvoren prilikom pada windowsa, - ponovnog logovanja ( restarta glupog windowsa) Klijent vise ne moze da pristupi tom fajlu - znaci nema r/w prava na taj fajl posto samba i dalje pila da taj fajl neko drzi otvoren... iako je taj klijent racunar nasilno restartovan.

Tada jedino pomaze restart sambe "etc/init.d/samba restart" ( restart windows klijenta ne pomaze).

samba cfg fajl mi izgleda otprilike ovako:

[global]
    workgroup = HIT
    server string = Server
    security = user
    username map = /etc/samba/smbusers
    encrypt passwords = yes
    map to guest = Bad User
    guest account = nobody
    log file = /hitard/logs/samba.log
    max log size = 100000
    local master = yes
    os level = 33
[final]
    path = /hit/http/final
    public =yes
    browseable  = yes
    write list = @korisnici
    vfs objects = recycle full_audit
    recycle:keeptree = yes
    recycle:versions = yes
    recycle:exclude = *.bak *.dwl *.log *.tmp *.dce *.dwl2 *.err
    recycle:touch = yes
    recycle:maxsize = 100000000
    recycle:directory_mode = 0775
    create mask = 0774
    directory mask = 0775
    force group = korisnici
    full_audit:prefix = %u|%I
    full_audit:success = write mkdir rename unlink rmdir pwrite
    full_audit:failure = none
    full_audit:facility = LOCAL7
    full_audit:priority = ALERT
[archive]
    path = /hit/http/archive
    public =yes
    browseable  = yes
    write list = @admini
    vfs objects = recycle full_audit
    recycle:keeptree = yes
    recycle:versions = yes
    recycle:exclude = *.bak *.dwl *.log *.tmp *.dce *.dwl2 *.err
    recycle:touch = yes
    recycle:maxsize = 100000000
    recycle:directory_mode = 0775
    create mask = 0774
    directory mask = 0775
    force group = admini
    full_audit:prefix = %u|%I
    full_audit:success = write mkdir rename unlink rmdir pwrite
    full_audit:failure = none
    full_audit:facility = LOCAL7
    full_audit:priority = ALERT

 

[GatZ]

Probaj da u [global] postaviš keepalive opciju na određeni broj sekundi. Samba će u zavisnosti od parametra, tj. broja sekundi koji postaviš čekirati status klijenta (active/inactive), i u teoriji, šutnuti usera koji se zombirao.

Npr,

[global]
    keepalive = 60

 

[zveckoo]

Hvala,

Probacu preko vikenda da eksperimentisem , sada mi je previse veliki cirkus u firmi da bi se igro sa serverom.

Nasao sam i neke parametre za ubrzanje sambe... pa cu i to da probam:

socket options = TCP_NODELAY IPTOS_LOWDELAY


The value of the parameter (an integer) represents the number of seconds between keepalive packets. If this parameter is zero, no keepalive packets will be sent. Keepalive packets, if sent, allow the server to tell whether a client is still present and responding.

Keepalives should, in general, not be needed if the socket has the SO_KEEPALIVE attribute set on it by default. (see socket options). Basically you should only use this option if you strike difficulties.

Default: keepalive = 300

Example: keepalive = 600

 

[zveckoo]

Izgleda je proradilo

Stavio sam

keepalive = 60
socket options = TCP_NODELAY IPTOS_LOWDELAY

A na windows klijentu sam ukljucio

enable: NetBios over TCP/IP ( posto koristim staticke IP adrese)

Probo sam da zakucam ( ugasim nasilno) windows klijent i upalim ga posle dva minuta... i sada bez problema pristupa fajlu za koje je pre inace samba i dalje tvrdila da ga klijent taj i taj drzi drzi otvoren....