[Samba] Samba (RHEL 3) hangs after few days or even hours

[Zdenek Pizl]

Hallo all,

 Last week I am trying to solve strange problem. Suddenly without no 
system config change the samba stops work properly.

Symptoms:
- it hangs after few days or hours (a worse case), there are many 
processes (from hundreds to thousands of smbd processes)
- no new connection can be established
- it happend regardless of samba package version, I've tried all 
available versions from 3.0.6-2.3E to 3.0.14a-1


System info:
- RHEL 3 Update 4 + available updates
- computer is a domain member controlled by another samba (2.2.8a) with 
ldap backend)
- problematic computer is a file server under medium/high load (about 50 
clients read and write to it gigabytes of data)

/etc/samba/smb.conf
# Global parameters
[global]
        workgroup = SYSTINET
        server string = Buildsrv
        security = DOMAIN
        password server = 10.0.0.3
        encrypted passwords = yes

        log level = 10
        log file = /var/log/samba/%m.log
        max log size = 5000

        wins server = 10.0.0.3
        smb ports = 139 445
        name resolve order = hosts wins bcast

        max smbd processes = 5000
        socket options = IPTOS_LOWDELAY TCP_NODELAY SO_RCVBUF=8192 
SO_SNDBUF=8192 SO_KEEPALIVE

        load printers = No
        dns proxy = No
        ldap ssl = no

        read raw = no
        level2 oplocks = true

        domain master = no
        domain logons = no
        local master  = no

        dont descend = /proc,/dev,/lost+found,/BUILD/lost+found

[buildsrv]
        comment = Share to store data etc.
        path = /BUILD
        force user = builder
        force group = users
        read only = No
        create mask = 0644
        directory mask = 0775
        force directory mode = 0775
        guest ok = Yes
        use sendfile = Yes
        write cache size = 65536

I noticed weird messages in nmbd.log:

[2005/05/15 02:50:27, 4] 
nmbd/nmbd_workgroupdb.c:find_workgroup_on_subnet(162)
  find_workgroup_on_subnet: workgroup search for SYSTINET on subnet 
UNICAST_SUBNET: found.
[2005/05/15 02:50:27, 4] 
nmbd/nmbd_workgroupdb.c:find_workgroup_on_subnet(162)
  find_workgroup_on_subnet: workgroup search for SYSTINET on subnet 
UNICAST_SUBNET: found.
[2005/05/15 02:50:27, 10] lib/util_sock.c:read_udp_socket(230)
  read_udp_socket: lastip 10.0.0.209 lastport 138 read: 215
[2005/05/15 02:50:27, 5] libsmb/nmblib.c:read_packet(757)
  Received a packet of len 215 from (10.0.0.209) port 138
[2005/05/15 02:50:27, 10] nmbd/nmbd_subnetdb.c:namelist_entry_compare(69)
nmbd_subnetdb:namelist_entry_compare()
-1 == memcmp( "SYSTINET<1d>", "SYSTINET<1e>", 84 )
[2005/05/15 02:50:27, 10] nmbd/nmbd_subnetdb.c:namelist_entry_compare(69)
nmbd_subnetdb:namelist_entry_compare()
1 == memcmp( "SYSTINET<1d>", "SYSTINET<00>", 84 )
[2005/05/15 02:50:27, 9] nmbd/nmbd_namelistdb.c:find_name_on_subnet(129)
  find_name_on_subnet: on subnet 10.0.0.56 - name SYSTINET<1d> NOT FOUND
[2005/05/15 02:50:27, 5] nmbd/nmbd_packets.c:process_dgram(1194)
  process_dgram: ignoring dgram packet sent to name SYSTINET<1d> from 
10.0.0.209


[2005/05/15 09:04:26, 4] nmbd/nmbd_workgroupdb.c:dump_workgroups(271)
  dump_workgroups()
   dump workgroup on subnet       10.0.0.56: netmask=  255.255.254.0:
        SYSTINET(1) current master browser = AUTHCZ
                BUILDSRV 40009b03 (Buildsrv)
                AUTHCZ 400c9b2b (AuthCZ PDC)
[2005/05/15 09:04:26, 4] nmbd/nmbd_workgroupdb.c:dump_workgroups(271)
  dump_workgroups()
   dump workgroup on subnet  UNICAST_SUBNET: netmask=        0.0.0.0:
        SYSTINET(1) current master browser = UNKNOWN
                BUILDSRV 40009b03 (Buildsrv)
[2005/05/15 09:04:26, 4] 
nmbd/nmbd_workgroupdb.c:find_workgroup_on_subnet(162)
  find_workgroup_on_subnet: workgroup search for SYSTINET on subnet 
UNICAST_SUBNET: found.
[2005/05/15 09:04:26, 4] 
nmbd/nmbd_workgroupdb.c:find_workgroup_on_subnet(162)
  find_workgroup_on_subnet: workgroup search for SYSTINET on subnet 
UNICAST_SUBNET: found.
[2005/05/15 09:04:36, 4] 
nmbd/nmbd_workgroupdb.c:find_workgroup_on_subnet(162)
  find_workgroup_on_subnet: workgroup search for SYSTINET on subnet 
10.0.0.56: found.
[2005/05/15 09:04:36, 10] 
nmbd/nmbd_sendannounce.c:announce_myself_to_domain_master_browser(382)
  announce_myself_to_domain_master_browser: t (1116140666) - 
last(1116139866) < 900
[2005/05/15 09:04:36, 4] nmbd/nmbd_workgroupdb.c:dump_workgroups(271)
  dump_workgroups()
   dump workgroup on subnet       10.0.0.56: netmask=  255.255.254.0:
        SYSTINET(1) current master browser = AUTHCZ
                BUILDSRV 40009b03 (Buildsrv)
                AUTHCZ 400c9b2b (AuthCZ PDC)
[2005/05/15 09:04:36, 4] nmbd/nmbd_workgroupdb.c:dump_workgroups(271)
  dump_workgroups()
   dump workgroup on subnet  UNICAST_SUBNET: netmask=        0.0.0.0:
        SYSTINET(1) current master browser = UNKNOWN
                BUILDSRV 40009b03 (Buildsrv)



Does anybody know what's going on there? I have no clue.
Thank you in advance, z.p.

-- 
Zdenek Pizl
Systinet Corporation
Vinohradska 190
13000 Praha 3