NET3126: The network NET1 ran out of NCBs. - OS2

This is a discussion on NET3126: The network NET1 ran out of NCBs. - OS2 ; I'm getting this on a WSeB (kernel 4.5x) system. I finally got around to installing my NetBIOS and TcpBEUI side by side together (on two different [logical] adapters) and at first it seemed to work fine. No erros in the ...

+ Reply to Thread
Page 1 of 2 1 2 LastLast
Results 1 to 20 of 25

Thread: NET3126: The network NET1 ran out of NCBs.

  1. NET3126: The network NET1 ran out of NCBs.

    I'm getting this on a WSeB (kernel 4.5x) system.

    I finally got around to installing my NetBIOS and TcpBEUI side by side
    together (on two different [logical] adapters) and at first it seemed
    to work fine.
    No erros in the Lantran.log file.
    But even so I can start the requester fine, I can't get the server to
    come up afterwards.
    It breaks with some weird error message about being unable to start the
    LOGON subservice and this is the error message I see first in the
    error log:

    >NET3126: The network NET1 ran out of NCBs. You may need to increase

    NCBs for
    >this network in the IBMLAN.INI file. The following information

    includes the
    >number of NCBs submitted by the server when this error occurred:


    This is the 2nd error in the log:
    >NET3101: The system ran out of a resource controlled by the NUMREQBUF

    option.

    Now no matter how I play with the IBMLAN.INI and PROTOCOL.INI values,
    the result is the same (if not making it all worse).
    And I already tried to set NUMREQBUF to the max (300) w/o success.
    Anybody got any ideas on this?

    DD


  2. Re: NET3126: The network NET1 ran out of NCBs.

    dumpfdumpf@yahoo.com wrote:
    > I'm getting this on a WSeB (kernel 4.5x) system.
    >
    > I finally got around to installing my NetBIOS and TcpBEUI side by side
    > together (on two different [logical] adapters) and at first it seemed
    > to work fine.
    > No erros in the Lantran.log file.
    > But even so I can start the requester fine, I can't get the server to
    > come up afterwards.
    > It breaks with some weird error message about being unable to start the
    > LOGON subservice and this is the error message I see first in the
    > error log:
    >
    >
    >>NET3126: The network NET1 ran out of NCBs. You may need to increase

    >
    > NCBs for
    >
    >>this network in the IBMLAN.INI file. The following information

    >
    > includes the
    >
    >>number of NCBs submitted by the server when this error occurred:

    >
    >
    > This is the 2nd error in the log:
    >
    >>NET3101: The system ran out of a resource controlled by the NUMREQBUF

    >
    > option.
    >
    > Now no matter how I play with the IBMLAN.INI and PROTOCOL.INI values,
    > the result is the same (if not making it all worse).
    > And I already tried to set NUMREQBUF to the max (300) w/o success.
    > Anybody got any ideas on this?
    >
    > DD
    >


    can you post your protocol.ini and ibmlan.ini?

    Rule of thumb: the values in IBMLAN.INI on the NET1= and NET2= lines
    should be smaller than the corresponding values in PROTOCOL.INI

  3. LAN Server diagnostics (was Re: NET3126: The network NET1 ran outof NCBs.)

    I'm having problem with a WSeB, after some time it stops working
    (vanishes from the LAN, refuses both local or remote logons), however I
    don't know how to diagnose what's wrong. Is there any way to diagnose
    out-of-the-box WSeB? Unfortunately, it is not reproducible and it occurs
    from time to time.

    Only the server stops working. NET STOP REQ - NET START SERVER resumes
    the networking. I don't know what exactly might me the problem as the
    TCP/IP keeps working - I can telnet into the stuck server and restart
    the networking. There's one NIC with TCP/IP and plain NETBIOS.

    Jeroen, can you recommend something, please?

    Cheers,
    Martin


    Jeroen Besse wrote:
    > dumpfdumpf@yahoo.com wrote:
    >
    >> I'm getting this on a WSeB (kernel 4.5x) system.
    >>
    >> I finally got around to installing my NetBIOS and TcpBEUI side by side
    >> together (on two different [logical] adapters) and at first it seemed
    >> to work fine.
    >> No erros in the Lantran.log file.
    >> But even so I can start the requester fine, I can't get the server to
    >> come up afterwards.
    >> It breaks with some weird error message about being unable to start the
    >> LOGON subservice and this is the error message I see first in the
    >> error log:
    >>
    >>
    >>> NET3126: The network NET1 ran out of NCBs. You may need to increase

    >>
    >>
    >> NCBs for
    >>
    >>> this network in the IBMLAN.INI file. The following information

    >>
    >>
    >> includes the
    >>
    >>> number of NCBs submitted by the server when this error occurred:

    >>
    >>
    >>
    >> This is the 2nd error in the log:
    >>
    >>> NET3101: The system ran out of a resource controlled by the NUMREQBUF

    >>
    >>
    >> option.
    >>
    >> Now no matter how I play with the IBMLAN.INI and PROTOCOL.INI values,
    >> the result is the same (if not making it all worse).
    >> And I already tried to set NUMREQBUF to the max (300) w/o success.
    >> Anybody got any ideas on this?
    >>
    >> DD
    >>

    >
    > can you post your protocol.ini and ibmlan.ini?
    >
    > Rule of thumb: the values in IBMLAN.INI on the NET1= and NET2= lines
    > should be smaller than the corresponding values in PROTOCOL.INI


  4. Re: NET3126: The network NET1 ran out of NCBs.

    My PROTOCOL.INI:
    [PROT_MAN]

    DRIVERNAME = PROTMAN$

    [VLAN_KERNEL]

    DRIVERNAME = VLANKNL$
    CFGTYPE = "LOCKED"
    MODE = "LEARN_WFLOWS"
    LANTYPE = "802.5"
    MAXADDRESSES = 3500

    [IBMLXCFG]

    landd_nif = landd.nif
    odi2ndi_nif = odi2ndi.nif
    vproto_nif = vproto.nif
    netbeui_nif = netbeui.nif
    tcpbeui_nif = tcpbeui.nif
    tcpip_nif = tcpip.nif
    FXWRAP_nif = fxwrap.nif
    E100BEO2_nif = e100beo2.nif
    IBMPRISM_nif = ibmprism.nif
    VSWITCH_nif = vswitch.nif
    GCIP100_nif = ntrace.nif
    GCPWIFI_nif = ntrace.nif
    GCIVPC_nif = ntrace.nif

    [NETBIOS]

    DriverName = netbios$
    ADAPTER0 = netbeui$,0
    ADAPTER1 = tcpbeui$,1

    [NETBEUI_nif]

    DriverName = netbeui$
    Bindings = FXWRAP1_nif
    ETHERAND_TYPE = "I"
    USEADDRREV = "YES"
    OS2TRACEMASK = 0x0
    SESSIONS = 70
    NCBS = 180
    NAMES = 22
    SELECTORS = 20
    USEMAXDATAGRAM = "NO"
    ADAPTRATE = 1000
    WINDOWERRORS = 0
    MAXDATARCV = 4352
    TI = 60000
    T1 = 10000
    T2 = 2000
    MAXIN = 1
    MAXOUT = 1
    NETBIOSTIMEOUT = 500
    NETBIOSRETRIES = 3
    NAMECACHE = 1000
    RNDOPTION = 1
    PIGGYBACKACKS = 1
    DATAGRAMPACKETS = 50
    PACKETS = 350
    LOOPPACKETS = 8
    PIPELINE = 5
    MAXTRANSMITS = 6
    MINTRANSMITS = 2
    DLCRETRIES = 10
    FCPRIORITY = 5
    NETFLAGS = 0x1000

    [landd_nif]

    DriverName = LANDD$
    Bindings = FXWRAP1_nif,FXWRAP2_nif
    ETHERAND_TYPE = "I","I"
    SYSTEM_KEY = 0x0,0x0
    OPEN_OPTIONS = 0x2000,0x2000
    TRACE = 0x0,0x0
    LINKS = 12,12
    MAX_SAPS = 6,6
    MAX_G_SAPS = 0,0
    USERS = 5,5
    TI_TICK_G1 = 255,255
    T1_TICK_G1 = 15,15
    T2_TICK_G1 = 3,3
    TI_TICK_G2 = 255,255
    T1_TICK_G2 = 25,25
    T2_TICK_G2 = 10,10
    IPACKETS = 250,250
    UIPACKETS = 100,100
    MAXTRANSMITS = 6,6
    MINTRANSMITS = 2,2
    TCBS = 64,64
    GDTS = 30,30
    ELEMENTS = 1100,1100
    NETFLAGS = 0x0,0x0

    [odi2ndi_nif]

    DriverName = odi2ndi$
    Bindings = FXWRAP1_nif
    NETADDRESS = "00028A35E66B"
    TOKEN-RING = "no"
    TOKEN-RING_SNAP = "no"
    ETHERNET_802.3 = "yes"
    ETHERNET_802.2 = "yes"
    ETHERNET_II = "no"
    ETHERNET_SNAP = "no"
    TRACE = 0x0

    [vproto_nif]

    DriverName = PSWCH00$
    Bindings = ,FXWRAP2_nif

    [tcpbeui_nif]

    DriverName = tcpbeui$
    Bindings = ,FXWRAP2_nif
    NODETYPE = "B-Node"
    OS2TRACEMASK = 0x0
    SESSIONS = 70
    NCBS = 64
    NAMES = 21
    SELECTORS = 15
    USEMAXDATAGRAM = "NO"
    NETBIOSTIMEOUT = 500
    NETBIOSRETRIES = 2
    NAMECACHE = 1000
    PURGECACHE = 0
    PRELOADCACHE = "NO"
    NAMESFILE = 2
    DATAGRAMPACKETS = 20
    PACKETS = 50
    ENABLEDNS = 0
    INTERFACERATE = 300

    [tcpip_nif]

    DriverName = TCPIP$
    Bindings = FXWRAP1_nif,FXWRAP2_nif

    [E100BEO2_nif]

    DriverName = E100B$
    SPEED = 0
    FORCEDUPLEX = 0
    TRANSMITBUFFERS = 20
    RECEIVEBUFFERS = 20

    [IBMPRISM_nif]

    DriverName = PRISMP$
    NETADDRESS = "0020E08F89B9"
    NODE_ID = "freetown"
    DESIRED_SSID = "homestead"
    STA_MODE = "BSS"
    CREATE_IBSS = "NO"
    WEP = "DISABLED"
    AUTH_TYPE = "OPEN_SYSTEM"
    RADIO_ON = "YES"

    [VSWITCH_nif]

    DriverName = DSWCH00$
    FEATURE = "NO"

    [GCIP100_nif]

    DriverName = GCIP100$
    ModuleType = "Service"
    Bindings = FXWRAP1_nif
    MaxBufSegments = 32
    RejectFrames = 5

    [GCPWIFI_nif]

    DriverName = GCPWIFI$
    ModuleType = "Service"
    Bindings = IBMPRISM_nif
    MaxBufSegments = 32
    RejectFrames = 5

    [GCIVPC_nif]

    DriverName = GCIVPC$
    ModuleType = "Service"
    Bindings = VSWITCH_nif
    MaxBufSegments = 32
    RejectFrames = 5

    [FXWRAP1_nif]

    Drivername = FXWRAP1$
    Bindings = E100BEO2_nif

    [FXWRAP2_nif]

    Drivername = FXWRAP2$
    Bindings = IBMPRISM_nif



    My IBMLAN.INI:

    ; IBM OS/2 Warp Server for e-business network initialization file

    [networks]

    NET1 = NETBEUI$,0,LM10,64,180,20
    NET2 = TCPBEUI$,1,LM10,64,64,20
    ;
    ParamsriverID,LogicalAdapterNr,??,NrOfNetBIOSsessions,Nr OfNetBIOScmds,NrOfNetBIOSnames
    (R.M.F.)
    ; All these values must be =< than their counterparts specified in
    PROTOCOL.INI (R.M.F.)
    ; This information is read by the redirector at device initialization
    time.

    [requester]

    Computername = DUMPF
    Domain = DUMPFDUMPF
    ; The following parameters generally do not need to be
    ; changed by the user.
    charcount = 16
    chartime = 250
    charwait = 3600
    keepconn = 600
    keepsearch = 600
    maxcmds = 16
    maxerrorlog = 100
    maxthreads = 10
    maxwrkcache = 64
    numalerts = 12
    numcharbuf = 10
    numservices = 16
    numworkbuf = 40
    numdgrambuf = 14
    othdomains =
    printbuftime = 90
    sesstimeout = 300
    sizcharbuf = 512
    sizerror = 1024
    ; sizworkbuf = 4096
    sizworkbuf = 2048
    useallmem = Yes
    ; The wrkheuristics parameter sets a variety of requester fine-tuning
    ; options. Brief descriptions are provided here for each digit. More
    ; complete descriptions, along with explanations of relationships
    between
    ; digits can be found in the LAN Server Network Administrator Reference
    ; Volume 2: Performance Tuning.
    ;
    ; When not otherwise specified, a value of 0 means off (inactive) and 1
    means
    ; on (active).
    ;
    ; 0 opportunistic locking of files (default is 1)
    ; 1 performance optimization for batch (.CMD) files (default is 1)
    ; 2 asynchronous unlock and asynchronous write-unlock
    ; 0 = never 2 = only on a LAN Server virtual
    circuit
    ; 1 = always (default)
    ; 3 asynchronous close and asynchronous write-close
    ; 0 = never 2 = only on a LAN Server virtual
    circuit
    ; 1 = always (default)
    ; 4 buffer all named pipes and serial devices (default is 1)
    ; 5 combined read-lock and write-unlock
    ; 0 = never 2 = only on a LAN Server virtual
    circuit
    ; 1 = always (default)
    ; 6 open and read optimization (default is 1)
    ; 7 reserved
    ; 8 use of the chain-send NETBIOS NCB
    ; 0 = never
    ; 1 = when a server's buffer is larger than the workstations
    buffer
    ; 2 = always (default)
    ; 9 buffer small read and write requests until the buffer is full
    ; 0 = never 2 = only on a LAN Server virtual
    circuit
    ; 1 = always (default)
    ; 10 buffer mode
    ; 0 = always read buffer size amount of data if the request is
    smaller
    ; than the buffer size (sizworkbuf) and data is being read
    ; sequentially
    ; 1 = use full buffer if file is open for reading and writing
    ; 2 = use full buffer if reading and writing sequentially
    ; 3 = buffer all requests smaller than the buffer size (if hits
    occur)
    ; (default)
    ; 11 RAW read and RAW write SMB protocols (default is 1)
    ; 12 RAW read-ahead buffer (default is 1)
    ; 13 RAW write-behind buffer (default is 1)
    ; 14 read multiplexing SMB protocols (default is 1)
    ; 15 write multiplexing SMB protocols (default is 1)
    ; 16 use of big buffers for large core (non-RAW) reads (default is 1)
    ; 17 same-size read-ahead or read-to-sector boundary (default is 1)
    ; 18 same-size small record write-behind or write-to-sector boundary
    ; (default is 0)
    ; 19 reserved and must be 0
    ; 20 flush pipes and devices on a DosBufReset or DosClose operation
    ; 0 = Flush only files and devices opened by the caller. Spin
    until
    ; flushed. Wait for confirmation before processing with
    other
    ; tasks. (default)
    ; 1 = Flush only files and devices opened by the caller. Flush
    only
    ; once. Do not wait for confirmation.
    ; 2 = Flush all files and all input and output of short-term
    pipes
    ; and devices. Spin until flushed.
    ; 3 = Flush all files and all input and output of short-term
    pipes
    ; and devices. Flush only once.
    ; 4 = Flush all files and all input and output of all pipes and
    devices.
    ; Spin until flushed.
    ; 5 = Flush all files and all input and output of all pipes and
    devices.
    ; Flush only once.
    ; 21 LAN Server encryption of passwords (default is 1)
    ; 22 control log entries for multiple occurrences of an error
    ; 0 = log all occurrences (default)
    ; 1-9 = limit occurrences that are logged (1-9 define size of
    table
    ; that is used to track errors)
    ; 23 buffer all files opened with deny-write sharing mode (default is
    1)
    ; 24 buffer all files opened with the read-only attribute set (default
    is 1)
    ; 25 read ahead when opening for execution (default is 1)
    ; 26 handle the interrupt (ctrl+c) key
    ; 0 = allow no interrupts
    ; 1 = allow interrupts only on long-term operations
    ; 2 = always allow interrupts (default)
    ; 27 force correct open mode when creating files on a core server
    (reserved
    ; for DBCS users) (default is 1)
    ; 28 NETBIOS NoAck mode
    ; 0 = NoAck is disabled.
    ; 1 = NoAck is set on send only. (default)
    ; 2 = NoAck is set on receive only.
    ; 3 = NoAck is set on send and receive.
    ; 29 send data along with SMB write-block RAW requests (default is 1)
    ; 30 send a message to the screen when the requester logs an error
    ; 0 = never
    ; 1 = on write-fault errors only (no time out) (default)
    ; 2 = on write-fault and internal errors only (no time out)
    ; 3 = on all errors (no time out)
    ; 4 = reserved
    ; 5 = on write-fault errors only (time out)
    ; 6 = on write-fault and internal errors only (time out)
    ; 7 = on all errors (time out)
    ; 8 = on all errors (time out) and pop up an error if a session
    ; to a server is abnormally disconnected
    ; 9 = on all errors (time out) and generate an INT24 (critical
    ; error) to applications running under DOS emulation if a
    ; network hard error is encountered
    ; 31 reserved
    ; 32 behavior of DosBufReset on a redirected file (not pipes or
    devices).
    ; When the call to the API returns, the data in the buffer has been
    ; handled as follows:
    ; 0 = Changed data in the buffer was sent from the requester to
    the
    ; server. The server has written the data to disk.
    ; 1 = Changed data in the buffer was sent from the requester to
    the
    ; server. The server has not yet written the data to disk.
    ; 2 = DosBufReset was ignored. (default)
    ; 33 time interval for performing logon validation from the domain
    ; controller
    ; 0 = 5 seconds 4 = 60 seconds 8 = 8
    minutes
    ; 1 = 15 seconds (default) 5 = 90 seconds 9 = 15
    minutes
    ; 2 = 30 seconds 6 = 2 minutes
    ; 3 = 45 seconds 7 = 4 minutes
    ; 34 date validation
    ; 0 = PCLP date format (default)
    ; 1 = MSNET date format
    ; 2 = assume date is correct
    ; 35 free disk space reported to DOS and Windows applications
    ; 0 = return true value (default)
    ; 1 = return a value less than two gigabytes
    ; 36 time and date synchronization with the domain controller at logon
    ; (default is 1)
    ; 37 type of verification for the LAN Server logon
    ; 0 = no verification
    ; 1 = verify against local NET.ACC
    ; 2 = verify against domain NET.ACC (default)
    ; 38 display warning messages for the LAN Server logon
    ; 0 = do not display warning messages
    ; 1 = display all warning messages (default)
    ; 2 = do not display LAN Server specific warning messages
    ; 3 = display all warning messages, including password
    expiration date
    ; 39 buffer all files opened in compatibility mode
    ; 0 = buffer only files opened for read access in compatibility
    mode
    ; 1 = buffer all files opened in compatibility mode (default)
    ; 40 allow a user to logon multiple times in the same domain (default
    is 1)
    ; 41 client ignores failures setting extended attributes
    ; 0 = Extended Attribute failures (e.g. to backlevel servers)
    are
    ; reported to the application (default)
    ; 1 = Failures setting extended attributes are ignored
    ; 42 boost priority of mailslot handling thread to time critical
    (default is 0)
    ;
    ; The next lines help you to locate bits in the wrkheuristics entry.
    ; 1 2 3 4
    ; 0123456789012345678901234567890123456789012
    wrkheuristics = 1111111121311111110001011120111221001211100
    WRKSERVICES =
    wrknets = NET1,NET2

    [messenger]

    logfile = messages.log
    sizmessbuf = 4096

    [netlogon]

    SCRIPTS = D:\IBMLAN\REPL\IMPORT\SCRIPTS
    pulse = 60
    update = yes

    [replicator]

    replicate = IMPORT
    IMPORTPATH = D:\IBMLAN\REPL\IMPORT
    tryuser = yes
    password =
    interval = 5
    guardtime = 2
    pulse = 3
    random = 60

    [dcdbrepl]

    tryuser = yes
    password =
    interval = 5
    guardtime = 2
    pulse = 3
    random = 60

    [server]

    alertnames =
    auditing = resource
    autodisconnect = 120
    maxusers = 16
    ; The following parameters generally do not need to be
    ; changed by the user. NOTE: srvnets= is represented in
    ; the server info struct as a 16-bit lan mask. Srvnet names
    ; are converted to indexes within [networks] for the named nets.
    guestacct = guest
    accessalert = 5
    alertsched = 5
    diskalert = 5000
    erroralert = 5
    keepdossearch = 600
    logonalert = 5
    maxauditlog = 100
    maxchdevjob = 48
    maxchdevq = 2
    maxchdevs = 16
    maxlocks = 64
    maxopens = 256
    maxsearches = 50
    maxsessopens = 64
    ; maxsessopens = 256
    maxsessreqs = 50
    maxsessvcs = 1
    maxshares = 192
    netioalert = 5
    numbigbuf = 31
    numfiletasks = 2
    ; numreqbuf = 96
    numreqbuf = 300
    othsrvnames =
    sizreqbuf = 2048
    srvanndelta = 3000
    srvannounce = 180
    srvhidden = no
    ; The srvheuristics parameter sets a variety of server fine-tuning
    ; options. Brief descriptions are provided here for each digit. More
    ; complete descriptions, along with explanations of relationships
    between
    ; digits can be found in the LAN Server Network Administrator Reference
    ; Volume 2: Performance Tuning.
    ;
    ; When not otherwise specified, a value of 0 means off (inactive) and 1
    means
    ; on (active).
    ;
    ; 0 opportunistic locking of files (default is 1)
    ; 1 read-ahead when the requester is performing sequential access
    ; 0 = do not use read-ahead
    ; 1 = use single read-ahead thread (default)
    ; 2 = use asynchronous read-ahead thread
    ; 2 write-behind (default is 1)
    ; 3 use of the chain-send NETBIOS NCB (default is 1)
    ; 4 check all incoming SMBs for correct format (default is 0)
    ; 5 support FCB opens (default is 1)
    ; 6 set priority of server (default is 4)
    ; 0-9 = 0 is highest priority, 9 is lowest priority
    ; 7 automatically allocate additional memory for directory searches
    ; (default is 1)
    ; 8 write records to the audit log only when the scavenger wakes up
    ; (default is 1)
    ; 9 do full buffering when a file is opened with deny-write sharing
    mode
    ; (default is 1)
    ; 10 set the interval for the scavenger to wake up
    ; 0 = 5 seconds 4 = 25 seconds 8 = 45
    seconds
    ; 1 = 10 seconds (default) 5 = 30 seconds 9 = 50
    seconds
    ; 2 = 15 seconds 6 = 35 seconds
    ; 3 = 20 seconds 7 = 40 seconds
    ; 11 allow compatibility-mode opens of certain types of files by
    translating
    ; them to sharing mode opens with deny-none
    ; 0 = Always use compatibility-mode opens.
    ; 1 = Use deny-none sharing mode if read-only access to .EXE or
    ..COM
    ; files.
    ; 2 = Use deny-none sharing mode if read-only access to .EXE or
    ..COM
    ; files. Use deny-write sharing mode if read-only access to
    ..BAT
    ; files is requested.
    ; 3 = Use deny-none sharing mode on all compatibility-mode opens.
    ; (default)
    ; 12 allow DOS LAN Services workstations to use a second NETBIOS
    session when
    ; sending printer requests (default is 1)
    ; 13 number of 64KB buffers used for read-ahead
    ; 0-9 = number of buffers (default is 1)
    ; 14 convert incoming path specifications into most basic format
    (default is 0)
    ; 15 Oplock Timeout and NETBIOS Acknowledgment Timeout (default is 0)
    ; 0 = 35 second Oplock Timeout; 34 second NETBIOS Acknowledgment
    Timeout
    ; 1 = 70 second Oplock Timeout; 69 second NETBIOS Acknowledgment
    Timeout
    ; 2 = 140 second Oplock Timeout; 127 second NETBIOS
    Acknowledgment Timeout
    ; 3 = 210 second Oplock Timeout; 127 second NETBIOS
    Acknowledgment Timeout
    ; 4 = 280 second Oplock Timeout; 127 second NETBIOS
    Acknowledgment Timeout
    ; 5 = 350 second Oplock Timeout; 127 second NETBIOS
    Acknowledgment Timeout
    ; 6 = 420 second Oplock Timeout; 127 second NETBIOS
    Acknowledgment Timeout
    ; 7 = 490 second Oplock Timeout; 127 second NETBIOS
    Acknowledgment Timeout
    ; 8 = 560 second Oplock Timeout; 127 second NETBIOS
    Acknowledgment Timeout
    ; 9 = 640 second Oplock Timeout; no NETBIOS Acknowledgment
    Timeout
    ; 16 validate IOCTLs (default is 1)
    ; 17 how long the server maintains unused, dynamic big buffers before
    freeing
    ; the memory.
    ; 0 = 0 seconds 4 = 5 minutes 8 = 1 hour
    ; 1 = 1 second 5 = 10 minutes 9 =
    indefinitely
    ; 2 = 10 seconds 6 = 20 minutes
    ; 3 = 1 minute (default) 7 = 40 minutes
    ; 18 how long the server waits after failing to allocate a big buffer
    before
    ; trying again.
    ; 0 = 0 seconds 3 = 1 minute (default)
    ; 1 = 1 second 4 = 5 minutes
    ; 2 = 10 seconds 5 = 10 minutes
    ; 19 RAW read and RAW write SMB protocols (default is 1)
    ; 20 server responds to announcement requests (default is 1)
    ;
    ; The next lines help you to locate bits in the srvheuristics entry.
    ; 1 2
    ; 012345678901234567890
    ; srvheuristics = 111101411113110913311
    ; srvheuristics = 011101411113130013311
    srvheuristics = 111101411113130913311
    SRVSERVICES = NETLOGON,LSSERVER
    ; SRVSERVICES = NETLOGON,LSSERVER,DCDBREPL,TIMESOURCE
    srvnets = NET1,NET2
    ; autopath = D:\IBMLAN\PROFILES\SRVAUTO.PRO
    ; autoprofile = load
    autopath =
    maxconnections = 300

    [alerter]

    sizalertbuf = 3072

    [browser]

    MaintainServerList = YES
    Quiet = NO

    [netrun]

    maxruns = 3
    runpath = D:\

    [lsserver]

    cleanup = no
    srvpipes = 5

    [services]

    ; Correlates name of service to pathname of service program.
    ; The pathname must be either
    ; 1) an absolute path (including the drive specification)
    ; OR
    ; 2) a path relative to the IBMLAN root
    alerter = services\alerter.exe
    dcdbrepl = services\dcdbrepl.exe
    genalert = services\genalert.exe
    lsserver = services\lsserver.exe
    messenger = services\msrvinit.exe
    netlogon = services\netlogon.exe
    netrun = services\runservr.exe
    replicator = services\replicat.exe
    requester = services\wksta.exe
    server = services\netsvini.exe
    timesource = services\timesrc.exe
    browser = services\browser.exe
    ups = services\ups.exe

    [ups]

    batterymsg = 600
    batterytime = 60
    cmdfile =
    cmdtimer = 30
    devicename = UPS_DEV
    lowbattery = warning
    messdelay = 5
    messtime = 120
    recharge = 100
    signals = 100
    voltlevels = 100
    ; IBM OS/2 Warp Server for e-business network initialization file

    [networks]

    NET1 = NETBEUI$,0,LM10,64,180,20
    NET2 = TCPBEUI$,1,LM10,64,64,20
    ;
    ParamsriverID,LogicalAdapterNr,??,NrOfNetBIOSsessions,Nr OfNetBIOScmds,NrOfNetBIOSnames
    (R.M.F.)
    ; All these values must be =< than their counterparts specified in
    PROTOCOL.INI (R.M.F.)
    ; This information is read by the redirector at device initialization
    time.

    [requester]

    Computername = DUMPF
    Domain = DUMPFDUMPF
    ; The following parameters generally do not need to be
    ; changed by the user.
    charcount = 16
    chartime = 250
    charwait = 3600
    keepconn = 600
    keepsearch = 600
    maxcmds = 16
    maxerrorlog = 100
    maxthreads = 10
    maxwrkcache = 64
    numalerts = 12
    numcharbuf = 10
    numservices = 16
    numworkbuf = 40
    numdgrambuf = 14
    othdomains =
    printbuftime = 90
    sesstimeout = 300
    sizcharbuf = 512
    sizerror = 1024
    ; sizworkbuf = 4096
    sizworkbuf = 2048
    useallmem = Yes
    ; The wrkheuristics parameter sets a variety of requester fine-tuning
    ; options. Brief descriptions are provided here for each digit. More
    ; complete descriptions, along with explanations of relationships
    between
    ; digits can be found in the LAN Server Network Administrator Reference
    ; Volume 2: Performance Tuning.
    ;
    ; When not otherwise specified, a value of 0 means off (inactive) and 1
    means
    ; on (active).
    ;
    ; 0 opportunistic locking of files (default is 1)
    ; 1 performance optimization for batch (.CMD) files (default is 1)
    ; 2 asynchronous unlock and asynchronous write-unlock
    ; 0 = never 2 = only on a LAN Server virtual
    circuit
    ; 1 = always (default)
    ; 3 asynchronous close and asynchronous write-close
    ; 0 = never 2 = only on a LAN Server virtual
    circuit
    ; 1 = always (default)
    ; 4 buffer all named pipes and serial devices (default is 1)
    ; 5 combined read-lock and write-unlock
    ; 0 = never 2 = only on a LAN Server virtual
    circuit
    ; 1 = always (default)
    ; 6 open and read optimization (default is 1)
    ; 7 reserved
    ; 8 use of the chain-send NETBIOS NCB
    ; 0 = never
    ; 1 = when a server's buffer is larger than the workstations
    buffer
    ; 2 = always (default)
    ; 9 buffer small read and write requests until the buffer is full
    ; 0 = never 2 = only on a LAN Server virtual
    circuit
    ; 1 = always (default)
    ; 10 buffer mode
    ; 0 = always read buffer size amount of data if the request is
    smaller
    ; than the buffer size (sizworkbuf) and data is being read
    ; sequentially
    ; 1 = use full buffer if file is open for reading and writing
    ; 2 = use full buffer if reading and writing sequentially
    ; 3 = buffer all requests smaller than the buffer size (if hits
    occur)
    ; (default)
    ; 11 RAW read and RAW write SMB protocols (default is 1)
    ; 12 RAW read-ahead buffer (default is 1)
    ; 13 RAW write-behind buffer (default is 1)
    ; 14 read multiplexing SMB protocols (default is 1)
    ; 15 write multiplexing SMB protocols (default is 1)
    ; 16 use of big buffers for large core (non-RAW) reads (default is 1)
    ; 17 same-size read-ahead or read-to-sector boundary (default is 1)
    ; 18 same-size small record write-behind or write-to-sector boundary
    ; (default is 0)
    ; 19 reserved and must be 0
    ; 20 flush pipes and devices on a DosBufReset or DosClose operation
    ; 0 = Flush only files and devices opened by the caller. Spin
    until
    ; flushed. Wait for confirmation before processing with
    other
    ; tasks. (default)
    ; 1 = Flush only files and devices opened by the caller. Flush
    only
    ; once. Do not wait for confirmation.
    ; 2 = Flush all files and all input and output of short-term
    pipes
    ; and devices. Spin until flushed.
    ; 3 = Flush all files and all input and output of short-term
    pipes
    ; and devices. Flush only once.
    ; 4 = Flush all files and all input and output of all pipes and
    devices.
    ; Spin until flushed.
    ; 5 = Flush all files and all input and output of all pipes and
    devices.
    ; Flush only once.
    ; 21 LAN Server encryption of passwords (default is 1)
    ; 22 control log entries for multiple occurrences of an error
    ; 0 = log all occurrences (default)
    ; 1-9 = limit occurrences that are logged (1-9 define size of
    table
    ; that is used to track errors)
    ; 23 buffer all files opened with deny-write sharing mode (default is
    1)
    ; 24 buffer all files opened with the read-only attribute set (default
    is 1)
    ; 25 read ahead when opening for execution (default is 1)
    ; 26 handle the interrupt (ctrl+c) key
    ; 0 = allow no interrupts
    ; 1 = allow interrupts only on long-term operations
    ; 2 = always allow interrupts (default)
    ; 27 force correct open mode when creating files on a core server
    (reserved
    ; for DBCS users) (default is 1)
    ; 28 NETBIOS NoAck mode
    ; 0 = NoAck is disabled.
    ; 1 = NoAck is set on send only. (default)
    ; 2 = NoAck is set on receive only.
    ; 3 = NoAck is set on send and receive.
    ; 29 send data along with SMB write-block RAW requests (default is 1)
    ; 30 send a message to the screen when the requester logs an error
    ; 0 = never
    ; 1 = on write-fault errors only (no time out) (default)
    ; 2 = on write-fault and internal errors only (no time out)
    ; 3 = on all errors (no time out)
    ; 4 = reserved
    ; 5 = on write-fault errors only (time out)
    ; 6 = on write-fault and internal errors only (time out)
    ; 7 = on all errors (time out)
    ; 8 = on all errors (time out) and pop up an error if a session
    ; to a server is abnormally disconnected
    ; 9 = on all errors (time out) and generate an INT24 (critical
    ; error) to applications running under DOS emulation if a
    ; network hard error is encountered
    ; 31 reserved
    ; 32 behavior of DosBufReset on a redirected file (not pipes or
    devices).
    ; When the call to the API returns, the data in the buffer has been
    ; handled as follows:
    ; 0 = Changed data in the buffer was sent from the requester to
    the
    ; server. The server has written the data to disk.
    ; 1 = Changed data in the buffer was sent from the requester to
    the
    ; server. The server has not yet written the data to disk.
    ; 2 = DosBufReset was ignored. (default)
    ; 33 time interval for performing logon validation from the domain
    ; controller
    ; 0 = 5 seconds 4 = 60 seconds 8 = 8
    minutes
    ; 1 = 15 seconds (default) 5 = 90 seconds 9 = 15
    minutes
    ; 2 = 30 seconds 6 = 2 minutes
    ; 3 = 45 seconds 7 = 4 minutes
    ; 34 date validation
    ; 0 = PCLP date format (default)
    ; 1 = MSNET date format
    ; 2 = assume date is correct
    ; 35 free disk space reported to DOS and Windows applications
    ; 0 = return true value (default)
    ; 1 = return a value less than two gigabytes
    ; 36 time and date synchronization with the domain controller at logon
    ; (default is 1)
    ; 37 type of verification for the LAN Server logon
    ; 0 = no verification
    ; 1 = verify against local NET.ACC
    ; 2 = verify against domain NET.ACC (default)
    ; 38 display warning messages for the LAN Server logon
    ; 0 = do not display warning messages
    ; 1 = display all warning messages (default)
    ; 2 = do not display LAN Server specific warning messages
    ; 3 = display all warning messages, including password
    expiration date
    ; 39 buffer all files opened in compatibility mode
    ; 0 = buffer only files opened for read access in compatibility
    mode
    ; 1 = buffer all files opened in compatibility mode (default)
    ; 40 allow a user to logon multiple times in the same domain (default
    is 1)
    ; 41 client ignores failures setting extended attributes
    ; 0 = Extended Attribute failures (e.g. to backlevel servers)
    are
    ; reported to the application (default)
    ; 1 = Failures setting extended attributes are ignored
    ; 42 boost priority of mailslot handling thread to time critical
    (default is 0)
    ;
    ; The next lines help you to locate bits in the wrkheuristics entry.
    ; 1 2 3 4
    ; 0123456789012345678901234567890123456789012
    wrkheuristics = 1111111121311111110001011120111221001211100
    WRKSERVICES =
    wrknets = NET1,NET2

    [messenger]

    logfile = messages.log
    sizmessbuf = 4096

    [netlogon]

    SCRIPTS = D:\IBMLAN\REPL\IMPORT\SCRIPTS
    pulse = 60
    update = yes

    [replicator]

    replicate = IMPORT
    IMPORTPATH = D:\IBMLAN\REPL\IMPORT
    tryuser = yes
    password =
    interval = 5
    guardtime = 2
    pulse = 3
    random = 60

    [dcdbrepl]

    tryuser = yes
    password =
    interval = 5
    guardtime = 2
    pulse = 3
    random = 60

    [server]

    alertnames =
    auditing = resource
    autodisconnect = 120
    maxusers = 16
    ; The following parameters generally do not need to be
    ; changed by the user. NOTE: srvnets= is represented in
    ; the server info struct as a 16-bit lan mask. Srvnet names
    ; are converted to indexes within [networks] for the named nets.
    guestacct = guest
    accessalert = 5
    alertsched = 5
    diskalert = 5000
    erroralert = 5
    keepdossearch = 600
    logonalert = 5
    maxauditlog = 100
    maxchdevjob = 48
    maxchdevq = 2
    maxchdevs = 16
    maxlocks = 64
    maxopens = 256
    maxsearches = 50
    maxsessopens = 64
    ; maxsessopens = 256
    maxsessreqs = 50
    maxsessvcs = 1
    maxshares = 192
    netioalert = 5
    numbigbuf = 31
    numfiletasks = 2
    ; numreqbuf = 96
    numreqbuf = 300
    othsrvnames =
    sizreqbuf = 2048
    srvanndelta = 3000
    srvannounce = 180
    srvhidden = no
    ; The srvheuristics parameter sets a variety of server fine-tuning
    ; options. Brief descriptions are provided here for each digit. More
    ; complete descriptions, along with explanations of relationships
    between
    ; digits can be found in the LAN Server Network Administrator Reference
    ; Volume 2: Performance Tuning.
    ;
    ; When not otherwise specified, a value of 0 means off (inactive) and 1
    means
    ; on (active).
    ;
    ; 0 opportunistic locking of files (default is 1)
    ; 1 read-ahead when the requester is performing sequential access
    ; 0 = do not use read-ahead
    ; 1 = use single read-ahead thread (default)
    ; 2 = use asynchronous read-ahead thread
    ; 2 write-behind (default is 1)
    ; 3 use of the chain-send NETBIOS NCB (default is 1)
    ; 4 check all incoming SMBs for correct format (default is 0)
    ; 5 support FCB opens (default is 1)
    ; 6 set priority of server (default is 4)
    ; 0-9 = 0 is highest priority, 9 is lowest priority
    ; 7 automatically allocate additional memory for directory searches
    ; (default is 1)
    ; 8 write records to the audit log only when the scavenger wakes up
    ; (default is 1)
    ; 9 do full buffering when a file is opened with deny-write sharing
    mode
    ; (default is 1)
    ; 10 set the interval for the scavenger to wake up
    ; 0 = 5 seconds 4 = 25 seconds 8 = 45
    seconds
    ; 1 = 10 seconds (default) 5 = 30 seconds 9 = 50
    seconds
    ; 2 = 15 seconds 6 = 35 seconds
    ; 3 = 20 seconds 7 = 40 seconds
    ; 11 allow compatibility-mode opens of certain types of files by
    translating
    ; them to sharing mode opens with deny-none
    ; 0 = Always use compatibility-mode opens.
    ; 1 = Use deny-none sharing mode if read-only access to .EXE or
    ..COM
    ; files.
    ; 2 = Use deny-none sharing mode if read-only access to .EXE or
    ..COM
    ; files. Use deny-write sharing mode if read-only access to
    ..BAT
    ; files is requested.
    ; 3 = Use deny-none sharing mode on all compatibility-mode opens.
    ; (default)
    ; 12 allow DOS LAN Services workstations to use a second NETBIOS
    session when
    ; sending printer requests (default is 1)
    ; 13 number of 64KB buffers used for read-ahead
    ; 0-9 = number of buffers (default is 1)
    ; 14 convert incoming path specifications into most basic format
    (default is 0)
    ; 15 Oplock Timeout and NETBIOS Acknowledgment Timeout (default is 0)
    ; 0 = 35 second Oplock Timeout; 34 second NETBIOS Acknowledgment
    Timeout
    ; 1 = 70 second Oplock Timeout; 69 second NETBIOS Acknowledgment
    Timeout
    ; 2 = 140 second Oplock Timeout; 127 second NETBIOS
    Acknowledgment Timeout
    ; 3 = 210 second Oplock Timeout; 127 second NETBIOS
    Acknowledgment Timeout
    ; 4 = 280 second Oplock Timeout; 127 second NETBIOS
    Acknowledgment Timeout
    ; 5 = 350 second Oplock Timeout; 127 second NETBIOS
    Acknowledgment Timeout
    ; 6 = 420 second Oplock Timeout; 127 second NETBIOS
    Acknowledgment Timeout
    ; 7 = 490 second Oplock Timeout; 127 second NETBIOS
    Acknowledgment Timeout
    ; 8 = 560 second Oplock Timeout; 127 second NETBIOS
    Acknowledgment Timeout
    ; 9 = 640 second Oplock Timeout; no NETBIOS Acknowledgment
    Timeout
    ; 16 validate IOCTLs (default is 1)
    ; 17 how long the server maintains unused, dynamic big buffers before
    freeing
    ; the memory.
    ; 0 = 0 seconds 4 = 5 minutes 8 = 1 hour
    ; 1 = 1 second 5 = 10 minutes 9 =
    indefinitely
    ; 2 = 10 seconds 6 = 20 minutes
    ; 3 = 1 minute (default) 7 = 40 minutes
    ; 18 how long the server waits after failing to allocate a big buffer
    before
    ; trying again.
    ; 0 = 0 seconds 3 = 1 minute (default)
    ; 1 = 1 second 4 = 5 minutes
    ; 2 = 10 seconds 5 = 10 minutes
    ; 19 RAW read and RAW write SMB protocols (default is 1)
    ; 20 server responds to announcement requests (default is 1)
    ;
    ; The next lines help you to locate bits in the srvheuristics entry.
    ; 1 2
    ; 012345678901234567890
    ; srvheuristics = 111101411113110913311
    ; srvheuristics = 011101411113130013311
    srvheuristics = 111101411113130913311
    SRVSERVICES = NETLOGON,LSSERVER
    ; SRVSERVICES = NETLOGON,LSSERVER,DCDBREPL,TIMESOURCE
    srvnets = NET1,NET2
    ; autopath = D:\IBMLAN\PROFILES\SRVAUTO.PRO
    ; autoprofile = load
    autopath =
    maxconnections = 300

    [alerter]

    sizalertbuf = 3072

    [browser]

    MaintainServerList = YES
    Quiet = NO

    [netrun]

    maxruns = 3
    runpath = D:\

    [lsserver]

    cleanup = no
    srvpipes = 5

    [services]

    ; Correlates name of service to pathname of service program.
    ; The pathname must be either
    ; 1) an absolute path (including the drive specification)
    ; OR
    ; 2) a path relative to the IBMLAN root
    alerter = services\alerter.exe
    dcdbrepl = services\dcdbrepl.exe
    genalert = services\genalert.exe
    lsserver = services\lsserver.exe
    messenger = services\msrvinit.exe
    netlogon = services\netlogon.exe
    netrun = services\runservr.exe
    replicator = services\replicat.exe
    requester = services\wksta.exe
    server = services\netsvini.exe
    timesource = services\timesrc.exe
    browser = services\browser.exe
    ups = services\ups.exe

    [ups]

    batterymsg = 600
    batterytime = 60
    cmdfile =
    cmdtimer = 30
    devicename = UPS_DEV
    lowbattery = warning
    messdelay = 5
    messtime = 120
    recharge = 100
    signals = 100
    voltlevels = 100


  5. Re: LAN Server diagnostics (was Re: NET3126: The network NET1 ranout of NCBs.)

    MMI wrote:
    > I'm having problem with a WSeB, after some time it stops working
    > (vanishes from the LAN, refuses both local or remote logons), however I
    > don't know how to diagnose what's wrong. Is there any way to diagnose
    > out-of-the-box WSeB? Unfortunately, it is not reproducible and it occurs
    > from time to time.
    >
    > Only the server stops working. NET STOP REQ - NET START SERVER resumes
    > the networking. I don't know what exactly might me the problem as the
    > TCP/IP keeps working - I can telnet into the stuck server and restart
    > the networking. There's one NIC with TCP/IP and plain NETBIOS.
    >
    > Jeroen, can you recommend something, please?
    >
    > Cheers,
    > Martin


    [C:\] type lantran.log
    This will show whether your network card is out of resources.

    [C:\] net error
    or the GUI version
    [C:\] lserror
    This will tell all errors the requester and the server encountered.

    In the latter: look for messages around the same time the server
    disappeared.

    Feel free to post some output; you can also contact me on the email
    address I use for posting.

    Best regards,
    Jeroen Besse

  6. Re: NET3126: The network NET1 ran out of NCBs.

    dumpfdumpf@yahoo.com wrote:

    Hi,

    Impressive protocol.ini... netware, wifi, virtual pc...!

    I'd suggest the following changes:

    PROTOCOL.INI:

    [tcpbeui_nif]

    SESSIONS = 70
    NCBS = 180
    NAMES = 22
    SELECTORS = 20


    (You might notice I made them the same as in [NetBEUI].)

    IBMLAN.INI:

    NET1 = NETBEUI$,0,LM10,64,180,20
    NET2 = TCPBEUI$,1,LM10,64,180,20

    The '64' is the number of sessions LAN Requester takes from those made
    available by PROTOCOL.INI.
    The '180' is the number of NCBs...
    The '20' is the number of names...

    In NETx = $,x,LM10,a,b,c
    Make sure that:
    a <= SESSIONS in PROTOCOL.INI
    b <= NCBS in PROTOCOL.INI
    c <= NAMES in PROTOCOL.INI

    When you issue a 'type lantran.log', you'll see how many sessions, ncbs
    and names are left over after LAN Requester took those defined in
    IBMLAN.INI away.

    BTW: I hope it was just an error in this message, pasting the content of
    IBMLAN.INI twice... if not: remove 50%! ;-)

    Hope this helps!

    Best regards,
    Jeroen Besse

  7. Re: NET3126: The network NET1 ran out of NCBs.

    dumpfdumpf@yahoo.com wrote on 24.01.05 14:57 :
    > My PROTOCOL.INI:


    [...]

    > [NETBEUI_nif]
    >
    > DriverName = netbeui$
    > Bindings = FXWRAP1_nif
    > ETHERAND_TYPE = "I"
    > USEADDRREV = "YES"
    > OS2TRACEMASK = 0x0
    > SESSIONS = 70
    > NCBS = 180


    Either increase this value to e.g. 188, or

    [...]

    > My IBMLAN.INI:
    >
    > ; IBM OS/2 Warp Server for e-business network initialization file
    >
    > [networks]
    >
    > NET1 = NETBEUI$,0,LM10,64,180,20


    decrease the 180 here to e.g. 172.

    [...]

    A change in Protocol.Ini needs a reboot, changing Ibmlan.Ini just
    requires a "net stop req /y" followed by a "net start srv".

    --
    Bye/2
    Meinolf

  8. Re: NET3126: The network NET1 ran out of NCBs.

    Thanks for *all* your suggestions folks, and yes there seems to have
    been a pasting error with IBMLAN.INI.
    I also have the Golden Code network tracer and the InJoy firewall
    running on my system.
    These tools are AMAZING and put any Windoze firewall and network tracer
    I've seen to shame.
    Jeez, my OS/2 box is now more powerfull than ever and if I could only
    get this part to work, too, then I'm "king of the network" again :-)

    I'll try all the suggestion in here and then I'll get back with you
    folks.
    Thanks to Bye/2 as well.

    PS: I recently purchased a high-resolution video in card for my laptop
    from a German outfit that comes with complete set of drivers and
    development tools for OS/2.
    Seems they're doing a lot of video scanning biz over there on OS/2
    systems for the manufacturing industry.
    I'm currently working on translating the PCMCIA drivers to CardBus
    standard and hope to post them to hobbes once I'm done.


  9. Re: NET3126: The network NET1 ran out of NCBs.

    Jeroen Besse wrote:
    > dumpfdumpf@yahoo.com wrote:
    >
    > Hi,
    >
    > Impressive protocol.ini... netware, wifi, virtual pc...!
    >
    > I'd suggest the following changes:
    >
    > PROTOCOL.INI:
    >
    > [tcpbeui_nif]
    >
    > SESSIONS = 70
    > NCBS = 180
    > NAMES = 22
    > SELECTORS = 20
    >
    >
    > (You might notice I made them the same as in [NetBEUI].)
    >
    > IBMLAN.INI:
    >
    > NET1 = NETBEUI$,0,LM10,64,180,20
    > NET2 = TCPBEUI$,1,LM10,64,180,20
    >
    > The '64' is the number of sessions LAN Requester takes from those made
    > available by PROTOCOL.INI.
    > The '180' is the number of NCBs...
    > The '20' is the number of names...
    >
    > In NETx = $,x,LM10,a,b,c
    > Make sure that:
    > a <= SESSIONS in PROTOCOL.INI
    > b <= NCBS in PROTOCOL.INI
    > c <= NAMES in PROTOCOL.INI
    >
    > When you issue a 'type lantran.log', you'll see how many sessions, ncbs
    > and names are left over after LAN Requester took those defined in
    > IBMLAN.INI away.
    >
    > BTW: I hope it was just an error in this message, pasting the content of
    > IBMLAN.INI twice... if not: remove 50%! ;-)
    >
    > Hope this helps!
    >
    > Best regards,
    > Jeroen Besse


    Oops, it was NET1 that was in error...

    Try values like these:
    protocol.ini, under [NetBEUI]:
    SESSIONS = 130
    NCBS = 225
    NAMES = 21
    SELECTORS = 50
    ibmlan.ini:
    net1 = NETBEUI$,1,LM10,120,210,14

    (increasing the NCBs made available by protocol.ini, then using more of
    them in ibmlan.ini)

  10. Re: LAN Server diagnostics (was Re: NET3126: The network NET1 ranout of NCBs.)

    Hi Jeroen,

    Jeroen Besse wrote:
    > MMI wrote:
    >
    >> I'm having problem with a WSeB, after some time it stops working
    >> (vanishes from the LAN, refuses both local or remote logons), however
    >> I don't know how to diagnose what's wrong. Is there any way to
    >> diagnose out-of-the-box WSeB? Unfortunately, it is not reproducible
    >> and it occurs from time to time.
    >>
    >> Only the server stops working. NET STOP REQ - NET START SERVER resumes
    >> the networking. I don't know what exactly might me the problem as the
    >> TCP/IP keeps working - I can telnet into the stuck server and restart
    >> the networking. There's one NIC with TCP/IP and plain NETBIOS.
    >>
    >> Jeroen, can you recommend something, please?
    >>
    >> Cheers,
    >> Martin

    >
    >
    > [C:\] type lantran.log
    > This will show whether your network card is out of resources.


    LANTRAN.LOG shows nothing except initial startup information, unfortunately.

    > [C:\] net error
    > or the GUI version
    > [C:\] lserror
    > This will tell all errors the requester and the server encountered.


    I cleared the logs, and now the NET ERROR returns

    [server]
    Program Message Time
    -----------------------------------------------------------------
    SERVER 5305 01/25/05 07:47pm
    NET5305: An NCB command timed out. The session may have ended abnormally.

    [client]
    Program Message Time
    -----------------------------------------------------------------
    REQUESTER 3194 01/25/05 07:47pm
    NET3194: A stalled session to WARPSERVER is being disconnected.
    REQUESTER 3195 01/25/05 07:47pm
    NET3195: An NCB error occurred: \\WARPSERVER\DISK160. The NCB is the
    data.
    REQUESTER 3195 01/25/05 07:47pm
    NET3195: An NCB error occurred: \\WARPSERVER\DISK160. The NCB is the
    data.

    I snipped the NCB hexa, but I do have them recorded.

    Cheers,
    Martin

    >
    > In the latter: look for messages around the same time the server
    > disappeared.
    >
    > Feel free to post some output; you can also contact me on the email
    > address I use for posting.
    >
    > Best regards,
    > Jeroen Besse


  11. Re: LAN Server diagnostics (was Re: NET3126: The network NET1 ranout of NCBs.)

    Hi Martin,

    Find and run NBJDSTAT (should be in the MPTS toolkit) and dump the
    output before the error, and after the error has occurred.

    >>> I'm having problem with a WSeB, after some time it stops working
    >>> (vanishes from the LAN, refuses both local or remote logons), however
    >>> I don't know how to diagnose what's wrong. Is there any way to
    >>> diagnose out-of-the-box WSeB? Unfortunately, it is not reproducible
    >>> and it occurs from time to time.
    >>>
    >>> Only the server stops working. NET STOP REQ - NET START SERVER
    >>> resumes the networking. I don't know what exactly might me the
    >>> problem as the TCP/IP keeps working - I can telnet into the stuck
    >>> server and restart the networking. There's one NIC with TCP/IP and
    >>> plain NETBIOS.

    NET3195: An NCB error occurred: \\WARPSERVER\DISK160. The NCB is the
    > data.


    Was is the data to the NET3195 error?

    What MPTS fixlevel are you at?
    What versions of server/HPFS386 are you using?
    >> In the latter: look for messages around the same time the server
    >> disappeared.


    Does the popuplog.os2 file show any processes abending? (e.g. SYS3175 in
    LRSD, NETSERVR, MUGLIB etc?)

    Also, please check your ibmlan.ini for a method of autodisconnection.
    Autodisconnection cleans up resources that are no longer required.

    Please make sure that scavange=yes and autodisconnect=xx, where xx is a
    value in minutes. If you have autodisconnect=-1 or 0, and do not have
    the scavange statement, then the server will eventually run out of
    resources...

    Regards,
    Stephen

  12. Re: LAN Server diagnostics (was Re: NET3126: The network NET1 ranout of NCBs.)

    Hi Steve,

    Steve Sharrad wrote:
    > Hi Martin,
    >
    > Find and run NBJDSTAT (should be in the MPTS toolkit) and dump the
    > output before the error, and after the error has occurred.


    I got the util, will do after the error occurs again, and post both dumps.

    >>>> I'm having problem with a WSeB, after some time it stops working
    >>>> (vanishes from the LAN, refuses both local or remote logons),
    >>>> however I don't know how to diagnose what's wrong. Is there any way
    >>>> to diagnose out-of-the-box WSeB? Unfortunately, it is not
    >>>> reproducible and it occurs from time to time.
    >>>>
    >>>> Only the server stops working. NET STOP REQ - NET START SERVER
    >>>> resumes the networking. I don't know what exactly might me the
    >>>> problem as the TCP/IP keeps working - I can telnet into the stuck
    >>>> server and restart the networking. There's one NIC with TCP/IP and
    >>>> plain NETBIOS.

    >
    > NET3195: An NCB error occurred: \\WARPSERVER\DISK160. The NCB is the
    >
    >> data.

    >
    >
    > Was is the data to the NET3195 error?


    First and second data:

    14 18 07 02 4C 45 5F 00 46 00 00 00 00 00 00 00 ....LE_.F.......
    00 00 00 00 00 00 00 00 00 00 4C 55 90 4E 00 00 ..........LU.N..
    00 00 00 00 00 00 00 00 00 00 00 00 94 DD 00 00 ................
    00 18 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................


    -------------------------------------------------------

    14 18 07 02 40 34 5F 00 46 00 00 00 00 00 00 00 ....@4_.F.......
    00 00 00 00 00 00 00 00 00 00 40 44 90 4E 00 00 ..........@D.N..
    00 00 00 00 00 00 00 00 00 00 00 00 94 DD 00 00 ................
    00 18 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................

    Can't read this :-)

    > What MPTS fixlevel are you at?
    > What versions of server/HPFS386 are you using?


    It's ACP2 with no addtional FP for any component, and I don't have HPFS386.

    >>> In the latter: look for messages around the same time the server
    >>> disappeared.

    >
    >
    > Does the popuplog.os2 file show any processes abending? (e.g. SYS3175 in
    > LRSD, NETSERVR, MUGLIB etc?)


    Nothing. I do have SUPPRESSPOPUPS=[bootdrive], but nothig is recorded.

    > Also, please check your ibmlan.ini for a method of autodisconnection.
    > Autodisconnection cleans up resources that are no longer required.


    Problem is this happens while traffic occurs.

    > Please make sure that scavange=yes and autodisconnect=xx, where xx is a
    > value in minutes. If you have autodisconnect=-1 or 0, and do not have
    > the scavange statement, then the server will eventually run out of
    > resources...


    I don't have scavange, but I do have autodisconnect at 120. But as I
    said, it happens while traffic occurs. I'd suspect NIC driver, but
    problem is that TCP/IP and simply the rest of os and networking works
    uninterrupted. Strange.

    Cheers,
    Martin

    >
    > Regards,
    > Stephen


  13. Re: LAN Server diagnostics (was Re: NET3126: The network NET1 ranout of NCBs.)

    MMI wrote:
    > Hi Steve,
    >
    > Steve Sharrad wrote:
    >
    >> Hi Martin,
    >>
    >> Find and run NBJDSTAT (should be in the MPTS toolkit) and dump the
    >> output before the error, and after the error has occurred.

    >
    >
    > I got the util, will do after the error occurs again, and post both dumps.
    >
    >>>>> I'm having problem with a WSeB, after some time it stops working
    >>>>> (vanishes from the LAN, refuses both local or remote logons),
    >>>>> however I don't know how to diagnose what's wrong. Is there any way
    >>>>> to diagnose out-of-the-box WSeB? Unfortunately, it is not
    >>>>> reproducible and it occurs from time to time.
    >>>>>
    >>>>> Only the server stops working. NET STOP REQ - NET START SERVER
    >>>>> resumes the networking. I don't know what exactly might me the
    >>>>> problem as the TCP/IP keeps working - I can telnet into the stuck
    >>>>> server and restart the networking. There's one NIC with TCP/IP and
    >>>>> plain NETBIOS.

    >>
    >>
    >> NET3195: An NCB error occurred: \\WARPSERVER\DISK160. The NCB is the
    >>
    >>> data.

    >>
    >>
    >>
    >> Was is the data to the NET3195 error?

    >
    >
    > First and second data:
    >
    > 14 18 07 02 4C 45 5F 00 46 00 00 00 00 00 00 00 ....LE_.F.......
    > 00 00 00 00 00 00 00 00 00 00 4C 55 90 4E 00 00 ..........LU.N..
    > 00 00 00 00 00 00 00 00 00 00 00 00 94 DD 00 00 ................
    > 00 18 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
    >
    >
    > -------------------------------------------------------
    >
    > 14 18 07 02 40 34 5F 00 46 00 00 00 00 00 00 00 ....@4_.F.......
    > 00 00 00 00 00 00 00 00 00 00 40 44 90 4E 00 00 ..........@D.N..
    > 00 00 00 00 00 00 00 00 00 00 00 00 94 DD 00 00 ................
    > 00 18 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
    >
    > Can't read this :-)
    >
    >> What MPTS fixlevel are you at?
    >> What versions of server/HPFS386 are you using?

    >
    >
    > It's ACP2 with no addtional FP for any component, and I don't have HPFS386.


    Does ACP(2) have updated MPTS & LAN Requester/Server? (I really don't know)

    You can see by issueing:
    [C:\] syslevel ibmcom
    and:
    [C:\] syslevel ibmlan

    My eCS 1.2 gives for MPTS:
    ibmcom\SYSLEVEL.TRP
    IBM OS/2 LAN Adapter and Protocol Support
    Version 6.01 Component ID 5639A5700
    Current CSD level: WR08708
    Prior CSD level: WR08701

    (AFAIK that's the latest & greatest)

  14. Re: LAN Server diagnostics (was Re: NET3126: The network NET1 ranout of NCBs.)

    Jeroen Besse wrote:
    > MMI wrote:
    >
    >> Hi Steve,
    >>
    >> Steve Sharrad wrote:
    >>
    >>> Hi Martin,
    >>>
    >>> Find and run NBJDSTAT (should be in the MPTS toolkit) and dump the
    >>> output before the error, and after the error has occurred.

    >>
    >>
    >>
    >> I got the util, will do after the error occurs again, and post both
    >> dumps.
    >>
    >>>>>> I'm having problem with a WSeB, after some time it stops working
    >>>>>> (vanishes from the LAN, refuses both local or remote logons),
    >>>>>> however I don't know how to diagnose what's wrong. Is there any
    >>>>>> way to diagnose out-of-the-box WSeB? Unfortunately, it is not
    >>>>>> reproducible and it occurs from time to time.
    >>>>>>
    >>>>>> Only the server stops working. NET STOP REQ - NET START SERVER
    >>>>>> resumes the networking. I don't know what exactly might me the
    >>>>>> problem as the TCP/IP keeps working - I can telnet into the stuck
    >>>>>> server and restart the networking. There's one NIC with TCP/IP and
    >>>>>> plain NETBIOS.
    >>>
    >>>
    >>>
    >>> NET3195: An NCB error occurred: \\WARPSERVER\DISK160. The NCB is the
    >>>
    >>>> data.
    >>>
    >>>
    >>>
    >>>
    >>> Was is the data to the NET3195 error?

    >>
    >>
    >>
    >> First and second data:
    >>
    >> 14 18 07 02 4C 45 5F 00 46 00 00 00 00 00 00 00
    >> ....LE_.F.......
    >> 00 00 00 00 00 00 00 00 00 00 4C 55 90 4E 00 00
    >> ..........LU.N..
    >> 00 00 00 00 00 00 00 00 00 00 00 00 94 DD 00 00
    >> ................
    >> 00 18 00 00 00 00 00 00 00 00 00 00 00 00 00 00
    >> ................
    >>
    >>
    >> -------------------------------------------------------
    >>
    >> 14 18 07 02 40 34 5F 00 46 00 00 00 00 00 00 00
    >> ....@4_.F.......
    >> 00 00 00 00 00 00 00 00 00 00 40 44 90 4E 00 00
    >> ..........@D.N..
    >> 00 00 00 00 00 00 00 00 00 00 00 00 94 DD 00 00
    >> ................
    >> 00 18 00 00 00 00 00 00 00 00 00 00 00 00 00 00
    >> ................
    >>
    >> Can't read this :-)
    >>
    >>> What MPTS fixlevel are you at?
    >>> What versions of server/HPFS386 are you using?

    >>
    >>
    >>
    >> It's ACP2 with no addtional FP for any component, and I don't have
    >> HPFS386.

    >
    >
    > Does ACP(2) have updated MPTS & LAN Requester/Server? (I really don't know)
    >
    > You can see by issueing:
    > [C:\] syslevel ibmcom
    > and:
    > [C:\] syslevel ibmlan
    >
    > My eCS 1.2 gives for MPTS:
    > ibmcom\SYSLEVEL.TRP
    > IBM OS/2 LAN Adapter and Protocol Support
    > Version 6.01 Component ID 5639A5700
    > Current CSD level: WR08708
    > Prior CSD level: WR08701
    >
    > (AFAIK that's the latest & greatest)


    Hi Jeroen, what I have is fairly new...

    E:\IBMCOM\SYSLEVEL.TRP
    IBM OS/2 LAN Adapter and Protocol Support
    Version 6.01 Component ID 5639A5700
    Current CSD level: WR08705
    Prior CSD level: WR08701

    Other syslevels:

    E:\IBMLAN\SYSLEVEL.REQ
    IBM OS/2 LAN Requester
    Version 5.20 Component ID 562294000
    Current CSD level: IP08605
    Prior CSD level: IP08600

    E:\IBMLAN\SYSLEVEL.SRV
    IBM OS/2 LAN Server
    Version 5.20 Component ID 562288900
    Current CSD level: IP08605
    Prior CSD level: IP08600



    However I expect some error exactly within the SERVER service, because
    after it hangs it refuses to be stopped. NET STOP SERVER doesn't work
    saying the service can't be stopped now.

    I don't expect much from replacing the MPTS alone. Or can it cure the hangs?

    Cheers,
    Martin

  15. Re: NET3126: The network NET1 ran out of NCBs.

    I tried your values and this is what I get now:
    0:0[D:\]net status
    The REQUESTER service is not started.
    The REQUESTER service is starting.......
    The REQUESTER service could not be started.
    An error occurred. Refer to the help for the following message:
    NET3065: The required NetBIOS group name cannot be added to the
    adapter.


  16. Re: NET3126: The network NET1 ran out of NCBs.

    I tried your suggestions and no longer net "out of NCBs" errors, but
    now it terminates like this:

    >0:0[D:\]net status
    >The SERVER service is not started.
    >The SERVER service is starting..
    >The SERVER service could not be started.
    >NET3057: An internal consistency error has occurred.
    >
    >For more information, type HELP NET3057.


    Even the LAN error log utility does not reveal any more info.
    I'm not sure if I'm now better or worse of, but nevertheless, got any
    more ideas?

    DD


  17. Re: NET3126: The network NET1 ran out of NCBs.

    dumpfdumpf@yahoo.com wrote:
    > I tried your suggestions and no longer net "out of NCBs" errors, but
    > now it terminates like this:
    >
    >
    >>0:0[D:\]net status
    >>The SERVER service is not started.
    >>The SERVER service is starting..
    >>The SERVER service could not be started.
    >>NET3057: An internal consistency error has occurred.
    >>
    >>For more information, type HELP NET3057.

    >
    >
    > Even the LAN error log utility does not reveal any more info.
    > I'm not sure if I'm now better or worse of, but nevertheless, got any
    > more ideas?


    That's weird. Now the server won't start at all. (IMHO that's worse than
    before; I'd suggest to revert to the vaules you had for now.)

    Does lantran.log now show some error?

    I'll install WSeB on a free partition, and try some values. I'll come
    back to you on this.

  18. Re: NET3126: The network NET1 ran out of NCBs.

    Well, there are *no* errors in LANTRAN.LOG and at least I'm ot getting
    those damn NCB errors no more.
    The server service didn't come up before either.
    The only bother is that now I'm getting no meaningful error message at
    all.
    Which really sucks ...
    Forget about typing >HELP NET3057<. That just gets me a "topic not
    found" error message :-(

    By the way, where can I get that MPTN applet mentioned somewhere in
    here called "nbjdstat" ?


  19. Re: NET3126: The network NET1 ran out of NCBs.

    It's part of some MPTS update/fixpak and is NOT installed automatically.
    However, you can just unzip the zip file (which I think is called applets.zip).

    Lars

    dumpfdumpf@yahoo.com schrieb:
    > Well, there are *no* errors in LANTRAN.LOG and at least I'm ot getting
    > those damn NCB errors no more.
    > The server service didn't come up before either.
    > The only bother is that now I'm getting no meaningful error message at
    > all.
    > Which really sucks ...
    > Forget about typing >HELP NET3057<. That just gets me a "topic not
    > found" error message :-(
    >
    > By the way, where can I get that MPTN applet mentioned somewhere in
    > here called "nbjdstat" ?
    >


  20. Re: LAN Server diagnostics (was Re: NET3126: The network NET1 ranout of NCBs.)

    Steve Sharrad wrote:
    > Hi Martin,
    >
    > Find and run NBJDSTAT (should be in the MPTS toolkit) and dump the
    > output before the error, and after the error has occurred.


    This was a good idea, however, after the server hung, it also vanished
    from the LAN, so NBJDSTAT couldn't check it, because the name wasn't in
    the network already. Any other idea?

    As usually, I tried to restart the network system and this is what I got:


    These services are active:

    MESSENGER SERVER NETLOGON
    LSSERVER DCDBREPL NETRUN
    Stopping the requester service also stops these services.
    These requesters have sessions on this server:
    WARPBOX
    These requesters have sessions with open files on this server:
    WARPBOX
    The SERVER service is stopping............
    The SERVER service was successfully stopped.
    Stopping the requester service logs you off.
    The REQUESTER service is stopping.
    The REQUESTER service was successfully stopped.
    NET2186: The service is not responding to the control function.

    For more information, type HELP NET2186.


    Of course, help message was unusable :-)

    Cheers,
    Martin



    > Regards,
    > Stephen


+ Reply to Thread
Page 1 of 2 1 2 LastLast