Conference virke::mrmemo

81.0. "FATAL INTERNAL ERRORS - HELP!!" by HSOMAI::WATKINS () Thu Apr 25 1991 19:09

    
    My customer installed MRMEMO a couple weeks ago and everything seemed
    okay, we could send messages back and forth with no problems. 
    Recently,  they have been having problems with the server getting fatal
    internal errors and doing a stack dump.  The events that occured prior
    to these errors is that both the VAX and IBM systems had problems and
    were rebooted.  Also there were problems with the VTAM application but
    they said that all the parameters for MEMO were okay and they could get
    MRMEMO going for a while.  This is a strange problem in that it only
    occurs sporadically, and when it happens, MRMEMO stays down for a long
    time.  When I visited the customer site yesterday, everything was
    working fine.  He could send and receive messages on both sides.  The
    problem occured again when he demoed for his managers and tried to send
    a message from MEMO to All-in-1.  The status in the server was
    'receiving from MEMO' and then a fatal internal error occurs.  The
    error codes and stack dump follows.  The site configuration is:
    
    
    E-net     ---------------------------------------------
                  |                           |
              ------------              -------------
    	      |   VAX	 |              | Interlink |
    	      |  8650    |              -------------
              ------------                    | <----- Channel connected
    					-------------
    				        |   IBM     |
    				        |  3090     |
    					-------------
      
    MRMEMO1.LOG (Excerpts of events)
    ============
    
    Time: 1991-04-11 17:25:28.94; message from server MRMEMO1:
    %MRMEMO-I-NEWACCOUNT, opening new accounting file:
    DISK$VAXEMAIL:[MRMEMO]MRMEMOACC1.DAT;14
    
    %SNA-E-LOGUNIDEA, SSCP has deactivated the session
    Time: 1991-04-12 02:32:15.81; message from server MRMEMO1:
    %MRMEMO-F-UNBIND, UNBIND request received from MEMO Gateway
    .
    .
    .
    %SNA-E-LOGUNIDEA, SSCP has deactivated the session
    Time: 1991-04-13 19:15:52.63; message from server MRMEMO1:
    %MRMEMO-F-UNBIND, UNBIND request received from MEMO Gateway
    
    %SNA-E-ACCINTERR, Gateway detected an error in the Gateway Access
    Routines
    .
    .
    .
    %SNALU0-E-NOSESN, failed to start session
    -SNA-E-FAIESTLIN, failed to establish DECnet link to the Gateway
    -SYSTEM-F-UNREACHABLE, remote node is not currently reachable
    %SNALU0-E-NETSHUT, network node is not accepting connects
    Time: 1991-04-14 02:33:23.67; message from server MRMEMO1:
    %MRMEMO-I-GWYCOMERR, error on DECnet link to DECnet/SNA Gateway
    .
    .  (IBM system went down but came back up)
    .
    %SNALU0-E-RCVFAIL, call to receive failed
    -SNA-E-FUNCABORT, access routine function aborted
    Time: 1991-04-14 02:33:23.77; message from server MRMEMO1:
    %SNALU0-E-RCVFAIL, call to receive failed
    
    %MRMRMO-W-TRACE, traceback forced from the Server Handler
    -MRMEMO-I-FSM, idle, connected to MR and MEMO after clock tick
    -MRMEMO-I-RING, ring: '1A 1A 1A 1A 1A 1A 1A 1A 1A 1A.', substates:
    00000030
    %TRACE-W-TRACEBACK, symbolic stack dump follows
     module name      routine name         Line     rel PC    abs PC
    
      SRVLOG            SRV$LOG_HANDLER    3943     000001C9  001E379D
    ---above condition handler called with exception 040080BA:
    %SNALU0-E-RCVFAIL, call to receive failed
    ---end of exception message
    .
    .
    .
    %SNALU0-E-NOSESN, failed to start session
    -SNA-E-FAIESTLIN, failed to establish DECnet link to the Gateway
    -SYSTEM-F-UNREACHABLE, remote node is not currently reachable
    %SNALU0-E-NOSESN, failed to start session
    -SNA-E-FAIESTLIN, failed to establish DECnet link to the Gateway
    -SYSTEM-F-UNREACHABLE, remote node is not currently reachable
    %SNA-E-LOGUNIDEA, SSCP has deactivated the session
    Time: 1991-04-14 02:35:49.92; message from server MRMEMO1:
    %MRMEMO-F-UNBIND, UNBIND request received from MEMO Gateway
    .
    . (The UNBIND messages occur every day around 2:30AM and retries for
    .      approximately 10 minutes)
    .
    . (Another RCVFAIL error and STACK dump occurs, current node
    .  unreachable, and SSCP deactivates session messages appear again)
    .
    .  (IBM and VTAM application brought back up)
    .
    Time: 1991-04-18 14:15:35.71; message from server MRMEMO1:
    %SNA-F-FATINTERR, internal error in Gateway access routines
    -SNA-F-UNKMSGREC, unknown message code received from Gateway
    
    %MRMRMO-W-TRACE, traceback forced from the Server Handler
    -MRMEMO-I-FSM, receiving from MEMO after clock tick
    -MRMEMO-I-RING, ring: '18 2A 27 1B 13 36 1A 13 39 3A.', substates:
    00000030
    %TRACE-W-TRACEBACK, symbolic stack dump follows
     module name      routine name         Line     rel PC    abs PC
    
      SRVLOG            SRV$LOG_HANDLER    3943     000001C9  001E379D
    ---above condition handler called with exception 0203827C:
    %SNA-F-FATINTERR, internal error in Gateway access routines
    -SNA-F-UNKMSGREC, unknown message code received from Gateway
    ---end of exception message
    .
    .  (Stack dump here)
    .
    Time: 1991-04-22 15:48:41.21; message from server MRMEMO1:
    %MRMEMO-I-SHUTTING, server shutting down
    
    Time: 1991-04-22 15:53:41.07; message from server MRMEMO1:
    %MRMEMO-I-SHUTTING, server shutting down
    .
    . (Customer shut MRMEMO down and brought back up after systems were
    .   stable)
    .
    Time: 1991-04-23 13:38:57.18; message from server MRMEMO1:
    %MRMEMO-I-NEWACCOUNT, opening new accounting file:
    DISK$VAXEMAIL:[MRMEMO]MRMEMOACC1.DAT;17
    
    Time: 1991-04-23 13:39:27.05; message from server MRMEMO1:
    %SNA-F-FATINTERR, internal error in Gateway access routines
    -SNA-F-UNKMSGREC, unknown message code received from Gateway
    
    %MRMRMO-W-TRACE, traceback forced from the Server Handler
    -MRMEMO-I-FSM, receiving from MEMO after clock tick
    -MRMEMO-I-RING, ring: '          4A 4D 52 13 39 3A.', substates:
    00000070
    %TRACE-W-TRACEBACK, symbolic stack dump follows
     module name      routine name         Line     rel PC    abs PC
    
      SRVLOG            SRV$LOG_HANDLER    3943     000001C9  001E379D
    ---above condition handler called with exception 0203827C:
    %SNA-F-FATINTERR, internal error in Gateway access routines
    -SNA-F-UNKMSGREC, unknown message code received from Gateway
    ---end of exception message
    .
    .  (Stack dump here)
    .
    . (SHUT DOWN MRMEMO AGAIN)
    .
    Time: 1991-04-23 13:52:21.72; message from server MRMEMO1:
    %MRMEMO-I-NEWACCOUNT, opening new accounting file:
    DISK$VAXEMAIL:[MRMEMO]MRMEMOACC1.DAT;18
    
    Time: 1991-04-23 14:01:48.63; message from server MRMEMO1:
    %SNA-F-FATINTERR, internal error in Gateway access routines
    -SNA-F-UNKMSGREC, unknown message code received from Gateway
    
    %MRMRMO-W-TRACE, traceback forced from the Server Handler
    -MRMEMO-I-FSM, receiving from MEMO after BID request from MEMO
    -MRMEMO-I-RING, ring: '1A 19 18 2A 27 1B 13 36 1A 33.', substates:
    00000070
    %TRACE-W-TRACEBACK, symbolic stack dump follows
     module name      routine name         Line     rel PC    abs PC
    
      SRVLOG            SRV$LOG_HANDLER    3943     000001C9  001E379D
    ---above condition handler called with exception 0203827C:
    %SNA-F-FATINTERR, internal error in Gateway access routines
    -SNA-F-UNKMSGREC, unknown message code received from Gateway
    ---end of exception message
    .
    .  (Stack dump here)
    .
    .  (Server shuts down)
    
    
    
    When I arrived at customer site on 4/24/91 everything was okay until
    after I left.  The customer thinks the problems occur when he tries to
    send a message from the MEMO side.  I have checked all my resources and
    have come to the conclusion that these things are happening due to
    problems within API.  The Gateway is fine because other applications
    are using it with no problems.  My customer will run traces from the
    IBM side but doesn't think that will tell him much.
    
    Has anyone experienced these weird occurences?  If so, what action did
    you take to resolve it?  If not, what can I tell my customer?  Is it
    a MRMEMO problem or not????
    
    Apologize for the LENGTHY note, just wanted all details here for HELP!!
    
    Tracey
    
    
    
    
    

    Hello Tracey,

    MRMEMO uses the SNA-API LU0 communication with Memo/Gateway and
    as far as I know, we don't support use of the API with anything but
    our own SNA/Gateways.
    The messages

	%SNA-E-ACCINTERR, Gateway detected an error in the Gateway Access
	Routines
    and
	%SNA-F-FATINTERR, internal error in Gateway access routines
	-SNA-F-UNKMSGREC, unknown message code received from Gateway

    certainly indicates that something is wrong in the lower layers of
    the protocol (i.e. the parts that are handled by the SNA API). I don't
    think MRMEMO is doing anything when this happens. MRMEMO has probably
    just called the API to start a session and for some reason the API
    failes to do that. So I think you are right in your conclusion that
    the problems occur whithin the API. (Which is not the same as to say
    there is something wrong with the API.)

    BTW, what versions are you using (MRMEMO, VMS, SNA-API) ?

    What "other" applications work whithout problems? Any other application
    that uses the API?

    We might be able to get some indication on what kind of message the
    API has trouble with if you could produce an SNA trace from an occasion
    when the problem occurs.

    I'm afraid I can't be of much help (and if this is an Interlink problem
    I'm not sure we should help).

    Anders    

    
    Re: -1
    
    In answer to your question, the versions of software are as follows:
    
    MRMEMO V2.1
    VMS    V5.3-1
    SNA-API V2.3
    
    There are no other API applications on the system, and since they are
    using an Interlink connection to the IBM system, there is no SNA
    management software to do a SNA trace.  Since this problem occurs
    intermitently (sp?), when a part of the 3090 goes down for maintenance
    or when a message is initiated from the MEMO side, I can only assume
    that something strange is happening on the IBM side.  My customer
    thinks this also, and will see what he can do to isolate the problem.
    
    I remember reading a previous note of another customer using the
    Interlink connection to implement MRMEMO.  Has that site been running
    successfully?
    
    Tracey