Conference virke::mrmemo

150.0. "Server craches" by PEARS::SONNTAG (Michael Sonntag, TSSC-OIS, @MUH) Fri Oct 09 1992 18:10

Hello

trouble on customer side, we start the MRMSRV 1, he fetches the message,
becomes status assembling and crashes. 
Here now the LOG-File (with MASK=FF)

Thanks
	Michael 

!
$! This command procedure is always run when anybody on the entire system
$! logs in. It is equivalent to LOGIN.COM except that the instructions
$! contained herein are executed everytime anyone on the VMS system
$! logs in to their account.
$!
$! For interactive processes, turn on Control T, and set the terminal type
$ set noveri
$ define tt nl:
$ run/nodebug sys$system:mrmsrv
Time: 1992-10-09 17:04:59.23; message from server MRMEMO1:
%MRMEMO-I-DECNETMSG, DECnet notification, message number is: 50

Time: 1992-10-09 17:04:59.38; message from server MRMEMO1:
%MRMEMO-I-FSM, starting, connecting to MR and MEMO Gateway after clock tick

Time: 1992-10-09 17:04:59.54; message from server MRMEMO1:
%MRMEMO-I-FSM, starting, connecting to MR and MEMO Gateway after start request f
rom MRMMAN

Time: 1992-10-09 17:04:59.90; message from server MRMEMO1:
%MRMEMO-I-NEWACCOUNT, opening new accounting file: MR$DISK:[MRMEMO]MRMEMOACC1.DA
T;380
Time: 1992-10-09 17:05:04.59; message from server MRMEMO1:
%MRMEMO-I-CONNECTION, connection established to Message Router:
-SYSTEM-S-SUCCESS, Successful completion

 9-OCT-1992 17:05:04.92: SNA event CONNECT, substate: 0050, info: 00
 9-OCT-1992 17:05:04.94: SNA event BINDOK, substate: 0050, info: 00
Time: 1992-10-09 17:05:04.94; message from server MRMEMO1:
%MRMEMO-I-CONNECTION, connection established to MEMO:
-SNALU0-S-NORMAL, normal successful completion

 9-OCT-1992 17:05:04.98: SNA event RCVMSG, substate: 0070, info: 00
 9-OCT-1992 17:05:05.09: SNA event BIDTST, substate: 0070, info: A0
Time: 1992-10-09 17:05:05.09; message from server MRMEMO1:
%MRMEMO-I-FSM, synchronising MEMO Gateway link after SDT request from MEMO

 9-OCT-1992 , substate: 0070, info: A0
 9-OCT-1992 17:05:05.23: SNA event RCVMSGx

 9-OCT-1992 17:05:06.30: SNA event BIDTST, substate: 047
 9-OCT-1992 17:05:06.30: SNA event BIDTST, substate: 047
..............
%MRMEMO-I-FSM, idle, connected to MR and MEMO after message indicated in MR mail
box

 9-OCT-1992 17:05:06.30: SNA event BIDTST, substate: 0470, info: C8
 9-OCT-1992 17:05:06.31: SNA event NEGRSP, substate: 1470, info: C8
Time: 1992-10-09 17:05:09.40; message from server MRMEMO1:
%MRMEMO-I-DECNETMSG, DECnet notification, message number is: 50

%ADA-F-PROGRAM_ERROR, PROGRAM_ERROR
-ADA-F-RECNOTPOS, Program is erroneous, error recovery by exception handling is
not possible
-ADA-F-IMAFOREXI, The image was forced to exit
-ADA-F-EXCDUREXC, An exception occurred in the VAX Ada run-time library while ha
ndling an exception
 9-OCT-1992 17:05:33.72: SNA event DISCON, substate: 1460, info: 00
%MRMEMO-I-CLOSEACC, closing accounting file: MR$DISK:[MRMEMO]MRMEMOACC1.DAT;380
%MRMEMO-I-BYE, VAX MAILGATE for MEMO Server Exit
%ADA-F-PROGRAM_ERROR, PROGRAM_ERROR
$ logout
  MRMEMO       job terminated at  9-OCT-1992 17:05:34.74

    
    Hi Michael,
    
    This has to do with DDS for sure. Are there any inconsistencies in 
    MBMAN> SHOW DDS NODE *
              
    cheers,
    
    	Stefan

    
    Another thing:
    
    - Is your node a DDS WS node (it should be)?
    
    - If so is it working properly?
    
    - What do the DDS logs say?

    
    Well, after some off-line work this was resolved like this:
    
    The problem had to do with a message with the syntax:
    
    <prefix>.""userid""..routes 
    
    The MRMEMO server address parser transformed the userid into an empty
    string. This resulted in a DDS search for ALL subscribers with 
    Userid = "", which meant that 4.000 records were found. This in turn
    resulted in heavy resource problems and in the end MRMEMO just died.
    Because of this the DDS output queue was never correctly purged and DDS
    started to run wild too. Great fun...
    
    We tried to release the message from the MEMO queue by tenmporarily
    disabling address translation and restart the server, but that wouldn't
    help since the Userid was empty, and consequently MRMEMO crashed when
    posting the message to Message Router with:
    
    MRIF-F-MISSUSERID
    
    A fix that will take care of this particular address syntax is available in
    5.18. 
    
    However please be aware that similar problems may occur if you use the
    MRMEMO addressing syntax too liberally.
    
    
    	Stefan