[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference help::decnet-osi_for_vms

Title:DECnet/OSI for OpenVMS
Moderator:TUXEDO::FONSECA
Created:Thu Feb 21 1991
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:3990
Total number of notes:19027

3859.0. "HELP! Network Server Object" by REFDV1::LANZILOTTI () Tue Feb 04 1997 10:15

Hello,

We have inherited a client server application which uses the NCP database 
to create its network object for the server.  Due to a lack of network 
expertise, especially with phase 5, undocumented code, and in general an 
unfamiliarity with this, we are unable to get the server working under the SEPS
environment....please help or advise!   I do not know if I am missing a step or
code, if the code just plain won't work, if "detached network processes" just 
won't work under the new decnet?  In the current environment, the client node 
just comes in and pings the command stream, creating the process id.  The code 
is written in DCL, COBOL, C and SPEEDWAY(CRINGE!).  
                                                 
Here are the particulars: 

The client modifies the following file when it is built on the client end:  
	
	SYS$LIBRARY:IMAGELIB.OLB 

The kit builds the following files on the client end:

	SYS$SHARE:APPIX$PPCS_CLIENT_SHR.EXE
	SYS$HELP:APPIX$PPCS_REFERENCE_MANUAL.PS
	SYS$TEST:APPIX$PPCS_IVP.COM 	

Although we are not PASE compliant, this server was tested under PASE and did 
work.  However, we are now upgrading to SEPS, and so far we have been 
unsuccessful at getting the server to work under the new version of DECnet.  We 
HAVE been successful at getting the client to work.  

The server was installed onto the SEPS test system as follows:

From old NCP definitions:  

$!
$ MCR NCP DEFINE OBJECT PPCS_PARTSRV NUMBER 174
$ MCR NCP DEFINE OBJECT PPCS_PARTSRV FILE -
      APPIX$SECURE_A:[PPCS.PART_SERVER]APPIX$PPCS_PART_SERVER.COM 
$ MCR NCP DEFINE OBJECT PPCS_PARTSRV USER APPIX_SERVER PASSWORD SUPERBOWL_BOUND
$ MCR NCP DEFINE OBJECT PPCS_PARTSRV ALL 
$!

To the new NCL code:

CREATE NODE 0 SESSION CONTROL APPLICATION PPCS_PARTSRV
SET NODE 0 SESSION CONTROL APPLICATION PPCS_PARTSRV -
ADDRESSES = {NUMBER=174 ,NAME=PPCS_PARTSRV} -
,CLIENT =  -
,INCOMING ALIAS = TRUE -
,INCOMING PROXY = TRUE -
,OUTGOING ALIAS = TRUE -
,OUTGOING PROXY = TRUE -
,NODE SYNONYM = TRUE -
,IMAGE NAME = APPIX$SECURE_A:[PPCS.PART_SERVER]APPIX$PPCS_PART_SERVER.COM - 
,USER NAME = "PPCS_PARTSRV" -
,INCOMING OSI TSEL =

The following command to INSTALL into memory: 
(Once this was executed on the SEPS test system, the CLIENT on the SEPS test 
system worked, BUT this INITIALLY came from the SERVER end, our PRODUCTION 
cluster...)

$       INSTALL
        REPLACE APPIX$SECURE_A:[PPCS.PART_SERVER]APPIX$PPCS_PART_SERVER.EXE -
        /PRIV=SYSNAM/HEADER/SHARED
        EXIT



Many Thanks/Regards,
Janet-Lee

T.RTitleUserPersonal
Name
DateLines
3859.1RMULAC.DVO.DEC.COM::S_WATTUMScott Wattum - FTAM/VT/OSAK EngineeringTue Feb 04 1997 11:1111
detached network processes work fine under DECnet/OSI.
More information is required.

Are you getting any network events on the server system when the client attempts
to connect?

Does the server get started and then fail?  Are there any .LOG files created.

If the server is failing, does it indicate why it failed?

--Scott
3859.2no logs...REFDV1::LANZILOTTIWed Feb 05 1997 11:5512
    Hello,
    
    no, we can't find any netwerver.log files, we just get a message from
    speedway saying the speedway$server object was not available
    (client end).  Thats why I must be defining the server incorrectly. The
    NET$SERVER.log would be located in the APPIX_SERVER accounts login
    directory on the client end, correct?   
    
    We really are unfamiliar with this, please bear with.
    
    Regards,
    Janet-Lee
3859.3RMULAC.DVO.DEC.COM::S_WATTUMScott Wattum - FTAM/VT/OSAK EngineeringWed Feb 05 1997 13:186
Well, you'd only see a .LOG file if you got far enough to login.

do a REPLY/ENABLE=(NETWORK,SECURITY)

and see if you are getting any network or security related messages about the
connection attempt.
3859.4how come....ipng::CARSONPete Carson, Networks for OpenVMS EngineeringThu Feb 06 1997 09:0715
	In the PhaseIV implementation you used
		OBJECT PPCS_PARTSRV USER APPIX_SERVER
	and in PhaseV (aka DECnet/OSI alias DECnet-Plus) you used
		USER NAME = "PPCS_PARTSRV" -

	????

	Did you mean to change this to use a different username?
	Note that in DECnet-Plus the password is no longer needed.
	Hence he who types NCL SET APPLIC X USERNAME has privs and
	is saying that this application may be run without knowing
	a password.

	-Pete
3859.5exitREFDV1::LANZILOTTITue Feb 11 1997 11:0274
Well, so far in the saga of the server, we have changed the server username, 
which was incorrect.  However, we are still getting the same error:

    %SPWY -F-NOSERVER, connection to remote server not available 

I did notice today however that the test server account is missing the full 
batch access in the sysuaf, unlike the real server account.  

Below is the results of the last try.  I enabled opcom and as you can see we 
have login failure and priv problems.  

Please let me know if you have any ideas. 

Manythanks/regards,
Janet-Lee


KWLITY_DRMMGR> 
%%%%%%%%%%%  OPCOM  11-FEB-1997 10:19:39.46  %%%%%%%%%%%    (from node KWDISK at
 11-FEB-1997 10:19:38.69)
Message from user AUDIT$SERVER on KWDISK
Security alarm (SECURITY) and security audit (SECURITY) on KWDISK, system id: 11
801
Auditable event:          Detached process login failure
Event time:               11-FEB-1997 10:19:38.64
PID:                      20404E90        
Process name:             _WSA1:          
Username:                 SYSTEM          
Process owner:            [SYSTEM]
Image name:               DSA0:[SYS16.SYSCOMMON.][SYSEXE]LOGINOUT.EXE
Status:                   %NONAME-F-NOMSG, Message number 02DB8204

KWLITY_DRMMGR> 
%%%%%%%%%%%  OPCOM  11-FEB-1997 10:21:18.75  %%%%%%%%%%%
Message from user AUDIT$SERVER on KWLITY
Security alarm (SECURITY) and security audit (SECURITY) on KWLITY, system id: 11
673
Auditable event:          Object access
Event information:        directory entry creation request (IO$_ACCESS, IO$_CREA
TE, or IO$_ENTER)
Event time:               11-FEB-1997 10:21:18.74
PID:                      202022B0        
Process name:             PPCS_P_143A0045 
Username:                 APPIX_SERVER    
Process owner:            [_322,APPIX_SERVER]
Image name:               DSA0:[SYS0.SYSCOMMON.][SYSEXE]LOGINOUT.EXE
Object class name:        FILE
Object owner:             [_322,APPIX_PROD]
Object protection:        SYSTEM:RWE, OWNER:RWE, GROUP:RE, WORLD:E
Directory name:           _$1$DKA400:[000000]APPIX_SERVER.DIR;1
Directory ID:             (19809,2,0)
Directory entry:          NET$SERVER.LOG;0
Access requested:         READ,WRITE
Sequence key:             02DF8FB6
Status:                   %SYSTEM-F-NOPRIV, insufficient privilege or object pro
tection violation

KWLITY_DRMMGR> 
%%%%%%%%%%%  OPCOM  11-FEB-1997 10:21:18.96  %%%%%%%%%%%
Message from user AUDIT$SERVER on KWLITY
Security alarm (SECURITY) on KWLITY, system id: 11673
Auditable event:          Network login
Event time:               11-FEB-1997 10:21:18.95
PID:                      202022B0        
Process name:             PPCS_P_143A0045 
Username:                 APPIX_SERVER    
Process owner:            [_322,APPIX_SERVER]
Image name:               DSA0:[SYS0.SYSCOMMON.][SYSEXE]LOGINOUT.EXE
Remote nodename:          AWARD           
Remote node fullname:     DEC:.NIO.AWARD
Remote username:          APPIX_DEV


3859.6RMULAC.DVO.DEC.COM::S_WATTUMScott Wattum - FTAM/VT/OSAK EngineeringTue Feb 11 1997 12:013
Why don't you fix the priv problems you have on the directory?  If the detached
process that's being created cannot access the files that it needs to get at,
would you normally expect it to work?
3859.7Problem eliminatedREFDV1::LANZILOTTIThu Feb 13 1997 12:5716
    Scott,
    
    You are correct in stating that the priv problem needed to be
    corrected, the issue was what privs and where.  From an application
    perepective, however, all of the correct acls were on all of the correct
    accounts and files.  Since I am doing this without any system
    management help, excuse the following oversight.   Not knowing much at all 
    about system management, network processes, etc, it was not clear to 
    me that issue the was related to the server account's SYSUAF.  In fact, I 
    found the obstacle and did get the server to work simply by comparing the 
    2 sysuafs for the server account.  The current production  account had 
    network and BATCH access, the new account did not.  The command stream 
    used to enable the server and create the process is run in batch mode.  I 
    corrected the privs on the account and the server works.   
    
    Thanks to those who helped.