Get
Ready for Insight 204!
CCA's Insight 204 Symposium
is fast approaching. If you haven't already registered, you should do so
soon. We've planned three days of information-packed sessions, each designed
to help you and your organization meet today's IT challenges by maximizing
your investment in Model 204. We want you and other members of your organization
to join us for this informative and educational event.
This article is a list
of the most frequently-asked questions about Insight 204. For additional
details on all the following answers, the most up-to-date information about
Insight 204, and to register on-line, please consult the Insight 204 Symposium
Web site at www.cca-int.com/insight/main.html.
Where and
when is the symposium being held?
Insight 204 is being held at The Westin Waltham-Boston in Massachusetts.
Kick-off is on Sunday, April 9th, at an evening welcome reception in the
hotel. The sessions run from Monday, April 10th through Wednesday, April
12th.
Is
the symposium really free of charge?
Yes, registration is
free! To ensure that you get the most out of your Model 204 investment,
we want to provide you with three full days of information and networking
opportunities. You pay for only your own travel and lodging expenses.
What
are some of the sessions?
We've got more than 20 different technical sessions planned, covering everything
from Model 204 performance tuning to building new solutions with new products.
Wednesday features mini training sessions that provide more in depth knowledge
than you could obtain in a 60 or 90 minute session.
Are third-party
vendors participating in Insight 204?
Absolutely. We value
our business partners and recognize their importance to your Model 204 applications.
Insight 204 features presentations from several business partners including
IBM, Sirius Software, ITS, and WebXi. We are also providing exhibit space
for partners as well. Visit the Partner Pavilion page of the Insight 204
Web site to keep abreast of partner activity at Insight 204.
I've heard
that there is a special training class on April 13th? Is that true?
Yes, it is true. Immediately
following Insight 204, there is a special, one-day Performance Tuning training
class that is being held at CCA on Thursday, April 13. The cost of the class
is $250.
Will I be
able to get any one-on-one time with Model 204 developers and experts?
Yes, Insight 204 hosts
a CCA Drop In Center that will be manned by product experts throughout the
week, allowing you to get some one-on-one time with the developers, and
to see product demonstrations in a more personal environment. You'll find
all the latest product literature here, as well as demonstration CDs, documentation,
and more.
If you have additional
questions about Insight 204, please submit them to us through the Feedback
form on the Insight 204 Web site, or contact Marie Kelly.
We look forward to seeing you in April!
Authorizing
Users for ODBC Driver
System
1032 — Frequently Asked Questions
By David Stone
Standard OpenVMS and System 1032 security
profiles determine ODBC Driver access to files and data at all levels. In
addition, you can restrict access to the exported datasets and views from
within the ODBC Dictionary utility, depending on your version of the ODBC
Driver. These security profiles assume an authorization process is in place.
This article addresses the complexities of authorizing and tracking ODBC
Driver users, as the context switches between server and client.
Tracking
context change
Q. When in the
process does the context change from the server to a given user or back
to the server?
A. The context changes
from server to user after OpenVMS validates the client user name and password,
but before System 1032 is started.
Q. What work does the
server do before changing the context to the user?
A. The initial server
context validates the user name and password, performs required network
tasks, and initializes internal data structures.
Q. What is done in
the user's context?
A. All access to System
1032, including the metadata dictionary, stored procedures, and the execution
of initialization files, is in user context, which is maintained until the
end of image.
Q. What work is done
after OpenVMS reverts back to server context?
A. For OpenVMS 7.0
and earlier, user context is used through the end of process.
With OpenVMS 7.1, ODBC
Driver reverts to server context only to rundown the process. This means
that all datasets, including the metadata dictionary, are closed in user
context. Additional work in server context can be performed by modifying
the S1032ODBC_LIB:S1032ODBC.COM and executing commands after the line that
reads:
$ S1032ODBC3 ODBC.INI
As shipped, the only
commands executed after the server process exits are to deassign the logical
names redefined earlier in the process. CCA recommends that you do not develop
an application that depends on this behavior, as this is subject to change
in a future release.
The
role of network services
Q. How much of the user's
environment is set up? Are SYLOGIN.COM and LOGIN.COM executed in the user's
context?
A. The user's environment
setup is dependent on your network service. The ODBC Driver server executes
as a TCP/IP for OpenVMS server process. When a network connect arrives at
the ODBC Driver port, the master server, maintained by your network service,
starts a new process using the privileged account that owns the ODBC Driver
server. The LOGIN.COM that belongs to the privileged account is executed,
not the one that belongs to the client user.
Depending on your network
stack, you cannot execute command procedures. This is a network configuration
issue. Currently the system-wide or user-specific login command procedures
cannot be rerun in user context.
Q. Is the system-wide
System 1032 initialization file used, if one is present? In what context,
user or server?
A. Yes, in the user's
context. Please be reminded that since the evaluation of S1032_INI logical
name occurs in user context, users who do not have read access to that file
might be excluded. In the following example, only users with access to DISK$PRIVATE[NO.PUBLIC.ACCESS]S1032.INI
can execute the commands in the S1032.INI file.
$ DEFINE S1032_INI DISK$PRIVATE[NO.PUBLIC.ACCESS]S1032.INI
To minimize access
problems you can define the S1032_INI logical name to execute in the user's
default directory, not the server's, as follows:
$ DEFINE S1032_INI SYS$LOGIN:S1032.INI
Note: Not all users name the initialization file S1032.INI or locate it
in their default directory.
What's the
OpenVMS version got to do with it?
Q. We are running OpenVMS 7.2. How can we tell which client
is using ODBC Driver since OpenVMS writes all accounting records to the
server process?
A. For OpenVMS 7.0
and earlier, the accounting record under the server context includes all
the work performed up to the context change. Then, a second accounting record
is written, this time in user context.
Beginning with OpenVMS
7.1, ODBC Driver uses the OpenVMS PERSONA system services rather than the
customized kernel code developed at CCA. The kernel data structures were
greatly modified to support threads in OpenVMS 7.1; Compaq has continued
modifications in subsequent releases of OpenVMS. For OpenVMS 7.1, process
rundown always occurs in the server context. ODBC Driver cannot write an
accounting record from within the server image.
In OpenVMS 7.2, an
application can instruct the PERSONA system services to maintain user context
through the end of the process, which would write the accounting record
for the entire process in user context. ODBC Driver does not yet implement
this functionality.
Authorizing
users
Q. How are users authorized?
A. ODBC Driver authorizes
a user name and password by retrieving the authorization record for that
user via the SYS$GETUAI system service, hashing the password using the SYS$HASH_PASSWORD
system service, and comparing the result against the hashed password stored
in the authorization record. This is exactly how the LOGINOUT service works.
ODBC Driver also obeys
the DISUSER flag that does not allow a disabled user to login. In addition,
ODBC Driver simulates the break-in detection process: a user with more than
ten successive failed logins is not allowed to connect via the client. The
high number of failures tolerated accommodates Microsoft Access and Microsoft
Query that can attempt to connect multiple times using various combinations
of no username and/or no password.
Q. Is external authorization
an issue for ODBC?
A. No.
Autofix:
Part 1
Model 204 — Frequently Asked Questions
by Donna Goodwin
Autofix
provides an interactive system to download and/or apply Model 204 and User
Language product Early Warnings. This two-part article addresses the questions
most frequently asked of CCA Customer Support regarding Model 204 nucleus
Early Warnings. You can define Autofix as an APSY subsystem that you invoke
using subsystem commands, or you can invoke Autofix by issuing the following
commands:
OPEN
EWDPROC
INCLUDE AUTOFIX
On
the Main Menu screen SCRN:1.0, enter the product, release number, and operating
system. Then, select Function 1: DISPLAY/APPLY Early Warnings to navigate
first to SCRN: 2.0, then to SCRN:3.0. Figure 1 shows Early Warnings 53 to
62, which include zaps; special handling zaps, indicated by a Y* in the
ZAP? column; and object replacement, indicated by a Y in the OBJECT? column.
Enter any character beside the Early Warnings you want to process; the example
uses an X.
Figure 1. A partial display of SCRN: 3.0 listing fixes for
Model 204 V4R2.0
What happens when I enter Apply (PF10 key)?
After selecting the
Early Warnings you want to download, Autofix writes the fixes to one or
more OUTxxxx files. OUTxxxx files are displayed on the SCRN:
9.0 screen, shown in Figure 2, which is accessible from most Autofix screens
by entering Usefile (PF4 key) at the ===> prompt.
Autofix
writes each of the zaps you selected to the OUTZAPS file. If the fix requires
special handling, the zap is written to the file, but it is commented out.
If the fix requires an object replacement, the object code is written to
an OUTcsect file, such as OUTBXCI, as shown in Figure 2.
Figure 2. A partial display of SCRN: 9.0 listing the OUTxxxx
files for zaps and object files
Most
nucleus Early Warnings have a zap portion that is written to the OUTZAPS
file. Some might also contain an object replacement that is written to an
OUTcsect file, such as OUTEVCP.
Note:
With each successive release of Autofix, additional OUTcsect files
might be added. Once an OUTcsect file is present, it remains there
for the life of that Model 204 release.
What are these OUTxxxx files?
The OUTxxxx files
are the input to jobs you must run to incorporate the fixes. For example,
in the case of an Early Warning that requires object replacement, you run
a zap job and a relink (or regen).
The
OUTZAPS file contains all the zaps for the Early Warnings you selected.
It is a sequential file that can be allocated as DSORG=FB, LRECL=80, BLKSIZE=8000.
The
OUTcsect files, such as, OUTBXCI, OUTEVSO, and OUTEVCP, are where
your object replacement modules are routed. In MVS and VSE, allocate these
files with the same characteristics as your original OBJLIB file created
during installation.
All
the OUTxxxx files must be accessible to the Model 204 Online where
you are running Autofix. Make them available to the run by adding DD cards
in MVS, FILEDEFs in VM, or DLBLs in VSE.
//OUTZAPS
DD DSN=prefix.OUTZAPS
//OUTBXCI DD DSN=prefix.OBJFIXES=BXCI
//OUTEVSO DD DSN=prefix.OBJFIXES=EVSO
ALLOCATE
OUTZAPS WITH DSN=prefix.OUTZAPS OLD SEQUENTIAL
ALLOCATE OUTBXCI WITH DSN=prefix.OBJFIXES OLD SEQUENTIAL
MEMBER(BXCI)
ALLOCATE OUTEVSO WITH DSN=prefix.OBJFIXES OLD SEQUENTIAL
MEMBER(EVSO)
- For
VSE, after assigning a punch or output file:
DEFINE
DATASET OUTZAPS WITH SCOPE=SYSTEM FILENAME=SYSnnn LRECL=80
Why have object replacements?
CCA sends an object
replacement when a fix is very large or complex. An object replacement has
an added advantage of performing better, because the patch area does not
need to be accessed.
What is a special handling Early Warning?
A special handling Early
Warning, identified in the ZAP? column with the value Y*, might indicate:
- Special
instructions or subsequent steps required, for example: apply the zap
and reorganize the file.
- Fix
applies to an optional product such as Horizon.
- Early
Warning contains an object replacement.
- Fix
is only available on the CCA Web site.
- Fix
is to source code.
- There
is more than one option for applying the fix.
To
properly manage a special handling Early Warning, you must take the following
steps:
- Examine
the reason for special handling by selecting the Early Warning and entering
DISPLAY (PF9 key) at the ===> prompt.
- Special
handling zaps written to the OUTZAPS file are commented out with an asterisk
(*) next to a white space( ). If you want to apply the zap, you must edit
the OUTZAPS file by removing the asterisk-space.
Next
question?
In Autofix: Part II, we will
discuss relinking and applying zaps, depending on your operating system.
Education
Schedule
April—May 2000
The
scheduled location for the following Model 204 courses is Framingham, Massachusetts.
| Course
|
Dates
|
Location |
|
Model
204
|
| Introduction
to User Language (UL150) |
3/22-24
|
Framingham,
MA
|
| Fine
Tuning Your Applications* |
4/13 |
Framingham,
MA |
| Implementing
Online Applications (UL250) |
4/24-27 |
Framingham,
MA |
| Application
Subsystem Facility (AP100) |
4/27 |
Framingham,
MA |
| Programmer’s
User Language (UL200) |
5/1-5
|
Framingham,
MA |
| System
Performance and Tuning (SM350) |
5/10-12 |
Framingham,
MA |
| Introduction
to System Management (SM100) |
5/22-24 |
Framingham,
MA |
| Problem
Determination & Resolutio (SM160) |
5/25-26 |
Framingham,
MA |
*Fine
Tuning Your Applications
Have you ever wondered
what techniques long-time Model 204 programmers use to improve the performance
of their applications? In this special one-day class, we'll look at those
techniques using a live application. Learn about available tools to measure
performance and retrieval efficiency and how to determine which approach
is best.
Topics
Include:
- Using
tools and parameters to measure performance
- Utilizing
AUDIT/204 and deciphering an audit trail
- Working
with DEBUG SUBSYSTEM and other commands
- Using
helpful parameters including DEBUGUL
- Analyzing
a FIND
- Identifying
optimal index structures for specific retrievals
- Counting
the I/Os -- Retrieval Efficiency
- Referencing
a Previous FIND
- Splitting
a FIND
- Sorting
records and the effects on buffers
- Using
sort record Keys
- Ordering
records with the ordered Index
- Ordering
records with FRV
- Sorting
Records -- Few vs. large number of records
- Using
FRV efficiently
- Retrieving
Data from a Record
- Writing
efficient code for multiply occurring fields
- Processing
multiply occurring groups
- Making
the most of image and screen processing
- Variable
Usage
Space
is still available. The cost is $250 per person and includes continental
breakfast and lunch.
Transportation
will be provided to and from the Westin Waltham-Boston, the Insight 204
hotel.
