CODECOVERAGE

•Measure quality of testing;

•What lines of code has been hit during testing

•Specify CODECOV option at compile time

•Specify -l pgo DLL at link time

•Perform GAT/PAT testing (measure)

•Run HTML reporter on PC

NQSQL TIPS

Select unknown values only

For performance reasons it is better to select as few columns as possible. The optimizer may choose techniques like Index-only access. Besides it makes the query easier to read and maintain. In the following query selects five columns of table FOTB100. The first two columns are also used in the where-clause with “=”-predicates. So the values of these columns are no surprise and therefore it is not necessary to select these columns. So leave them out of the column-list.  SELECT         REK_NR       , PE_TX_REF       , ST_TS       , ST       , PRI  INTO        :REK-NR          OF FINTXSTH      , :PE-TX-REF     OF FINTXSTH      , :ST-TS              OF FINTXSTH      , :ST                   OF FINTXSTH      , :PRI                  OF FINTXSTH FROM  FOTB100  WHERE  REK_NR     = :REK-NR    OF FINTXST    AND  PE_TX_REF  = :PE-TX-REF OF FINTXST

Unicode - UTF 8, 16 and 32 Details

• UTF8 is variable 1 to 4 bytes.
• UTF16 is variable 2 or 4 bytes.
• UTF32 is fixed 4 bytes.

In brief, UTF32 uses 32-bit values for each character. That allows them to use a fixed-width code for every character.

UTF16 uses 16-bit by default, but that only gives you 65k possible characters, which is nowhere near enough for the full Unicode set. So some characters use pairs of 16-bit values.

And UTF8 uses 8-bit values by default, which means that the 127 first values are fixed-width single-byte characters. (the most significant bit is used to signify that this is the start of a multi-byte sequence, leaving 7 bits for the actual character value) All other characters are encoded as sequences of up to 4 bytes (if memory serves).

Unicode is a standard and about UTF-x you can think as a technical implementation for some practical purposes:

• UTF-8 - "size optimized": best suited for Latin character based data (or ASCII), it takes only 1 byte per character but the size grows accordingly symbol variety (and in worst case could grow up to 6 bytes per character)
• UTF-16 - "balance": it takes minimum 2 bytes per character which is enough for existing set of the mainstream languages with having fixed size on it to ease character handling (but size is still variable and can grow up to 4 bytes per character)
• UTF-32 - "performance": allows using of simple algorithms as result of fixed size characters (4 bytes) but with memory disadvantage

What is INSPECT?

Inspect is an interactive symbolic debugger that enables you to isolate program bugs quickly.

Few High-Level Inspect Commands which are specific to 100 coded object files are in detail below.

IDENTIFIER
Displays information about the internal characteristics of a given data location or of all data locations in one or more scope units. The IDENTIFIER command is a synonym for the INFO IDENTIFIER command.

INFO IDENTIFIER
The INFO IDENTIFIER command displays information about the internal characteristics of a given data location or of all data locations in one or more scope units, including native symbols.

Usage Considerations

• The INFO IDENTIFIER command can be used for active and for inactive scope units in processes, PATHWAY servers, and save files. In a PATHWAY requester program, the INFO IDENTIFIER command can be used only for active scope units.
• Types of Entities - The form of the output produced by INFO IDENTIFIER for a given identifier depends on the type of entity the identifier denotes.
• INFO IDENTIFIER Presentation for Variables - A typical program contains more identifiers for variables than for any other entity class.

MATCH
The MATCH command searches for scope-unit names or other identifiers in the current program.

Default Values

• If the MATCH IDENTIFIER command is entered with only one pattern, Inspect will display only the matching symbols, and the scopes they were found in.
• If the VERBOSE clause is omitted, scope names are displayed only if they have a matching pattern.

Usage Considerations

• Matching Uppercase and Lowercase Letters - The MATCH command does not distinguish between uppercase and lowercase letters in the pattern unless the current language is C.
• Alias Restrictions with the MATCH command - Aliases are not be expanded in the pattern with the MATCH IDENTIFIER or the MATCH SCOPE commands.
• Matching Scopes With a Leading “#” - You must include a “#” to match scopes which have a leading “#”. For example, to match the scope, “#global”, enter: MATCH SCOPE ##global

Impacts and Top Recommendations for HP Integrity NonStop Users

Different Object File and Native Processor Types

Object File Code	Object Type	Native Processor Types
100	TNS	CISC (Complete Instruction Set Computing) for the Tandem Processor. - TNS objects (COBOL, TAL, etc) will be interpreted on Integrity Platforms
700	TNS/R	RISC(Reduced Instruction Set Computing) for the MIPS Processor - TNS/R Native object files cannot be executed on a TNS/E Platform. All TNS/R  Native Programs must be recompiled using a TNS/E Native Compiler(for Native 800 Object) OR the TNS Compiler(for interpreted 100 Object)
800	TNS/E	EPIC(Explicitly Parallel Instruction Computing) for the Intel Itanium Processor - TNS/E Native Object Files cannot be executed on a TNS/R Platform. - OSS Programs must be migrated to TNS/E Platform - All TNS/E Native Code is position independent code (PIC)

Corporates today depend on HP Integrity NonStop mission-critical servers

Modern HP Integrity NonStop Servers continue to deliver on hallmark attributes -
   - Availability,
  - Scalability and
  - Data Integrity

The HP NonStop Server continues to be the industry leader in mission-critical availability and reliability, with an extensive footprint in financial services and telecommunications marketplaces, and the customers who deploy NonStop have been spared the ignominy that can come with unwanted media coverage.

Few recent wins.

- In America, PULSE selects four NB54000c for Credit Scoring

- In India, Mobile phone operator depends on 300 NonStop processors

- In Norway, Healthcare is operating with NonStop

- In Singapore, Shippers move containers NonStop

- In Germany, auto manufacturer never unseated

HP NSS Vs HP NSB(Blade Servers)

For Migration from COBOL to Java - POC Details

HP and Blu Age® completed a multi-step migration and modernization Proof of Concept (POC) project that took a legacy OLTP application from COBOL and DB2 on an IBM mainframe,

 - First to Java and DB2 on the HP ProLiant DL360 platform, &

 - Ultimately to Java and SQL/MX on an HP NonStop Server.

The project included comparisons of transaction execution times in the

    - ProLiant environment,

    - a hybrid NonStop and

    - Windows environment, and

    - a “pure NonStop” environment.

The NonStop Server was clearly the overall winner for the set of transactions comprising this POC.

The Blu Age1 software suite implements

- Model Driven Architecture (MDA) principles &

- Offers appropriate methods to deliver Service Oriented Architecture (SOA) application programs written in Java.

The Reverse Modeling component facilitates the extraction of business logic from the legacy code and constructs UML (Unified Modeling Language) models that can be used to generate the application in a Java framework.

The Blu Age software is packaged as a plugin for Eclipse, an open source IDE (integrated development environment) frequently used to develop Java applications.

POC configuration

Configuring the hardware and software for this POC required no extraordinary efforts.

Standard 1 GB Ethernet connections were used to connect these hardware components used for testing:

• NonStop Server: NS16004, with 10 mirrored data drives, running H06.25, SQL/MX 3.2, and NonStop Java 6.0 (1.6).

• Windows server (for Java middle layer): HP ProLiant DL360 G5, running Windows Server 2008, Java 1.6 and Blu Age 3.7.

The middleware setup required installing JBoss 7.1.1 and its dependencies.

A few minor (and common) DDL adjustments were made when the NonStop SQL/MX tables were created:

- VARCHAR columns with lengths less than 20 were changed to fixed CHAR columns of the same length. (This is a typical optimization practice for NonStop SQL.)

- BIGINT columns were changed to the equivalent NonStop SQL/MX data type, LARGEINT.

- Columns declared with DEFAULT NOT NULL syntax were given a default value of 0.

Availablity , Statistics and Performance (ASAP)

ASAP Monitors hundreds of Objects in Real-Time.

It Provides goals and actions to repair or restart system and Application Components.

ASAP has a wide range of visual and programmatic interfaces,
       - Including EMS,
       - Conversational interface,
       - Fat and thin client interfaces,
       - Provider/Consumer APIs
       - Email
       - Phone Notifications with color-coded HTML reports
       - Customizable GUIs with user-defined Dashboards and
Long term Historical Database with up to eve 10 years worth History of all Availablity,
statistics and performance data.

ASAP Can integrate with HP operations manager.

ASAPX in an API into the ASAP infrastructure to monitor the availablity and performance ob abstract application domains.

ASAP Hybrid plug-in for Linux integrates Linux Application metrics into ASAP, providing a visa of the entire application spanning across Nonstop and Linux systems.
       

Checklist - To migrate into OSS - VI - And Thats it :)

After System and Database updates, Next step would be Application Installation.

ID	Task Description	Prior Task	Resource	Remarks
24	Place Application in Dir. Structure	11,19	Applications	Expand the Application Delivery Package and Place into the proper delivery structure. This often requires utilities like TAR,GZIP and/or Jar.
25	Configure Pathway and Serverclass	24	Applications	Configure Pathmon/Pathway Naming, Locations and server classes(including environment variables) - If needed
26	Deploy .war files for web interfaces	25	WebAdmin	Deploy distributed .war files
27	Configure Externalized properties for web interface(s)	26	WebAdmin	Modify Externalized properties for IP addresses, catalogs, schemas and user-id.

ID	Task Description	Prior Task	Resource	Remarks
28	Configure Online Dumps	21	Database, Systems, Operations	Use TMFCOM to configure online dump schedules for database protection
29	Configure Scheduled Backups	27	Database, Application, Operations, Systems	Determine backup fileset, schedule and methods; Implement in batch scheduler.(Each OSS pathname has a corresponding Guardian Filename)
30	Identify Logs to Monitor	27	Database, Applications, Operations, Systems, Interface Group, WebAdmin	Identify Logs to Monitored across the enterprise. Ported Applications will NOT typically use VHS or EMS.

Checklist - To Migrate into OSS - V

When the System is READY, the next steps is for DBA Team to prepare and execute DDL Scripts and prepare database for use.

ID	Task Description	Prior Task	Resource	Remarks
20	Modify DB scripts for Disk Layout	3	Database	Edit SQL DDL Scripts to distribute the database.
21	Create the Application Schemas and Tables	20.21	Database	Obey the scripts in MXCI to create the Application schemas and Tables
22	Create MetaData (catalog) convenience views{optional}	21	Database	Download and run in OSS the mxschema utility to create views of database metadata tables
23	Create datasource for J/ODBC access	18,21	Database	Use mxcs or NSM/Web to create a data source.

Checklist - To Migrate into OSS - IV

ID	Task Description	Prior Task	Resource	Remarks
16	Create Resolver Links for TCP/IP	15	Systems	Edit, In Guardian Space, $system.ZTCPIPHOSTS & RESCONF, In OSS, create symbolic links for HOSTS, RESOLV.CONF, PROTOCOLS, NETWORK and SErvices in the /etc directory
17	Create Data service	16	Systems	Use SCF to add a Data Service
18	Add "DBA.manager" to MXCS	17	Systems	Use MXCS mode of MXCI to grant OPERATOR privileges to the dba.manager user.
19	Install any open source utilities	18	Systems	Identify the useful utilities and install those. Modify the /etc/profile and users .profile files.

Checklist - To Migrate into OSS - III

ID	Task Description	Prior Task	Resource	Remarks
11	Create Application Directories	10	Systems	Establish OSS directory for Application like APPL, BIN etc
12	Install SQLMX	11	Systems
13	Install Java	12	Systems
14	Install Java-addons and other software required for Application	13	System	Following Components needed - J/ODBC drivers - JToolKit- Contains APIs - Javaservlets - Required for WebApps - HP Jmeter for Nonstop Java - Performance Anlayse Tool
15	Initialize JDBC	14	Systems

Checklist - To Migrate into OSS - Part II

ID	Task Description	Prior Task	Resource	Remarks
6	Move QIO to KSEG2 (Required for Java)	3	Systems	Use SCF to move QIO to KSEG2. And Reboot.
7	Configure TMF	6	Systems	Use TMF to configure TMF components and Objects: Audit Trails, Dump and restore volumes etc
8	Create Mount Point Dir	8	Systems	Use OSS to establish mount point directory
9	Create Disk Pool	9	System	Create Disk pool Definition Files / Scripts
10	Start the Filesets	10	Systems	Use SCF to add and start a name server If needed. Then, add and start a fileset for the new mount point.

Checklist - To Migrate into OSS - Part I

Tokenization : Creating Secure Test Data

PCI DSS Section 6.4.3 requires that "Production Data (live PANs) are NOT used for Testing or Development".

If we need a Huge Dataset of real looking PANs for a Test. How do we create those?

A good Tokenization will allow you to do "BATCH COPY" where you start with PAN Data and end up with Tokens.

Doing this in production once is a fast and secure way to create test data: when configured properly, the tokens can look like "REAL" PANs and can even pass Luhn check.

Thus creating a huge set of Test data which is NOT tied to real PAN data at all becomes easy exercise.

Challenges of being a Tandem DBA

• Repartition
• Reviewing and updating key values
• Merging Partitions
• Reuse Partitions
• Update statistics
• Gathering SQL/MX statistics
• Copy Statistics from One Table (or system) to Another
• Index maintenance
• Create
• Populate
• Drop
• Alter
• Specifying Access Previleges
• Grant
• Revoke
• Query Performance Tuning
• Building and Deploying Standard Queries
• Dataload/Unload
• Moving Data from MP to MX for Testing

NSSQL : Varchar columns

Varchar in COBOL

Invoke of a varchar in COBOL gives the following result:_

invoke testtab format cobol;

* Record Definition for table

* Definition current at 16:08:08 - 09/06/01

01 TESTTAB.
    02 NAAM PIC X(4).
    02 LANG.
             03 LEN PIC S9(4) COMP.
             03 VAL PIC X(255).__

Assigning a value to a VARCHAR column in COBOL

On insert, the length-field of the VARCHAR-columns must be specified by the program as well
as the value-field.

The length-field will be filled automatically when the TRIM-function is used._

Example:_
INSERT INTO =EMS-LG
(
. . .
, TXT
. . .
)
VALUES
(
. . .
, TRIM(:VAL OF TXT OF WS-EMS-LG)
. . .
) _

Note: when the source-string is nullable, an indicator-field has to be used as well._
Example_
TRIM(TRAILING " " FROM ( :VAL OF TGV-REK-NR OF table_
INDICATOR :INDICATOR OF TGV-REK-NR OF table))

Assigning a value to a VARCHAR column in SQLCI

If you want to assign a value to a varchar-column with a string that is longer than 80
characters using SQLCI, then use the following syntax:_

insert into =test-tabel values ("test",
"1234567890123456789012345678901234567890123456789012345678901234567
890" ||
"1234567890123456789012345678901234567890123456789012345678901234567
890" ||
"1234567890123456789012345678901234567890123456789012345678901234567
890" ||
"123456789012345678901234567890123456789012345");_
--- 1 row(s) inserted._

Select in SQLCI

Selecting a varchar column in SQLCI only shows the first 80 characters._

select long from testtab;_

LONG
-------------------------------------------------------------------------------------------------_
123456789012345678901234567890123456789012345678901234567890123456789012345
67890_
--- 1 row(s) selected._

If you are interested in all characters, use the report writer command SET STYLE. Change the
style-option VARCHAR_WIDTH. Maximum is 255._

set varchar_width 255;

select langejan from testtab;_

LANGEJAN_
--------------------------------------------------------------------------------_
--------------------------------------------------------------------------------_
---------------_
123456789012345678901234567890123456789012345678901234567890123456789012345
67890
123456789012345678901234567890123456789012345678901234567890123456789012345
67890
123456789012345678901234567890123456789012345678901234567890123456789012345
67890
123456789012345_

--- 1 row(s) selected._

For columns even longer than 255 bytes the LIST_COUNT-option of the SQLCI-session can be
used:_

set list_count 1;_

select constrainttext from =constrnt
where tablename like upshift(“%BTCHFILE%”);_
detail constrainttext as C3000.75;_

list first; (or short: l f )

(also list next / list all can be used)_

**** next = enter_
**** to cancel the SELECT when not all result rows are shown
cancel;_
set list_count all;_

NSSQL : Nullable columns

Nullable columns in search-conditions Sometimes the use of nullable columns may lead to unexpected results. Note that a row is not selected when the expression in the search-condition evaluates to NULL. Example: Table definition: -- Definition of table    (      ACT_NR                           PIC 9(2) COMP NO DEFAULT NOT NULL   , KANTOOR                        PIC S9(4) COMP NO DEFAULT NOT NULL   , LOCATIE                           PIC 9(2) COMP NO DEFAULT NOT NULL   , VP                                      PIC S9(4) COMP NO DEFAULT   ) Actual data select * from rekening ACT_NR  KANTOOR  LOCATIE  VP    ------        -------          -------    ------       1              1                  1         11       6              1                  1          ? --- 2 row(s) selected. The following query might not deliver the result you expect select * from rekening where vp <> 55; REK_NR  KANTOOR  LOCATIE  VP   ------           -------        -------    ------      1                 1                1         11 --- 1 row(s) selected. The row with a NULL value in the vp-column is not selected. NULL is not different  from 55 but also not equal to 55. It is simply unknown. If the row with the NULL-value has to be selected also, then following syntax has to be used: select * from rekening where vp <> 55 or vp is NULL; REK_NR  KANTOOR  LOCATIE  VP   ------          -------         -------    ------       1               1                1         11       6               1                1          ? --- 2 row(s) selected. Nullable columns in COBOL Another issue to watch carefully is the use of a nullable column in a search-condition in a COBOL program. If the value of a nullable column has to be compared to a host-variable in a search-condition, and the host-variable must be able to contain NULL, then the following construction has to be used: Select ……………   from ……………………   where <other selection criteria>   and (column_name = :valu of column-name and :indicator of column-name > -1            or column_name is null                          and :indicator of column-name <  0) Do not use the following construction (SQL does not recognize the indicator): Select ……………   from …………………… where <other selection criteria>    and (column_name = :valu      of column-name             indicator :indicator of column-name) Suppose a table TEMP. The invoke under SQLCI gives the following result: invoke testtab; -- Definition of table -- Definition current at 16:19:11 - 09/06/01   (     TEMP-ID                        PIC X(5) NO DEFAULT NOT NULL   , OMS                            PIC X(80) NO DEFAULT   , OMS-LANG                       VARCHAR(255) NO DEFAULT NOT NULL   ) Invoke in a COBOL program results in: invoke temp as temp 01 TEMP.     02 TEMP-ID                          PIC X(5).     02 OMS-I                            PIC S9(4) COMP.     02 OMS                              PIC X(80).     02 OMS-LANG.        03 LEN                           PIC S9(4) COMP.        03 VAL                           PIC X(255). Using this type of invoke in COBOL the nullable column cannot be treated as one object. It results in two host-variables on the same level (OMS-I en OMS). This structure also differs from the host-variable structure for a varchar column. A host-variable for a varchar has two levels: the column-name on the highest level and the “LEN” and  “VAL” fields at the lower level. Rather use the next syntax in COBOL programs: invoke temp as temp null structure; Result: 01 TEMP.     02 TEMP-ID                          PIC X(5).     02 OMS.        03 INDICATOR                     PIC S9(4) COMP.        03 VALU                          PIC X(80).     02 OMS-LANG.        03 LEN                           PIC S9(4) COMP.        03 VAL                           PIC X(255).  Now the host-variable for the nullable column has two levels, with at the lowest level two fields with the names “indicator” en “valu”.