APPX System Administration

2009-Sep-04 5:34pm

                     export APPX_IMPORT_COMMIT=5000

                     export APPX_IMPORT_COMMIT=5000      

                      $APPXPATH/appx.env
                      $HOME/appx.env

                      ORACLE SID=APPX
                      APPX DBG CODE=OxFFFFFFFF
                      Etc.
The --- COMMIT MODE pdf has four valid values.  (You may see AUTOCOMMIT as 
well, but it's not implemented yet):

 MANUAL     -  Appx will not automatically COMMIT your changes.  You must use
               the COMMIT statement.

 END        -  Appx COMMITs your changes when the process ends.

 PCF CYCLE  -  Appx COMMITs your changes as you progress from PCF record to 
               PCF record.

 COMPATIBLE -  Appx COMMITs your changes as soon as -all- record locks are 
               released.

Appx never automatically issues a ROLLBACK.  Based on the structure of your 
application, there may be other places where you want to COMMIT a transaction.  
That's what the COMMIT statement is for.  One thing to keep in mind is that the
frequency of COMMITs will affect performance.  The more COMMITs you do, the 
slower your application will be.  For interactive processes, you can probably 
ignore the performance issue, but for batch-mode operations, COMMITing too 
often can really slow things down.

Another issue to be aware of is that AppxIO files do not support transaction 
processing.  If you ROLLBACK a change to an RDBMS-hosted file, any changes made 
to AppxIO will not be rolled back.  Related to this is the issue of visibility.  
Changes made to an RDBMS-hosted file are invisible to other users until they 
are committed.  Changes made to an AppxIO-hosted file are visible immediately.

Appx also supports save-points.  A save-point is a partial transaction.  You 
mark the beginning of a save-point with the SAVEPNT statement - the only 
argument is a name of your choosing. Executing a SAVEPNT tells the RDBMS that 
it should write a marker into the current transaction.  If you find that you 
want to ROLLBACK part of a transaction, you can specify a save-point name in 
the ROLLBACK statement.  The transaction is still open and you can commit other 
changes, but changes made since the save-point was created are lost.
Example: moz /users/pat> appx –u

Usage: -u                          list these options
       -k                          define a new keymap
       -m=[user_id:]keymap_name  override keymap ID
       -d=<database_id>            override startup database
       -a=<application_id>         override startup application
       -t=<process_type>           override startup process type
       -p=<process_name>           override startup process name
       -i=<flat_file_name>         convert a flat file into UNIXIO format
       -e=<data_file_name>         convert a UNIXIO file into a flat format
       -v=<data_file_name>         verify the integrity of a UNIXIO file
       -an=<data_file_name>        analyze a UNIXIO file
       -l=<log file name>          start a license server

Restrictions:

An application_id and process_type must always be included when specifying 
a process_name.  If an application_id is included without specifying a 
process_type/name, the startup menu for the application will be used.

If no application_id is specified, the startup application and menu for the 
database will be used.  If no database_id is specified, the startup database 
defined for the user within System Administration is used.  If the startup 
database is also unspecified, the Appx main menu is used.

The valid process_types are Menu, Job, Input, Output, Update, Action, Inquiry, 
Query, Status, and Subroutine.  Use underscores (_) in place of any embedded 
spaces within parameters.
 1) Enter a server name in the Appx/Oracle FMS group (either an Oracle
    SID, or the name of a remote server as described below).
 2) If you leave the server name blank, the OCI uses the environment
    variable $TWO_TASK to locate a remote server.
 3) If you leave the server name blank and $TWO_TASK is not defined, the
    OCI uses the $ORACLE_SID environment variable to locate a local
    server.
       $ORACLE_SID:$ORACLE_HOME:<N|Y>:

     oracle:/opt/oracle/product/9.2.0:N
      ORACLE.INTERNAL.APPX.COM =
      (DESCRIPTION =
      (ADDRESS_LIST =
      (ADDRESS = (PROTOCOL = TCP)(HOST = badger.appx.com)(PORT = 1521)))
      (CONNECT_DATA =
      (SERVER = DEDICATED)
      (SERVICE_NAME = oracle)))
 1) If you know your Oracle server is installed on the same machine as
    your Appx server,

       a) Examine the /etc/oratab file to find your Oracle SID

       b) Examine the /etc/oratab file to find your $ORACLE_HOME

       c) Make sure $ORACLE_HOME is defined in your start-appxd.sh 
          script

       d) Create an Oracle FMS group (type 5) and enter the Oracle SID
          in the server name field

 2) If you know your Oracle server is installed on a remote host,

       a) Examine the /etc/oratab file to find your $ORACLE_HOME.   (In 
          this case, $ORACLE_HOME will tell the OCI where to find config 
          files.)

       b) Examine the 'tnsnames.ora' file to find your remote server 
          name. 
       c) Make sure $ORACLE_HOME is defined in your start-appxd.sh 
          script.

       d) Create an Oracle FMS group (type 5) and enter the remote
          server name in the server name field.
         www.appx.com/assets/asp/dynamic_generator.asp?pageid=605

         www.appx.com/assets/asp/dynamic_generator.asp?pageid=682
      BEG READ 0AD PROCESS                HOLD 1 KEY IS  PROCESS KEY
      END READ 0AD PROCESS

 Waiting for lock held by APX in /usr/appx/data/00/SUR/Data/PROCESS.dat
 Waiting for lock held by session 9999 in /appx/data/00/SUR/Data/PROCESS.dat

 Windows - use SysInternals.com's Process Explorer utility:
           http://www.sysinternals.com ... look for "Process Explorer"

 AIX     - http://aixpdslib.seas.ucla.edu/packages/lsof.html
 HP/UX   - http://hpux.cs.utah.edu/hppd/hpux/Sysadmin/lsof-4.70
 Red Hat Linux - part of the lsof RPM package
 Solaris - use the fuser command instead of lsof

How can I easily reprint a file that's has been requeued (saved in the print
directory) without using the winprint from the dos prompt?  The file in
question is checks.  Once you print them you can't rerun the process.  
I cannot use the winprint command from dos because I need to grab the form 
file so it spaces everything correctly.

When you requeue an APPX report, it leaves the print file and a configuration 
file in the user's print directory after they've been submitted to the print 
spooler.

For example, when printing a System Administration list, the print file might  
be SYSADMI0 and the configuration file SYSADMI0.cfg

If you wanted to reprint this report, you might submit commands like the  
following:

If you are on Windows:

     cd  C:\appx\data\PRINT\USR
     C:\appx\data\winprint.exe  -config=SYSADMI0.cfg  SYSADMI0
or
     C:\appx\data\winprint.exe  -config SYSADMI0.cfg  SYSADMI0

The general form is:

     winprint.exe  -config=<config.file>  <print file>
or
     winprint.exe  -config <config.file>  <print file>

For those on Unix/Linux:

/usr/local/appx/data/appx_print -config=/usr/local/appx/data/PRINT/USR/SYSADMI0.cfg
/usr/local/appx/data/PRINT/USR/SYSADMI0

You can construct and issue shell (DOS) commands by means of the RUN command. 
You'll find documentation for it in the Designer Reference Manual, and at 
www.appx.com/pub/documents/HowTo/run.Windows.Functions.html 

The print file name can be found in --- PRINT FILE PATHNAME.  You'd be well
advised not to operate on the file pointed to by --- PRINT FILE PATHNAME, 
until after you've exited the OUTPUT process that generated the print file.  
This means passing the contents of --- PRINT FILE PATHNAME to a subsequent 
job step, possibly by means of a related or detached work field.
       Building Dynamic Index
       You are building a dynamic index for
       a possibly large number of records.
       This process may take a long time.
       Please CANCEL to stop building index.
       Press RETURN to continue.
        APPX_DYNAKEY_PROMPT=5000
        APPX_DISABLE_DYNAKEYS=yes
  C:\Documents and Settings\{yourWindowsID}\.appx\cache\

  C:\WINNT\.appx\cache\

  $HOME/.appx/cache/

Method #1:  Start > Settings > Control Panel > Administrative Tools >
                    Data Sources (ODBC) > Drivers Tab > Version

Method #2:  Windows Explorer > {your Windows Directory} > System32 >
                    AppxODBC.dll > Properties > Version Tab > "File
     IF  you are on 4.1:
     recover  FILE.bad.dat  FILE.dat > FILE.corrupted

     If you are on 4.2:
     appx -recover FILE.bad.dat FILE.dat > FILE.corrupted
Bruce Johnston Jan 14, '04:
FOR THOSE OF YOU ON WINDOWS, here're a few .bat files in a .zip to help you do this.

They assume that they and the recover.exe reside in D:\APPX\Tools and that D:\APPX\Data is the starting path for your .dat and .key files so you might have to edit them accordlingly.

One is standalone, rebuild.bat, and it is for doing the cmd line work on a single file.  It's syntax is
rebuild ver_or_db app datfilenamewithoutextension
No wildcards allowed for this one!

The other four are invoked in sequence rebuild1 ... rebuild4 with rebuild1 being invoked from the command line as
rebuild1 ver_or_db app datfilenamewithoutextension
but, in this case, all three parameters may contain dos wildcards
e.g.
rebuild1 01 T?? *
should run recover against all of the design files in the T-Apps, assuming no file locking issues.
and,
rebuild1 TST * C* would run recover againt the TST database for all files starting with "C" across all applications.

Hopefully, I'll be able to upload rebuildx.zip after completely this...
That didn't work so here're the .bat files:

rebuild.bat
>>>
d:
cd\appx\data
cd %1
cd %2
cd data
del %3.ke0
ren %3.key %3.ke0
del %3.bad
del %3.del
del %3.da0
ren %3.dat %3.da0
d:\appx\tools\recover %3.da0 %3.dat %3.del > %3.bad
cd..
cd..
cd..
<<<

rebuild1.bat
>>>
REM     Receives %1 = 1st level below APPXPATH (Company or Version)
REM              %2 = 2nd level below APPXPATH (Application)
REM              %3 = .DAT file spec WITHOUT the extension
REM     Wildcards permitted at all three levels.

REM     Note that the "/d" usage in the FOR commands implies
REM     that the command line extensions in NT server be supported.

REM     Make sure that "d:\appx\tools" is changed to match the path to
REM     you own .BAT files. AND RECOVER.exe

d:
cd\appx\data
for /d %%d IN (%1) DO CALL d:\appx\tools\REBUILD2.bat d:\appx\tools %%d %2 %3
<<<

rebuild2.bat
>>>
REM     REBUILD2.bat
REM     CALLed from REBUILD1.bat 
REM     Receives %1 = path to REBUILD3.bat and RECOVER.exe
REM              %2 = 1st level below APPXPATH (Company or Version)
REM              %3 = 2nd level below APPXPATH (Application)
REM              %4 = .DAT file spec WITHOUT the extension

cd .\%2
for /d %%d in (%3) do CALL %%1\REBUILD3.bat %1 %%d %4
REM     In case dir structure or wildcards were wacky
REM     Force a return to where REBUILD1.bat started us.
cd\appx\data
<<<

rebuild3.bat
>>>
REM     REBUILD3.bat
REM     CALLed from REBUILD2.bat 
REM     Receives %1 = path to REBUILD4.bat and RECOVER.exe
REM              %2 = 2nd level below APPXPATH (Application)
REM              %3 = .DAT file spec WITHOUT the extension
if not exist .\%2\data goto ENDBATCH

cd .\%2
cd .\data
for %%f in (%3.dat) do CALL %1\REBUILD4.bat %1 %%f
cd..
cd..

:ENDBATCH
<<<

rebuild4.bat
>>>
REM     REBUILD4.bat
REM     CALLed from REBUILD3.bat 
REM     Receives %1 = path to RECOVER.exe
REM              %2 = .DAT file spec
REM
REM     Thanks to Paul Peterson, (ppeterson@utahsbr.edu)
REM     I know that %~n2 returns the file name
REM     WITHOUT the extension which he learned from
REM     (the?) Windows NT Scripting Administrators Guide
REm             by William R. Stanek

del %~n2.ke0
ren %~n2.key %~n2.ke0
del %~n2.bad
del %~n2.del
del %~n2.da0
ren %~n2.dat %~n2.da0
%1\recover %~n2.da0 %~n2.dat %~n2.del > %~n2.bad
<<<
 a) Verify that the files are indeed OK.

 b) cd $APPXPATH/vv/aaa/ where "vv" is the version and
      "aaa" is the application id

 c) chmod 777 Resource

 d) chmod 777 Resource/*

 e) rm -f Em/*

 f) rm -f Dbg/*

 g) On your local client system, remove the ".appx" cache directory
    tree.  It should be in your home directory.  If you look at 
    home\appxlog.txt file in your client directory, it will 
    have a log line with the path to your local .appx cache directory.

 h) Check that the case of the filenames in the resource file exactly
    match the case of the filenames on disk.

 i) Try to load the screen again in the image editor

 j) If it still fails, email home\appxlog.txt as it exists 
    right after killing the client from a lockup, to techsupp@appx.com.

  If OS is set, then your operating system is NT, 2000, or XP

  If OS is not set then your operating system is 95/98/ME

  If OS=Windows_NT, and the variable "ProgramFiles" is set then you have
  an operating system that is 2000 or XP, else your operating system is   
  NT.
     Seq No  Field Name          Field Type Format
     ======= =================== ========== ======================
     100     VERSION SIZE        DOMAIN     S9(10)
     200     VERSION MAJOR       DOMAIN     S9(10)
     300     VERSION MINOR       DOMAIN     S9(10)
     400     VERSION BUILD       DOMAIN     S9(10)
     500     VERSION PLATFORM    DOMAIN     S9(10)
     600     VERSION CSD         ALPHA      X(128)
      SET      --- TEMP 80                    =      kernel32.dll,
      APPEND   --- TEMP 80                    0      GetVersionExA
      *
      SET      1EX VERSION SIZE               =      148
      PASS     1EX VERSION                    RECORD           SHARE? Y
      CALL     --- TEMP 80                    RESIDENT? N  END? N  FAIL 0
> # which appxps
/bin/appxps

> # cat /bin/appxps
#!/bin/sh
ps -axwo ruser,pid,ppid,c,stime,tty,time,args

> # appxps|grep appx
root     1573     1 0 Jan11 ?     00:00:00 /appx/tools/appxd -s=8066
root     1576     1 0 Jan11 ?     00:00:00 /appx/tools/appxd -s=8090
appx     1589     1 0 Jan11 ?     00:00:00 /appx/appx -l=/appx/license.log
nobody   1660  1573 0 Jan11 ?     00:00:00 /appx/data/../appx -f=4 -fms_server
al      11565  1576 0 Jan12 ?     00:00:01 /appx/data/../appx -f=4 -d=   
brenda  22224  1576 0 Jan13 ?     00:00:01 /appx/data/../appx -f=4 -d=   
al      24736  1576 0 Jan13 ?     00:00:10 /appx/data/../appx -f=4 -d=   
al       4786  1576 0 10:22 ?     00:00:04 /appx/data/../appx -f=4 -d=   
pat      6181  6180 0 12:42 pts/1 00:00:00 /bin/bash  /home/pat/.appx_profile
pat      6183  6181 0 12:42 pts/1 00:00:01 /appx/appx -d=CTS
root     6574  6522 0 13:17 pts/0 00:00:00 /bin/sh  /bin/appxps
root     6575  6522 0 13:17 pts/0 00:00:00 grep appx

> # 
          appx -l=logfile
               APPX_LS_DBG=0xFFFFFFFF
               APPX_LS_LOG={some.file.location.and.name}
 ftp://customerftp.appx.com/hostinfo/
          cd c:\appx\data\00\TAP\Portdata
          FOR %%f IN (*.) DO c:\appx\tools\appxutil -i=%%f 

   The name of an executable module is derived from the date that the PROCESS                                                                             
   record was added for that process.  If you happen to get two processes with 
   the same "date added", they'll have the same EM name and nasty things will 
   happen.  "Reset Process Date Add"  assigns a unique "date added" to every 
   process record.
   This resets the Date Compiled fields for all data Field and Domain  
   definitions in an application to NULL and also resets the internal Appx Area 
   Number for each field to zero.

   Resetting the compile date to null forces all fields and domains to 
   recompile.

   The Appx Area Number is how Appx locates data in memory.  We have had bugs in
   the past that caused two or more fields to have the same area number meaning 
   they shared the same memory space, which is a very bad thing.  Setting the 
   area number to zero for all fields forces appx to assign all new area numbers
   and thus fix the memory corruption problem.
   It deletes and then re-creates the given files.  
   When the data dictionary compiler compiles a dictionary, it writes records 
   into those four files.  When the process compiler compiles a process, it 
   reads from those files, not the FIELD and RECORD file.  So, initializing 
   those files is basically the same thing as deleting your data dictionary and 
   typing it all back in again - you end up with an uncompiled dictionary.

   Analyzes the source form of the data dictionary for errors and 
   inconsistencies and produces a more directly executable form (the compiled 
   form is stored in BITMAP, ELEMENT, FILE, and INITIAL).

   Restructure changes existing data records into a new form after you've 
   changed (and compiled) a data dictionary.

     The appx.stk dump, 
     The version# of APPX under which you are running, 
     The release level of your Operating System, 
     A brief description of what the error looks like from the operator's
     perspective, and a brief description of the Application Design elementes 
     in which the error occurred. 
     LogonUser() failed with 9999
 /usr/local/appx/appx -recover -c INDEXED.dat CONSEC.dat INDEXED.deleted
First, before you run asql.exe, set the expert mode environment variable as follows:

ASQL_EXPERT_MODE=1

1) Run asql and select the file that you wish to import the schema for.
2) Click on the "Show Script" button. The "create table" script should be displayed in the lower window.
3) Remove all of the comment lines from the script
4) Change field names to suit your purposes
5) Remove the last ";" from the very end of the script
6) Click on the "Try Again" button.
7) If "OK" is displayed, then everything probably worked as expected.
8) If there is an error, drop the table and try again.
9) WinSQL can be used to verify that the table, field names, and indexes were created as expected.

            APPX_LABEL_MEMORY=1

   Application Design > Utilities > Toolbox > Conversion 
   Utilities > Check for Duplicate ANO and EMs.
 1) Run:  
      Conversion Utilities > Reset Process Date Add Fields
      Conversion Utilities > Reset DD for full process

 3) Go into System Admin > Source File Management for your 
    application, initialize the 4 DD object files (BITMAP, 
    ELEMENT, FILE, INITIAL), and recompile the Data 
    Dictionary.

 4) Restructure your enduser datafiles.  
    (This shouldn't take long.)

 5) Delete your Application's EMs

          export APPX_TRAILER_COUNT=40

    http://msdn.microsoft.com/library/default.asp?url=/library/en-us/architec/8_ar_ts_8dbn.asp 

  $(host)         -  host name
  $(engine)       -  Appx engine version
  $(endian)       -  'intel' or 'risc'
  $(root)         -  $APPXPATH
  $(uid)          -  3-character Appx User ID
  $(user)         -  OS User ID
  $(pid)          -  OS Process ID
  $(ap)           -  File's application ID
  $(ver)          -  File's application Version
  $(db)           -  File's Database ID
  $(scope)        -  Process nesting level
  $(file)         -  8-character file name
  $(type)         -  "Data", "Struct", "Portdata", etc.
  $(anything-else)-  An environment variable

        $(root)/$(db)/$(ap)/$(type)/$(file)

        $(root)/$(ap)/$(db)/$(type)/$(file)

        /tmp/$(user)/$(type)/$(file)

 jdbc:odbc:appxodbc
 jdbc:odbc:appxodbc;UID=appx;PWD=password

        if the field is a key segment , Appx builds a CHAR(n) column
        if the field is less than 6 character long, Appx builds a CHAR(n) 
        otherwise, Appx builds a VARCHAR(n) column
        for CCYYMMDDhhmmss key segments, Appx builds a VARCHAR(14) column
        for CCYYMMDDhhmmss non-key fields, Appx builds a DATE column
        for all other date components,  Appx builds a VARCHAR(n) column
        if field length = 1, Appx builds a NUMBER( 3, 0 ) column
        if field length = 2, Appx builds a NUMBER( 5, 0 ) column
        if field length = 4, Appx builds a NUMBER( 10, 0 ) column
        Appx builds a NUMERIC( x, y ) column 

        http://support.microsoft.com/?kbid=899976

        http://troels.arvin.dk/db/rdbms

        http://troels.arvin.dk/db/rdbms/#cli
These should appear in 

HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\<service name>\Config

TCPEnableKeepAlive Y

TCPKeepIdle 300

TCPKeepInterval 60

TCPKeepCount 8

Tthe values shown are the default values, those values are installed when you create a new service 
and they are also the values that we use if the entries are missing from the registry.

The AppxDSvc installer only writes a Config entry if it differs from the default value.  
That means that the installer won't write out entries for the TCPKeepXXX values, 
but AppxDSvc will pretend like those values are present (and set to their default values).

You can add the entries manually if, for some reason, you want to change the defaults.
Setting                 Argument        Default What it does
----------------------- --------------- ------- -----------------------------------------------
TCPEnableKeepAlive      true or false   true    Enable dead connection detection
TCPKeepIdle             (seconds)       300     Idle time before starting to send pings
TCPKeepInterval         (seconds)       60      Interval between keepalive pings
TCPKeepCount            (count)         8       Count of keepalive ping attempts before closing

The time before a server session closes would be the combination of all of these.  
Here is what happens.

After 300 seconds of idle time the server TCP stack will start trying to ping the client TCP stack.  
The server will make 8 attempts to ping the client and get a response before giving up and closing.  
It will wait for 60 seconds after each ping attempt for a reply before sending the next ping.

So in this case it will be 300 seconds + ( 8 * 60 seconds ) = 780 seconds = 13 minutes 
before the server session goes away.
appxd has environment variables for these settings as follows:

APPXD_DISABLE_KEEPALIVE set to anything will turn off the keepalive code.

APPXD_KEEPIDLE     = TCPKeepIdle
APPXD_KEEPINTERVAL = TCPKeepInterval
APPXD_KEEPCOUNT    = TCPKeepCount
        1) the windows event log
        2) a pipe
        3) a gzipped flat ascii file
        4) an AppxIO data file
        5) stdout
        6) a flat ascii file

         =>

        /usr/lib/libc.2 =>      /usr/lib/libc.2

        /usr/lib/libdld.2 =>    /usr/lib/libdld.2

        /usr/lib/libc.2 =>      /usr/lib/libc.2

        /usr/lib/libm.2 =>      /usr/lib/libm.2

        /usr/lib/libnss_dns.1 =>        /usr/lib/libnss_dns.1

        /usr/lib/libdld.2 =>    /usr/lib/libdld.2

        /usr/lib/libc.2 =>      /usr/lib/libc.2

        /usr/lib/libpthread.1 =>        /usr/lib/libpthread.1

        /usr/lib/librt.2 =>     /usr/lib/librt.2

        /usr/lib/libcl.2 =>     /usr/lib/libcl.2

        /usr/lib/libisamstub.1 =>       /usr/lib/libisamstub.1

        /usr/lib/libdld.2 =>    /usr/lib/libdld.2

     For details on reading the log, see http://board.appx.com/faq/cache/633.html 

     For details on setting the variable to control the contents of the log,  see http://board.appx.com/faq/cache/634.html
 00 - 0000072.0797031 0000000.0057123 05000
 01 - 0000076.0737460 0000000.0043817 10001
 02 - 0000000.0000000 0000000.0000000 00000
 03 - 0000000.0000000 0000000.0000000 00000
 04 - 0000000.0022587 0000000.0022573 00002
 05 - 0000000.0000000 0000000.0000000 00000
 06 - 0000026.0281965 0000000.0031189 05001
 07 - 0000002.0012514 0000000.0012999 15003
 08 - 0000000.0000000 0000000.0000000 00000
 09 - 0000000.0000000 0000000.0000000 00000  
 10 - 0000000.0017440 0000000.0017440 00001
 11 - 0000024.0951585 0000000.0022819 05000
        operation type
        cumulative time spent performing that operation for that table
        longest operation of that type for that table
        number of operations of that type for that table
        00 - Read operations
        01 - Readnext operations
        02 - Write operations
        03 - Rewrite operations
        04 - Open operations
        05 - Reread operations
        06 - Search operations
        07 - Release operations
        08 - Search by primary key operations
        09 - Search by alternate key(s) operations
        10 - Position to start of table operations
        11 - Search by row identifier operations
         00 - 0000072.0797031 0000000.0057123 05000
  As mentioned above, a Read is performed whenever Appx needs to retrieve
  a record by a specific key value.  A Read is accomplished by performing a
  Search followed by a Readnext, so you should have at least as many Searches
  and Readnexts as you have Reads.
  A Readnext is performed whenever Appx has established a Search position
  within a file and then needs to retrieve the record following that position.
  A Readnext will be performed for READNEXT, BEG/END READ, and READ statements
  as well as for most PCF operations.
  A Write operation is equivalent to a WRITE statement - it causes a row to be
  INSERTed into a table.
  A Rewrite operation is equivalent to a REWRITE statement - is causes an SQL
  UPDATE statement to be executed.
  This entry reflects the amount of time required to open a table.  It includes
  querying the RDBMS data dictionary and performing all internal bookkeeping.
  The time required for this operation is usually negligible.
  Since Sybase does not perform row-level locking, Appx must simulate locking
  outside of the Sybase server. The algorithm used to simulate locking requires
  reading every locked record twice: once to retrieve the key values and once
  to read the rest of the record.  A Reread operation reflects the second read.
  There should be one Reread operation for each record held.
  A Search operation is performed whenever Appx needs to position to a specific
  key value.  Searches are usually (but not always) performed by a Readnext
  operation and are usually part of a BEG AT or READ operation.  Search
  operations are further divided into "Search by primary key", "Search by
  alternate key(s)", "Position to start of table", and "Search by row dentifier
  operations".  In the above example, you can see that we performed 5001
  searches: 1 position to start-of-table and 5000 read by row-id's.
  A Release operation is performed whenever Appx releases a hold on a locked
  record.  Appx will try to release holds on records which were not locked -
  the overhead is negligible and does not cause any i/o.
  This a subdivision of the Search operation.  It reflects the number of
  Search operations which specify a primary key value.  In Sybase - primary
  keys are implemented as a clustered index.
  This a subdivision of the Search operation.  It reflects the number of
  Search operations which specify an alternate key value.

  This a subdivision of the Search operation.  It reflects the number of
  Search operations which position to the first row in an index (as opposed
  to a specific key value).
  This is a subdivision of the Search operation.  Whenever Appx reads a
  row from a table, it remembers the row-identifier (in the case of Sybase,
  the row-id is the primary key).  If Appx needs to reposition to that row,
  it will use the row-id.  An Appx query process works by recording the 
  row-id's of all records which meet the query constraints and then re-reading
  the results using the row-id's.   The above example was produced by selecting
  5000 out of 5000 records in a query process.
        1)  The amount of time spent on 00 (Read), 01 (Readnext), 02 (Write),
            03 (Rewrite), 05 (Reread), and 06 (Search) operations.
        2)  The number of operations performed against each file compared to
            the number of operations performed against the PCF.
The APPX_DBG_CODE environment variable controls the amount of information written to the APPX_SQL_CMD log file.  APPX_DBG_CODE is a 32-bit bit mask - if a bit is on, the corresponding information will be logged.

Here are the current (01/22/98) bit definitions:

0x00000001 -    Log all generated SQL commands

0x00000002 -    Log status codes returned from server (*)

0x00000004 -    Log all fetch attempts (*)

0x00000008 -    Log all engine level operations (*)

0x00000010 -    Log all connection-related operations (*)

0x00000020 -    Log binding summary (*)

0x00000040 -    Log binding details (*)

0x00000080 -    Log column/field mappings

0x00000100 -    Enable server logging mechanism (Oracle only)

0x00000200 -    Enable result (null/non-null) logging (*)

0x00000400 -    Log negotiations for server options (such as packet size)

0x00000800 -    Log all records returned from server

0x00001000 -    Currently undefined

0x00002000 -    Log proxy locking operations (Sybase/ODBC only)

0x00004000 -    Time individual operations (use 0x10000 instead)

0x00008000 -    Enable server-side statistics (if capable)

0x00010000 -    Gather per-table timing statistics

0x00020000 -    Enable "show plan" feature on server, generates voluminous logs

0x08000000 -    Write server-based log to "/tmp/syblog" (Sybase only)

0xFFFFFFFF -    Logging all activity

Definitions marked with (*) are for internal debugging purposes only - they 
provide very little useful information.
Information regarding Language Translation and Language Translation Manual is 
available on our website:

http://www.appx.com/assets/asp/dynamic_generator.asp?pageid=558

Previous:	APPX Runtime Environment
Next:	APPX Installation & Configuration