Dump & Load versus conv910

[microedgenh]

We have historically converted our customers' database from progress 9 to 10 using a conv910 batch file:

@echo off
rem conv910.bat
rem This Batch file converts a Progress 9.x database to Progress 10.1.
rem Input Parameters:
rem 1. Path to 9.x dlc directory.
rem 2. Path and name for Progress database.
rem 3. Path to 10.1 dlc directory.
rem
rem Example d:\upgrade\Conv910.bat d:\dlc9 c:\npo\found\found d:\dlc
rem
rem History:07/25/06:RPS:Created from conv89.bat. Readme# 3692.
rem 07/23/07:RPS:Add enableseq64 and enablelargekeys. Readme# 4112.
rem
if '%3' == '' goto help
@echo on
set dlc=%1
set promsgs=%3\promsgs
"%dlc%\bin\_proutil.exe" %2 -C truncate bi
set dlc=%3
"%dlc%\bin\_proutil.exe" %2 -C conv910
"%dlc%\bin\_proutil.exe" %2 -C enableseq64
"%dlc%\bin\_proutil.exe" %2 -C enablelargekeys
goto exit
:help
echo .
echo Syntax: [path to FIMS]\conv910.bat [Progress 9 DLC] [database] [Progress 10 DLC]
echo Example: d:\npo\found\fims\conv910.bat d:\npo\dlc9 d:\npo\found\found d:\npo\dlc
echo .
:exit

However, while I was at a DBA class I was told by the instructor that this is not the way to upgrade b/c you lose all the benefits of the new database structure and features such as clustering. Now, my programmers are questioning the "new" way I am upgrading; dump and load, because there have been some data issue due to different .df files (the customer has tables, fields, indexes in different orders than my new .df). I need more to convince them dumping and loading is the way to go. The programming manager wants to do this instead:

[FONT='Arial','sans-serif']1. Run [/FONT][FONT='Arial','sans-serif']proutil found –C cnv910 [/FONT]
[FONT='Arial','sans-serif']2. Create incremental df with new db and apply to found.db.[/FONT]
[FONT='Arial','sans-serif']3. Apply new structure.(run prostrct).[/FONT]
[FONT='Arial','sans-serif']4.Run other proutils to enable new features ( including auditing).[/FONT]
[FONT='Arial','sans-serif']proutil found truncate bi[/FONT]
[FONT='Arial','sans-serif'] proutil found -C enableseq64 [/FONT]
[FONT='Arial','sans-serif'] proutil [/FONT][FONT='Arial','sans-serif']found[/FONT][FONT='Arial','sans-serif'] -C enablelargekeys[/FONT]
[FONT='Arial','sans-serif'] proutil found -C enableauditing area "auditarea"[/FONT]

What are your thoughts about this? Thank you, Ann.

 

[TomBascom]

You can only take advantage of many new features of the database, such as type 2 storage areas, by dumping and loading. (Ok, you could do table moves to get there but that's pretty much the same thing and it takes even longer...)

You don't get those features for free though even if you do dump & load -- You also have to actually define storage areas, assign rows per block and cluster size values and so forth...

Are you distributing r-code to your customers?

If so then that is likely the reason that your programmers are complaining. The .df files need to match -- but you should be able to manage that by either using the same .df file for everyone or by using a common empty db.