Migration from BS2000 to Linux

[magsbg]

Because of the different sorting order, you also have to change starting-and-ending-values.
For example:
On BS2000 you wrote READ … = ‘A’ THRU ‘9999’
On Linux you have to write READ … = ‘0’ THRU ‘zzzz’

[Ralph_Zbrog]

@Sascha Wiegandt

It was also interesting for me because i got from this discussion the ENDIAN Natural Parameter which will solve the default Problem. But i’m not sure if i better write Programs which converts the Workfiles or just use the ENDIAN Parameter.

Rather than worry whether the soure and target systems have the same endian mode, move the Integer value to a temporary Binary variable, and write the Binary to the WORK file. The target system will need to read the Binary value and move it to an Integer variable. In this way the target system will convert the binary value to the appropriate Integer endian format. No need to write a conversion program; the conversion is done as needed, on the fly, by the writing and reading programs.

[Matthias1 (Matthias)]

@Ralph Zbrog

move the Integer value to a temporary Binary variable, and write the Binary to the WORK file

You’re right! That would be an easy way.

Meanwhile I found another way. You can set the workfile-Type to “Portable”. The Natural documentation says:

PORTABLE performs an automatic endian conversion of a work file when it is transferred to a different machine. For example, a work file written on a PC (little endian) can be read correctly on an RS6000 or HP machine (big endian). The endian conversion applies only to field formats I2, I4, F4 and F8.

[system]

I tried to write down the differences i have found depending the Development side (Not Administrative) at first Post. If someone finds an Error or please write it down there so i could correct it.

Greetings
Sascha Wiegandt

[system]

Hello There,

i have a new question about Workfiles. If i write a Workfile at BS2000 which would have the following fields in it:

[code]

01 WF-STRUCTURE
02 #FIELD1 (B4) INIT <8256>
02 #FIELD2 (N9.2) INIT <123456789.12>
02 #FIELD3 (I4) INIT <-65535>
02 #FIELD4 (P12.2) INIT <-234567890123.12>
02 #FIELD5 (A32) INIT <’

[Wilfried_B]

You have to convert from EBCDIC to ASCII. But you won’t be lucky in converting the file in one piece. The reason is:

1. You have binary fields
2. you have packed numbers
3. you have signed integers

So the only thing is to convert the file on field basis (EXAMINE TRANSLATE USING). Packed and unpacked numbers dont need to be changed, the binary and interger values may need to be changed (whether low and high byte order is swapped or not).

[system]

I hoped that i could use a SAG-Tool which is shipped with Natural or so. Like using the Natural ENDIAN Parameter.

If not i have to find all WRITE Work’s at all Programs we have then i have to look at the Parameters written with all redefinitions and then i could write a program which converts the ENDIANs at all binary Types, convert Text with using USR0620N and then i could hope that i have 99% of all Data convertet. Not included those bad Programms which use an redefinition at an other sub/program level. (Alphafields with Binary redefiniton and so on - very bad old stuff)

But thanks anyway

[fibag]

Hi All,

I’m new to this forum. I’ve read all posts on the subject until now and it’s valuable info.

2 additional questions however:

• how would you size an open system (Windows or Linux) to provide the same perf. as the BS2000 you’re migrating from :?
• how do you estimate the migration effort ? :? :?

Thanks in advance for any hint.

fibag