I did a sequential dump of one of our production companies and for the first time ended up with a dump file > 2GB. In fact, I ended up with 3 files:

CO160.dmp, size 2,097,152KB - The file specified in the bdbpre script
CO160.dmp.1 size 2,097,152KB
CO160.dmp.2 size 133,472KB

If I try to load the data using:

D:\Baan\bin\bdbpost -kmnif -I L:\CO160.DMP -q L:\CO160LOAD.ERR

it only loads the data actually in DO160.dmp. If I try and load the data using:

D:\Baan\bin\bdbpost -kmnif -I L:\CO160.DMP.1 -q L:\CO160LOAD.ERR

I get the message:

Default separator ('\0') taken
Reading dump header failed
BDBPOST: FATAL ERROR: 'Not a valid dump file'

I presume I am missing a flag or something. Any help or suggestions would be appreciated.

Thanks

Stephen

Stephen,

How did you split the files in 3 parts. I think there lies the problem, I can't see how many Baan's bdbpost would be able to 'pick up' where the previous file ended.

Why don't you use a table based export? In that way you will get a separate file for each table and you will have less chance of running into the 2GB limit.

The problem is BDBPRE made the "decision" to create the .1 and .2 files (I did not split them). I started the dump Sunday morning. When I got into the office Monday morning. I had the three files. Since this is the first time I have had a dump file exceed 2GB, I'm not really certain what or how.

To your point, I understand how, and why, to create two different dump files based on two or more different "table lists". But that doesn't solve my immediate problem. Short of doing another dump, how do I get BDBPOST to load the .1 and .2 files that BDBPRE created.

Thanks

Stephen

Hi,

Try to use the parameter -M (capital 'm') also into your command.
(It is possible by the bdbpre command to use -M and then followed by a number (i.e. -M 500M) which will give a file of about 500 Megabites (or 500K = a file of 500 kilobites).

GN

Stephen,

This would be new functionality of bdbpre and bdbpost to me. Does your first file have a 'header' section? I know Baan came up with a new dump format a while ago and possibly introduced the split-feature then as well.

If you were on UNIX you could use a named pipe to pass the files through one by one but on NT....

Hi,

If you do a bdbpre with the -M 500M flag it will creates files of about 500 Megabites. (So not only one file, if you have more data in your company or table))
If i do a dump from our tfgld106 it then creates about 10 files of 500 MB.
As Patvdv already proposed, it is better to do it by table, because if there is a problem with bdbpost, you are able to start with the table which gave an error and it is not necessary to do a bdbpost for the other tables.

GN

So, if you use the -M flag how does bdbpost know to link all the files back together? Is there a flag you need to use on bdbpost? If so, would it work in my situation?

Hi sgabor!

I suppose that you didn't run bdbpre for quite a while - to me it looks as if your dump totals > 4GB
what happens if you simply say "bdbpost6.1 .... < CO160.dmp*"?
By the naming convention, the first file should be used first (the one with the header and no filename suffix), then the one with suffix .1, then .2 ?
Please try and give a feedback

good luck

Norbert

Hi,

According to me is the -M flag a signal in the program, which recognize that there should be more dmp-files then only 1(one).
I use th bdbpost en bdbpre commands on UNIX in a script, because of the 2.14GB limit. (I use it at this moment for 3 TF tables)
But, according to me, if you did the dump by the BAAN-session, it should be that you also can import the dump by the BAAN-session. Isn't it?

GN

Check out Baan case number 107090.

Basically, the file that has been created, should (ideally) not exceed 2Gb. The solution (listed in case number 107090) details how to create multiple files from the same table. NOTE: this is for porting sets 6.1c.05 and up.

To create the files, use a syntax similar to;
bdbpre6.1 -M1G -O/home/bsp/g_l -Ntfgld410 -C100
This will create multiple dump files (from the tfgld410 table of company 100, each dump file being no greater than 1Gb), named;
g_l
g_l.1
g_l.2
....

To import the files into a different company, use a syntax similar to;
bdbpost6.1 -A -M -c900 -I/home/bsp/g_l
where the company you are importing into, is company 900.

Please refer to solution # 107090 for a more detail description.

I usually do:
bdbpre6.1 -I tfgld106 -t"|" -M 1024M -o tfgld106_out -q

then

bdbpost6.1 -n -m -k -M -D ./tfgld106_out -t "|" -q log_tfgld106

The output you would get...
tfgld106_out.S
tfgld106_out.1
tfgld106_out.2
tfgld106_out.3
tfgld106_out.4
etc

Based on the size of the table, you can adjust the -M option to the size you would like.

The first file that is outputted (.S) will have 3 rows of header information. The rest of the files will not.

If for some reason the bdbpost fails, and it has already loaded the first file, you can copy tfgld106_out.1 to tfgld106_out.S, and continue your data load.

If you are back porting the pre files, you can strip out the 3 lines of header information to load into an older porting set.

Hope that helps...

hallo,


I'm having the same problem.
I creating dump with ttaad4226m000.
when the data is exceded 2 Gb it will create another file.
example: com.777 and com.777.1

right now I've try using tdaad4227m000 to create table from dump.
but there only one file is being used while there is 2 dump file.

how can I import the second file

please advice


thanks

correct me if i am wrong but i guess you have to do this on a commandline since the -M parameter can be turn on in the session.

Rgrds,
Han

I think there is a patch to tdaad4227m000 that fixes that.

Dave

I am on BaanV sp17 but that doesn't show an option. Is it hidden? Or does this program check your os? (I have to live with windows).

Hi, long time... no see...

Han Brinkman, I know that ttaad4226m000 by default "calls" bdbpre6.1 with the option "-M 2G".
I even have the impression that also in the command line the bdbpre6.1 has a default setting "-M 2G".

So, for wei_smooth, the files were split due to these default settings.

Basically, BAAN should be able by itself to import from all files it has generated. But there are also bugs, as dave_23 has mentioned.

So you have to patch your SW and install latest version for session ttaad4226m000 and ttaad4227m000 (in fact latest scripts)

But since these sessions use bdbpre/bdbpost you must also upgrade to the latest porting set.

P.S. I remember that there were bugs in PS 6.1c.05.02 when you explicitly mentioned "-M some_size". If you did that, bdbpost was not able to import anymore all generated export files.

Hi
I did the following
1- use bdbpre6.1 to dump a table making the file size no more than 200k
the table was 370k thus I have 2 files for the same table
bdbpre6.1 -Ntfgld410 -t'|' -M200K -C444

2- put the 2 files in the same directory and used bdbpost6.1

bdbpost6.1 -t'|' -R -D/dump/san/bk380/test -C444

it worked just fine, try it with your files may be it will work