Baan: IVc4
UNIX: HPUX11
Oracle: 8.0.6

Can you help?

I've just gone through an install of Baan on a test server. I have done the following:-

1. Loaded Baan software
2. Loaded Service packs 3 and 4
3. Upgraded to porting set 6.1c.05.01

I have now noticed that the following tables have error 512:-

ttadv304
ttadv335
ttadv364
ttiex188
tttss010

When you look at these tables using the database tools you can see that columns over 255 in length have been split. i.e. In table ttadv304 you see columns t$layo#1 (255) and t$layo#2 (5)

Two questions:-

1. Which process caused this to happen. I know that you encounter this issue when upgrading from Oracle 7 to Oracle 8 but I already had 8 installed.

2. What's the resolution to this problem?

Thanks in advance

PJ

Insert this line in $BSE/lib/defaults/db_resource file:

ora_column_format:7

I think it will help.

PJ,

Did you install using the ora7 portingset database driver? If so then I think this would explain why your table fields got chopped up.

The way to fix it would be to dump the tables using ORA_COLUMN_FORMAT=7 and to re-import them using ORA_COLUMN_FORMAT=8

Hi Pat,

I pretty sure that we selected the Oracle 8 driver when going through the Baan install script.

How can I tell which Oracle driver is being used now?

Where in the file system can the Oracle driver be found?

As I'm not sure when the corruption of these tables actually happened I can't put my finger on the process that caused the issue. I think it was when I upgraded the porting set to 6.1c.05.01. Maybe this port set replaced the Ora8 driver with the Ora7 one?? we had a similar incident on our production server two years ago. This happened whilst I was very new to Baan so I'm having trouble recalling what was done to resolve the issue then.

Just to confirm you fix Pat.... I need to enter ORA_COLUMN_FORMAT=7 into the db_resource file dump the tables, enter ORA_COLUMN_FORMAT=8 into db_resource and then create table from seq dump.

Cheers

PJ

PJ,

Your last comment is correct.

As to db drivers: that you can tell from the $BSE/lib/tabledef6.1 files. The drivers themselves are located in $BSE/bin as ora7_srv6.1 and ora8_srv6.1

I am not sure why this happened but possibly your install (master) tape didn't really contain an oracle8 driver and used an oracle7 one. You would need a pretty new master tape for that.

Pat,

If if look at the tabledef6.1 file for both the original porting set and the updated porting set it has the following entry:-

*:*:oracle8(ORACLE_HOME=/oracle/product/8.0.6,ORACLE_SID=baan):N

The ora8_srv6.1 file was also available under the original porting set.

The master tape is IVc4, and about two years old. Our production install was on Oracle 805.

Another Baan mystery.........

Thanks

PJ

PJ,

Looks like that indeed. The only other thing I can think of is that the ORA_COLUMN_FORMAT=8 needed to be set manually at install time.

Hi,

Could somebody please confirm that the db_resource file is read at the time of client login and that I don't need to rc.stop and rc.start for new changes to take effect.

Pat: You suggested that I should dump the tables and re-import. When I attempt to do this I just get a 512 read error. Any other suggestions?

Thanks

PJ

Hi pjohns,

db_resource is read on an ad-hoc basis - you don't need an rc.stop or rc.start.

Remember, you need to set ORA_COLUMN_FORMAT=7 while exporting the table, then set ORA_COLUMN_FORMAT=8 while doing the import. (Note also this is a Unix variable, not the setting in the db_resource file)

Like Patricks comment, I think the only way you've ended up with the hash columns is due to the install running via an Oracle7 database driver.

Thanks James,

Can you give any additional advice on how I can get a dump of this table. If I attempt to take a dump I get a 512 read error.

Thanks

PJ

If you've exported ORA_COLUMN_FORMAT=7 before bdbpre, then you should not get error 512. Can you confirm this?

James,

I initially mis-understood where ORA_COLUMN_FORMAT=7 was entered and I was entering it into db_resource. :confused:

I have now successfully ran a bdbpre for one of my corrupted tables.

When I bdbpost should I drop the table first? If so should I do this from the Oracle end or use the -k option in bdbpost?

When I look at the dump the data looks a bit strange....Is this correct??

tt|aad|0100m0001|2|B40|c4||1|^C^BNM-^KM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^H
M-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^H
M-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^H
M-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^H
M-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^H^E-^AM-^K:^F-^AM-^K:^G-^AM-^K:M-^?|
tt|aad|0150m0001|2|B40|c4||1|^C^BNM-^KM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^H
M-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^H
M-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^H
M-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^H
M-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^HM-^?|

Regards

Philip

Hi Philip,

Good to hear it is now working.

Makes no difference where you drop the table - using -k is fine.

Regarding the dump data - I've no idea, you'll just have to try it I guess.

James,

Okay I got our DBA to drop the table form the Oracle end when I perfom a bdbpost I get the following error:-

Can't create table (error 2017)

I also got this error when using the -k option in bdbpost.

When I look at the log.oracle8 I see the following:-

2002-11-04[13:59:29]:E:bsp: ******* S T A R T of Error message *******
2002-11-04[13:59:29]:E:bsp: Log message called from /port.6.1c.05.01/vobs/tt/ser
vers/ORACLE_2/ora_native.c: £1867 keyword: Oracle Error
2002-11-04[13:59:29]:E:bsp: Pid 631 Uid 125 Euid 125 Gid 125 Egid 125
2002-11-04[13:59:29]:E:bsp: user_type S language 2 user_name bsp tty ote locale
ISO88591/NULL
2002-11-04[13:59:29]:E:bsp: Errno 25 (Not a typewriter) bdb_errno 0
2002-11-04[13:59:29]:E:bsp:
2002-11-04[13:59:29]:E:bsp: dbs_errno = 0Error 2017 occurred:
2002-11-04[13:59:29]:E:bsp: Error ORA-1017 occurred during logon.
2002-11-04[13:59:29]:E:bsp: ORA-01017: invalid username/password; logon denied
2002-11-04[13:59:29]:E:bsp:
2002-11-04[13:59:29]:E:bsp: Error BDB-2017 returned.
2002-11-04[13:59:29]:E:bsp: Logon failed; errno 2017
2002-11-04[13:59:29]:E:bsp: Flushed at /port.6.1c.05.01/vobs/tt/servers/ORACLE_2
/ora_table.c : £706.
2002-11-04[13:59:29]:E:bsp: ******* E N D of Error message *******

The problem appears to be the linking between the baan user and Oracle. I've tried to run ttdba0110m000 - Maintain Links between Baan users and Oracle users. When I try to run this session I get a 'sh: ttdba0110m000 not found' Any ideas?

I should say that this is a frsh Baan IVc4 install with only the 000 company running.

Thanks

PJ

ORA-01017 - Looks like you've got an incorrect password or authentication problem somewhere. Either the bsp user, or the baan user.