Hi Guys,

I am getting the following errors when some sessions are being executed...

I shows it as error-254 from that BW message.

From the help manual -

254 EAUDABORT This indicates that a commit transaction has failed in the audit server. See the log.audit file for more information.

I couldnt find any info from the log.audit file. I have also checked the $BSE/audit permissions, all the directories and subdirectories have the 777 permissions.

Searching on the forum I found one posting where someone advised to remove the specific audit file maybe if its corrupted. I tried that too but it still gives the same fatal error.

Any clues ??? :confused:

Thanks!

Are the tables on which you have the error, also been specified as that they should be auditted?

Yes Hans, the tables are having the entries with "Y" in the tabedef file as well as checked in the audit sessions...

I couldnt get a clue to fix it yet...

:confused:

It seems that there is an error in the retry and commit mechanism... I am not really sure on how to solve, but I think that something like this is happening.

There is a transaction pending on the audited table. The transaction must be audited, so it is also written into the audit files. But if the pending transaction is aborted (due to any DB error: locking, db outage etc.) the audit record should not be written.

Now I expect that the writing of the audit record itself is based upon the correct commit of the first transaction. So the writing of the audit is a kind of post-commit commit. But if the second commit fails... then the system should jump to its retry point, which then could be: the db.retry on the first transaction.

But, as lined out... it is only a clue. What happens if you run your session with "-- -set TEST_RETRY=2" which forces the db.retry mechanism to be triggered so the failing transaction is forced.

If the audited session then runs fine, the above story is crap :(

Hi,
What is exactly the message called in the log.audit file??
GN

I am still yet to fix this mysterious problem... :mad:

Omeluuk, we tested with ur idea "-- -set TEST_RETRY=2" and it still gave the same error.

GN, the problem is that I couldnt find any info in the log.audit file

The only information I could get from the audit sessions are shows in the document.

Are you on Baan ERP 5.0c/7.1b tools or newer versions? Audit changed greatly at that level and my answer will be dependent on what version you are on...

-- Jim

James,

We are on the latest BaaN version - 5.2 the iBaaN series.

Check the permissions on the file...are they ok..

I had a similar problem , i had to delete the file in the $BSE/audit/ccc/atccom/atccom000670.000 and
$BSE/audit/ccc/atccom/atccom000670.inf
(u can replace with your table!!)

and the make changes to the table through Baan, the problem will be solved and the two files u have just deleted will be recreated!!!

The file permissions are OK...

I have also checked the $BSE/audit permissions, all the directories and subdirectories have the 777 permissions.

I was hesitant to delete the files, so I have moved them to a temporary directory and the developer executed the sessions.
We got 2 new *.000 and *.inf files, but got the fatal BW error message again...


>pwd
/app/baan/sy16/bse/audit/055/awhinp
>ll
total 8
drwxrwxrwx 3 bsp bsp 96 Jun 4 15:44 .
drwxrwxrwx 115 bsp bsp 2048 Jan 11 08:44 ..
-rw-rw-rw- 1 svpat bsp 813 Jun 5 14:04 ainp100055.000
-rw-rw-rw- 1 svpat bsp 274 Jun 5 14:04 ainp100055.inf
drwxrwx--x 2 nprao bsp 96 Jun 4 15:43 bak


Do you have any idea what does the Unterminated status on the Display audit sequence means?

:confused:

Hi,

I mean the log.audit - file in the $BSE/log directory.
Or is there any other log......-file in that directory where you can see any error-message?

The UNTERMINATED status means that the sequence file untill now not has been terminated by you or anybody.
in BaanIVc4 it is session: ttaad4160m000 where you can terminate the current sequence file.

GN

Hi all!
Did someone solve this problem? I am facing similar problem in Baan 5.0c. When we started session, session immediatelly terminated and we got error 254. The first message in log.bshell was: "Insert into ttaud110 compnr 000 failed".
Best regards,
jc

The fix is not yet available in the porting sets. Here is more info -

the problem is in the audit_srv6.2 or bshell6.2. (portingset problem).
The problem occurs when doing a db.clear.table with NO.ROLLBACK.
So, probably the other sessions do also such db.clear.table.
For session whinp1200m000:

in the source code pwhinp12000 the following function causes the error:

function clear.inp100.B()
{
commit.transaction() |* Leave it this way.
db.retry.point() |* It's the only way in which
db.clear.table(twhinp100, NO.ROLLBACK) |* it is working correctly.
commit.transaction() |* Else error 203.
}

when changing this to:

function clear.inp100.B()
{
commit.transaction() |* Leave it this way.
db.retry.point() |* It's the only way in which
db.clear.table(twhinp100, WITH.ROLLBACK) |* it is working correctly.
commit.transaction() |* Else error 203.
}

problem does not occur, so this would be a workaround.

For the other sessions, a reproduction scenario is needed as well. When reproducing the error with the other sessions, you can enable dbgflow tracing to find out the function from which the db.clear.table is called.

There are 2 workarounds in the meantime:

1.Change the db.clear.table function, from "db.clear.table(<table>, NO.ROLLBACK)" to "db.clear.table(<table>, WITH.ROLLBACK)". However, this change will cause trouble with the new portingset and the fix for this problem. So if you change these functions, you will need to change them back before upgrading the portingset.

2. Set "Audit" to "No" for these tables, however, you can also have some problems with some sessions, specially if they are "parameters" sessions.

For BaanERP5.0/5.1 the fix should be in porting set 7.1c.05.
For BaanERP5.2 the fix will become available in porting set 7.3a.02