We have three Baan servers. On one server the ttiex1285 session is very slow abd I can't figure out why. The log file shows that it requires 3 minutes to do a 'Repair cpac vrc structure'. What does that mean ?

Part of the log.ttiex1285 logfile :
Installation on date: 100406113119
Install Data Dictionary ${BSE_TMP}/tmpsyncsoft 100406113119
Repair cpac vrc structure 100406113421
Install Data Dictionary forms ${BSE_TMP}/tmpsyncsoft 100406113424
Install Data Dictionary reports ${BSE_TMP}/tmpsyncsoft 100406113424
Install Data Dictionary scripts ${BSE_TMP}/tmpsyncsoft 100406113424
Install Data Dictionary libraries ${BSE_TMP}/tmpsyncsoft 100406113424
Install Runtime Data Dictionary ${BSE_TMP}/tmpsyncsoft 100406113424
Install objects ${BSE_TMP}/tmpsyncsoft 100406113424

Any ideas ?

Walter

Hi,

I have found the following thread:
http://www.baanboard.com/baanboard/showthread.php?t=11620

I hope it helps,

- George

We don't use pmc.
Also the problem server is a recent copy of our production server. We do this every now and then to get more recent data on the second server for testing.
I have done all this several times before but this is the first time we have this problem. As usual I copied company 000 plus our production companies and several directories and runtime files from $BSE.
I'd appreciate it if someone could tell me what the session does when it is
'repairing the cpac vrc structure' ?

Are the object ttiex1285 on the several servers the same?
The same Portingset? The other TOOLS-objects the same?

The second server has always been a copy of the first. So everything is the same (or it should be). I checksummed the ttiex1285 object and they are the same on both servers. Portingset, Oracle, tools etc are all the same.

10G does install with a standard job to analyze the tables. I have noticed that in our environment I have to disable this job and to delete the statistics of sys.cdef$ sys.con$ and sys.user$

That solved my problem with installing patches (with pmc).

I am having this same problem. The repair cpac vrc structure is taking 15-20 minutes. We also have a test server that we refreshed with production data prior to this happening. I think it also started around the time we installed service packs 19 & 20. Any help on this would be much appreciated.

Hi Steve,

I had one of our consultants look at it and he solved this for us. It has to do with a known problem in Oracle 10 where something goes wrong when you have the statistics switched on for the sys.cdef$ sys.con$ and sys.user$ tables.
When inserting records in one of these tables Oracle uses recursive sql which doesn't work well when the statistics are switched on.
It was solved by removing the statistics and locking the job that periodicaly creates/updates the statistics. I'm not an Oracle man but this is more or less the story. The repair session is now back to a few seconds.
I think Han Brinkman also solved this.

Walter

Thank you very much for the quick reply. I saw Han's response, but I thought it might be specific to PMC. I'll ask our Oracle admin to look into this.

Thanks again,
Steve

Can you tell me the detail of solution. How can I imlement this.

Thnaks in advance

Hello sheetu,

Unfortunately I'm no Oracle expert but my post on the 10th of january should be enough for any Oracle dba. Also read Han's post of the 6th of october 2006.

Walter