Starting last Thursday this DDC function server started having performance problems. This session is used with our Autotime software in conjunction with a tmhra100 table. What happened was that all of a sudden the session could not handle a hundred records - it would just stall. If we process 50 records they run, but about 4 times slower than before.

We do not own source - I have it to view, but I can not compile it. Would anyone have any suggestions for something to check? Any suggestion would be helpful. Also, so system changes or any patches have been applied.

Clues, suggestions?

Thanks

Mark

Hi Mark -
I am not sure about Autotime, but for us this 1 question has to be answered first - is the problem in the DDC somewhere or is it the functions themselves.
You can find this out in a couple of ways. What happens when you run things through the DDC session "Test Function Server"? If this is also slow, it's the functions.

You can also check the DDC Transaction logs in Baan - they will tell how much time it took to process each transaction once it reaches Baan. If this is very small, it's another clue that the functions are ok, it's the ddc or other influences. Does Autotime have tracing logs for their side? For instance, our logs can tell where exactly in the several steps that the slow down is occuring (PC/handheld to Local controller, Local controller to Baan DDC Driver, Baan DDC Driver to Baan function, then back). It could be ddc, network related or Baan itself.

If it turns out to be the Baan code, try a database trace and see if any of the updates are taking a long time. We do not use the standard DDC hours accounting code (tihra1101.b), does Autotime? I am betting not...we have our own version tihra1101900.b. Any coding questions would have to be referred to the Autotime developers if they are not using the standard tihra1101.b code. If they are, though, you can probably get standard support for this.

Thanks Lisa. I will look for some log files. one of my Oracle people(actually the person running the session) already put a trace on it and it looks like all of the queries that make it to Oracle are running just fine. You are correct the script for the session is tihra1101.b2. Everything looks fine until it hits the tihra1101 funtion server, then it just crawls or dies if there are too many records in the tmhra100 table. Not sure where I will go at this time, but I am just looking for any suggestions.

Thanks again!

Mark

Okay - we rebuild a couple of indexes - one on tihra100 and one on tuddc100. We are not sure yet, but it looks like everything is running again with no problems. This confuses me - why would it work if we put 50 records into tmhra100(notice this really is tmhra100) and not when we put in 100? Maybe I just do not understand this process, but if 50 worked why would not 100.

Mark