Hello everyone.
I have a kind of fact table, containing just 4 small columns, which I filled with about 1.1 billion rows of test data. The table is essentially filled with random double values which I created in a procedure which totally ran successfully.
However soon after it ran the hana database started to do the delta merge. Unfortunately it seems like the database runs out of memory during the delta merge, so the specific thread fails. Soon after that, the database starts a delta log replay and afterwards again a delta merge - which fails again and again. So actually the whole system is stuck in this kind of loop of trying to recover the delta log and doing the delta merge. The whole behaviour even still prevails, after I configured the indexserver not to do auto merge. I also restarted the machine itself.
I wonder how I can escape this endless loop? I dont mind losing the data in the delta buffer, nor the main store of the fact table itself, since its test data only, but I don't want to lose all the other stuff / tables since the last system backup.
Is there a way to delete the delta buffer for this specific table or preventing the system to do a delta replay automatically after restart? Actually the whole system is very sluggish during the heavy job execution, but there's a small timeframe between the jobs (maybe 1-2 minutes) in which I can execute everything relatively fast and it seems like theres no lock on the problem causing table in this timeframe. Also there's no way to cancel the jobs during their execution, that's why I always have to wait a long time (~1-1.5 hours) to get to this "free" timeframe and do the next attempt. I already tried to delete a few million rows of that table during this timeframe, but it seemed that didn't help at all or even made the problem worse. Next I'll try to drop the whole table during this timeframe - maybe that helps.
Anyway, can anybody give me another hint how to solve that problem plz?
Thanks in advance
Christian