Dear Rahul. thanks for the information. We haven't ran the TPC-H out of the box on this scale factor yet. The restart time might be related to cleaning out the transaction log, which it does upon restart. regards, Martin Rahul Chopra wrote:
I ran into some problems loading data for doing some TPC-H tests. After a successful test for a factor 1 database, I moved on to test a factor 20 database.
mserver5 appeared to crash when loading partsupp, orders, or lineitem. On a hunch, I ran split on these to ensure that each file was less than 2 GB each, splitting partsupp and orders into two files, and orders into 10. After doing this, partsupp and orders loaded without any problems. However, lineitem would cause mserver5 to crash after loading 2 or 3 of the 1.5 GB chunks. Also, in restarting mserver5, I noticed that CPU usage was sometimes heavy for a while, before starting loading. I am guessing it is doing some recovery or cleanup?
Is loading an issue? Since it was busy upon restart, do I need to wait a while before loading a new file to give the system some time to do any housekeeping or something?
I am using the latest released sources (not from HEAD), compiled on CentOS 5.1 (64 bit).
Thanks,
Rahul
= Live Answering Service Answering. Live operators 24/7. Rates start at $39/100 calls. http://a8-asy.a8ww.net/a8-ads/adftrclick?redirectid=202d1706cc4cee71bfb381fc...