
www.Usenet.com
| <-- __Chronological__ --> | <-- __Thread__ --> |
I have an application which needs to query across 200 different Dbs. It runs of of virtual memory. I'm using Java which I think is significant. I tried a similar test in C++ and no problem. Each database consists of a 500 MB BTree and another Btree which is only 128K in size (stores metadata). Therefore there are 400 Db files to open although only half of them are 500 BM in size. Anyway : cache is 128MB. I monitored the memory usage: It starts off at about 171M. Then by the time we are trying to open the 172th pair of Dbs it has gone up to 1.2 Gigs !!! About 5 MB per pair of Dbs. I would envisage needing to be able to query across more Dbs than 200 so can anybody point out whats causing these allocations ? Cheers Chad.
| <-- __Chronological__ --> | <-- __Thread__ --> |