Not sure what you mean with "Total VM usage or just mserver5 process" What I can tell you today is the following: - Created a VMWare workstation system under KUbuntu 64bit 8.04 - latest updates - VM Has - 1Processor, 2GB memory, 30GB disk space - OS = Ubuntu Server 64bit 8.04 with latest updates - Compiled monet from Jun Super Tar with --enable-optimise --enable-sql --nightly-stable - Created database with exact same scripts as used for 32bit systems - started copy into process using script... <code> copy 6157114 records into PrimalPool from '/home/cfourie/temp/PrimalPool.csv' delimiters '|', '\n' null as '' GO --score copy 12240201 records into score from '/home/cfourie/temp/score.csv' delimiters '|', '\n' null as '' GO --mailpot copy 10078100 records into mailpot from '/home/cfourie/temp/mailpot.csv' delimiters '|', '\n' null as '' GO </code> - Got back to the office this morning. The first copy of 6mil records was still running, mserver5 consuming 19GB VM according to top I'm stumped :) Next test - - keep existing ubuntu server install within VM - recompile monet same as before BUT with 32bitOID (2 billion record limit, I think we'll be ok for now :) - recreate DB same as before - attempt to load data again As a point of interest I'll enable memory debug and major IO mask. Maybe that gives me a pointer. Also related to this...Is there a known limit of the number of columns in a monetdb table? Could this be an issue? That will be my next step ;) -- View this message in context: http://www.nabble.com/COPY-INTO-on-Ubuntu-64bit-is-slow-tp18833879p18868004.... Sent from the monetdb-users mailing list archive at Nabble.com.