John, Going off the top of my head and what you posted I would suggest a dual processor machine with at least 1 GB of RAM. (Make sure you compile MySQL with threads) SCSI is going to be much faster than IDE. If this is a production machine I would highly recommend SCSI. Finally, I would recommend RAID 0+1 (Striped and mirrored) As that gives you the best overall combination of performance in a database server.
If you have to start cutting then you could drop the RAID and one processor and still have a well-performing machine. Next thing you need to think about is how to configure MySQL. Give it as much RAM as you can. If you are using InnoDB tables then check out http://www.innodb.com/ibman.html#InnoDB_tuning for more suggestions. They used to have several versions of my.cnf suggestions (small, medium and large) but I can't seem to find them anymore. HTH, =C= * * Cal Evans * The Virtual CIO * http://www.calevans.com * -----Original Message----- From: John D Siirola [mailto:[EMAIL PROTECTED]] Sent: Wednesday, July 17, 2002 2:35 PM To: [EMAIL PROTECTED] Subject: Help with spec-ing new MySQL server Hello, I have been asked to spec out a new computer that will act as a dedicated MySQL server for a research cluster (running under RH Linux). To do that, I am trying to find information on the performance trade-offs between multiple processors (2 vs 4), the disk system (IDE vs SCSI; single vs RAID), and the memory system (1 GB, 2 GB, >2 GB). Specifically, is there a point above which you (typically) don't see a significant improvement? Is there a "rational" upper limit to the amount of physical memory that MySQL can effectively use or benefit from? Does an IDE RAID system significantly improve the database performance? Does SCSI RAID significantly outperform IDE? Should we concentrate on a system with an enormous amount of memory, or a very fast disk system? Based on current usage, I am expecting the database load to float around 200-400 simultaneous connections to a small number of databases (1-10). Each database will have several tables with a couple hundred thousand records each (each table will float between a MB and a GB). Thanks for any input, john ----- john Siirola Department of Chemical Engineering Carnegie Mellon University [EMAIL PROTECTED] --------------------------------------------------------------------- Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail <[EMAIL PROTECTED]> To unsubscribe, e-mail <[EMAIL PROTECTED]> Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php --------------------------------------------------------------------- Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail <[EMAIL PROTECTED]> To unsubscribe, e-mail <[EMAIL PROTECTED]> Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
