I have several production MySQL servers, but I’m having a strange issue. Both are comparable systems (although not hardware identical) running Ubuntu 8.10 with the exact same software and databae types (and indexes). Both were deployed with apt-get, git, and RoR db migrations, and phpMyAdmin shows indexes match. However one is tasking significantly longer to import data.

Box A
Takes under 15 minutes to import 100k records (I do some RoR based validation)
Has more column rows
Server version: 5.0.67-0ubuntu6
Box B
Is still chugging along after 5+ hours with a 300k record import
Has fewer column rows
Server version: 5.0.67-0ubuntu6-log

Examining the Processes tab with phpMyAdmin Box B always shows a
SELECT COUNT( * ) AS count_all
FROM `contacts`
WHERE (
user_id
IN ( 5 )
AND phone = '44113293XXXX'
);

Query as its current process (my RoR program FIRST checks for a duplicate, and if none is found, adds it). Doing an explain on this statement at the command line shows that both boxes use the correct index. I cannot figure out why one machine performs the import so well, and the other so poorly. The one item I noticed was the server version. Box B is running 5.0.67-0ubuntu6-log and Box A is running the same version w/o log. What does this mean? I’m out of ideas of how to solve this, so your input would be greatly appreciated.

--matt
hello hunter corp
Hello Hunter: Flash-based Hosted Dialer for Sales
Hosted Predictive Dialer & Voice Broadcasting