Repeating replication error

[rloef]

all,

I'm replicating several databases. The most active one has one table
with close to five million rows, one has about half a million and the
rest are pretty small. The master is version 5 on an ubuntu-5.10 box
and the slave is version 5 on a FreeBSD-6.0 box. Once a day the biggest
table gets an infile of several thousand lines. It always works fine.

Things will be great for about a week and then I'll get an error on the
slave, like today's:

060201 11:02:50 [ERROR] Key in wrong position at page 8563712
060202 10:54:25 [ERROR] /usr/local/libexec/mysqld: Table './test/test'
is marked as crashed and last (automatic?) repair failed

The slave process stops, I'll repair the table, restart the slave
process and things will be fine for some number of days before it
happens again. What's going on? The slave tables are never touched (I
have learned that running OPTIMIZE TABLE... on the slave will cause
trouble, so I don't touch anything on it.) Is this bad memory, or a
config issue?

thx,

r