[Date Prev][Date Next]
> Of course, slapadd doesn't have any bearing on what we were discussing. The
> difference observed here is probably just due to the added data volume of
> storing normalized attribute values.
The 5.7% difference between 2_1 and 2_2_5 can be attributed the increased disk I/O.
The difference in slapadd code between the two versions are just a few if statements.
Do you also have some data with the concurrency test program (test008) ?
> By the way, the input LDIF for the test was 97260309 bytes, to give you an
> idea of the expansion going on in the database layouts.