[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: 2.1 & 2.2 statistics, and some odd behavior that needs to be examined.





--On Wednesday, February 11, 2004 3:37 PM -0500 Jong <jongchoi@OpenLDAP.org> wrote:

Oh... I was referring to the difference in the slapadding performance -
7~8 hours vs. 2 hours
in your old mail. But the same assesment can be applied to the read
application as well, because BDB needs to write to the memory pool even
when read is performed -> the entries read from disk drive need to be
written to the BDB memory pool, and from the viewpoint of the BDB memory
pool, it is that continuous writes are taking place when entries are
brought from the database file.

Hm... well, with the way it currently behaves, it is quite frankly pointless for anyone to use the file-backed DB caching if they will be doing more than simple binds -- Which is probably a large number of people and there are probably a number of them that can't use shared memory DB caching, so this might become a rather critical issue.


--Quanah

--
Quanah Gibson-Mount
Principal Software Developer
ITSS/TSS/Computing Systems
ITSS/TSS/Infrastructure Operations
Stanford University
GnuPG Public Key: http://www.stanford.edu/~quanah/pgp.html