Ulrich Windl wrote:
Howard Chu <[email protected]> schrieb am 07.11.2014 um 11:16 in Nachricht
<[email protected]>:
Ulrich Windl wrote:
Howard Chu <[email protected]> schrieb am 06.11.2014 um 14:42 in Nachricht
<[email protected]>:
Just recently measured: when doing a bulk-load of a DB that's larger than
RAM,
it's faster to turn off WRITEMAP and just use regular writes.
Doesn't that depend on the ratio of change compared to the size of RAM or
DB?
Read and think before you post. You have no idea WTH you're talking about,
but
you could easily have read the code or docs or papers before posting and
actually known something.
I cannot follow your reasoning why you think that I don't think, and what
convinced you that I have no idea what I'm taking about.
Wasn't it you who made some vague performance statement/recommendation
without presenting any numbers or details on the tests one might assume you
have made before writing?
There is nothing vague about "bulk-load a DB that's larger than RAM."
I also don't see why I should read your code just because you made some
statements I cannot follow.
If you aren't going to do the basic homework then you have nothing to
contribute.
--
-- Howard Chu
CTO, Symas Corp. http://www.symas.com
Director, Highland Sun http://highlandsun.com/hyc/
Chief Architect, OpenLDAP http://www.openldap.org/project/
