[Monotone-devel] Re: file-system archaeology

[Lapo Luchini]

address@hidden wrote:
> The page size was probably changed for performance reasons.  However, 
> the large small database phenomenon does seem to interfere with several 
> BestPractices:
> 
> http://monotone.ca/wiki/DifferentDbsForServeAndWork/
> http://monotone.ca/wiki/OneBranchPerDbForServe/

Err... I guess we're talking about a server that serves several hundreds
branches, or I fail to see how an overhead of 304KiB per-DB can be a big
issue...

Also, those 304KB are the overhead of *empty* DBs, as the size for DBs
with many revision is actually smaller with a bigger page size, AFAIR
the optimization was really a win in total file size (which fit better
in caches).

To experiment page_size, a simple shell script can be used, such as:

( echo 'PRAGMA page_size = 1024;' ; echo .dump | sqlite3 old-db ) \
| sqlite3 new-db

-- 
Lapo Luchini - http://lapo.it/

“The best way to predict the future is to implement it.” (David
Heinemeier Hansson)