Jim,
Thanks for the fast response.  As to why.  Part of the reason is ignorance. 
When I started working with web applications I had to learn javascript,
python, qooxdoo, cherrypy, zodb, etc.  So when things worked they stayed as
simple as possible.  I would think that there are better ways to accomplish
what I am doing, but what I am doing seems to work and works quite well.

My application is an online marketplace consisting of distinct stores (
books, cars, fabric, music, etc.) in which the client is constantly reading
and writing to the zodb database.  I am using cherryPy on the server side.

1) Actually my app is about to be a collection of marketplaces, each with a
set of stores.  It is feasible that I could have 20 marketplaces and
hundreds of stores in the future.  It makes management very easy by putting
each store in its own file.
 
2) I write alot to my fs files so I worry about size.  The end users create
and manage their stores and their data and they work on the live database
when they do so.  I think that it may be necessary to change a store to a
(forgive me!) relational database if my files get too big.  For reading and
writing.  Having them broken up into separate files makes this easier also.  

3) It is easier to manage.  I can move each store to separate marketplaces
if I wish.  I am trying to make it to where I can change any part of my
system without affecting another. 

I have been using this database for about three years now.  Mostly part time
when not doing my real job, but I have recently started to work on software
fulltime.  This database has NEVER corrupted any data or as far as I can
tell screwed up in any way.  It is downright amazing.  You did one hell of a
job.  

tom



Jim Fulton wrote:
> 
> On Tue, Oct 27, 2009 at 12:17 PM, tsmiller <tsmil...@gnixterhouse.com>
> wrote:
>>
>> Zodb developers,
>> When I open my database I am mapping individual files in the following
>> way.
> 
> Why?
> 
>> I have about 20 fs files mapped in this fashion.  My concern is that I am
>> going to hit some kind of limit.  Is accessing these fs files going to
>> slow
>> down as the number grows?  Am I going to hit an upper limit - say 256 or
>> something like that?
> 
> There collection of databases are managed with a Python dictionary, so
> you should be able to have many thousands of databases without feeling
> pain due to the number. You'll probably run out of file descriptors
> first.
> 
> Why are you splitting your database up so much (or even at all)?
> 
> Jim
> 
> -- 
> Jim Fulton
> _______________________________________________
> For more information about ZODB, see the ZODB Wiki:
> http://www.zope.org/Wikis/ZODB/
> 
> ZODB-Dev mailing list  -  ZODB-Dev@zope.org
> https://mail.zope.org/mailman/listinfo/zodb-dev
> 
> 

-- 
View this message in context: 
http://www.nabble.com/Limits-for-mapping-fs-files-tp26080548p26081951.html
Sent from the Zope - ZODB-Dev mailing list archive at Nabble.com.

_______________________________________________
For more information about ZODB, see the ZODB Wiki:
http://www.zope.org/Wikis/ZODB/

ZODB-Dev mailing list  -  ZODB-Dev@zope.org
https://mail.zope.org/mailman/listinfo/zodb-dev

Reply via email to