[OT] Working with File System

[Jacob Hookom <ho...@uwec.edu>]

Hey All,

If we are storing large amounts of files on the server, is it acceptable
(performance wise) to be querying directories straight from the file
system?  I realize Tomcat does this for directory browsing, but I don't
know if it will scale well enough that I would need to cache the
results?

My idea is to query the file system and use their hash code as a look up
in the DB for extra properties as added by the user (description, etc).

Thanks in advance for any information,
Jacob



--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>

Re: [OT] Working with File System

["Craig R. McClanahan" <cr...@apache.org>]

On Mon, 21 Oct 2002, Jacob Hookom wrote:

> Date: Mon, 21 Oct 2002 05:45:39 -0500
> From: Jacob Hookom <ho...@uwec.edu>
> Reply-To: Struts Users Mailing List <st...@jakarta.apache.org>
> To: 'Struts Users Mailing List' <st...@jakarta.apache.org>
> Subject: [OT] Working with File System
>
> Hey All,
>
> If we are storing large amounts of files on the server, is it acceptable
> (performance wise) to be querying directories straight from the file
> system?  I realize Tomcat does this for directory browsing, but I don't
> know if it will scale well enough that I would need to cache the
> results?
>

Assuming that there is a filesystem at all would make your application
non-portable, because app servers are *not* required to support that mode
of application execution.  For example, Tomcat gives you the option to run
directly from a WAR.

I would suggest using the context-relative path to the resource as the key
to your database lookups.  That way, it will work either inside or outside
a WAR.  If you need metadata about the resource itself (such as the file
size or the last modified date), you can acquire that by using the
ServletContext.getResource() method to get a URL for the resource, and
then using URL.openConnection() to get a URLConnection for it.  From this
you can read the various headers associated with the file.

> My idea is to query the file system and use their hash code as a look up
> in the DB for extra properties as added by the user (description, etc).
>

There's no general purpose answer to whether you need to cache or not.
The only way to find out is try it.

> Thanks in advance for any information,
> Jacob

Craig


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>