Apache::AutoIndex - Perl replacment for mod_autoindex

[Petyr David]

anyone have an experience using this?

The real question:

does using this speed the creation of a directory index in Apache
significantly? We have directories with thousands of small files.

Thanks!

 

[J. Gleixner]

anyone have an experience using this?

The real question:

does using this speed the creation of a directory index in Apache
significantly? We have directories with thousands of small files.

It's more likely that it'll be slower because mod_autoindex is
written in C and compiled into the Apache daemon, you're not
going to get much faster than that.

Possibly you could list 500 at a time, or something, which
would be faster, however having thousands of files in a directory
isn't typically a good design.

 

[Petyr David]

It's more likely that it'll be slower because mod_autoindex is
written in C and compiled into the Apache daemon, you're not
going to get much faster than that.

Possibly you could list 500 at a time, or something, which
would be faster, however having thousands of files in a directory
isn't typically a good design.

agreed, but the nature of the data forces us to store the files in
this fashion so there's some sense of meaningfulness: every file is
named and then has a sequential number appended to it - there 64K
potential numbers. we might have to start something like breaking it
into 1000 files/directory.

either way - thanks for your opinion

 

[smallpond]

agreed, but the nature of the data forces us to store the files in
this fashion so there's some sense of meaningfulness: every file is
named and then has a sequential number appended to it - there 64K
potential numbers. we might have to start something like breaking it
into 1000 files/directory.

either way - thanks for your opinion

Why not use a real database instead of making one out of a
filesystem? Directory searches are slow, sequential string
compares. Database lookups use fast hash techniques.

 

[J. Gleixner]

agreed, but the nature of the data forces us to store the files in
this fashion so there's some sense of meaningfulness: every file is
named and then has a sequential number appended to it - there 64K
potential numbers. we might have to start something like breaking it
into 1000 files/directory.

either way - thanks for your opinion

You could also generate one or more static HTML files, once a day or
whenever new files are added, which would display the filename
and contain a link to the file. Heck, you could probably display
5-10 per line, too. That way you'd avoid having to stat
the thousands of files every time the directory is accessed.

 

[xhoster]

Why not use a real database instead of making one out of a
filesystem?

Why not use a file system rather than making one out of a database?
Admitted, appending a sequential number to the file name is rather strange,
but that still doesn't mean that he really wants a database rather than a
file system.

Directory searches are slow, sequential string
compares.

Not if you already know the exact name of the file. At least, not on any
reasonable file system. They use trees or hashes or something to find the
file quickly.

Database lookups use fast hash techniques.

Same as reasonable file systems, given an exact name. Of course, if you
are doing globs, or just extracting all entries, then neither file systems
nor databases will use fast hash techniques.

Xho

--
-------------------- http://NewsReader.Com/ --------------------
The costs of publication of this article were defrayed in part by the
payment of page charges. This article must therefore be hereby marked
advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate
this fact.