Bill Stoddard wrote:
Please, no more specialized knobs which 99.9% of the world cares
nothing about.
How do you define that percentage? By domains? In that case 99.999%
probably care nothing about what we are doing. If you look at total
traffic, however, would not options that help the to
Bill Stoddard wrote:
To get mod_cache/mod_mem_cache (I know little or nothing about
mod_disk_cache) really performing competatively against best-of-breed
caches will require bypassing output filters (and prebuilding headers)
and possibly bypassing or at least reworking input filters.
The initial
Mads Toftum wrote:
On Mon, Aug 02, 2004 at 04:54:07PM -0400, Brian Akins wrote:
Any reason md4 was not used in mod_cache? In my ad hoc tests, it seems
much faster. I do not know the in-and-outs of encryption, but is there
any compelling reason to use md5 over md4 in this case? We don't care
i
Brian Akins wrote:
On one hand, I think doing MD5 is sort of silly - just use the URL
itself. *shrug* -- justin
One of the goals of the new mod_cache was to support the caching of URL
variants.
I actually have somewhat of a solution:
URL encode the uri and any vary elements:
www.cnn.com/index.
On Mon, Aug 02, 2004 at 04:54:07PM -0400, Brian Akins wrote:
> Any reason md4 was not used in mod_cache? In my ad hoc tests, it seems
> much faster. I do not know the in-and-outs of encryption, but is there
> any compelling reason to use md5 over md4 in this case? We don't care
> if someone "
Justin Erenkrantz wrote:
MD5 has the possibility for collisions, too.
Any reason md4 was not used in mod_cache? In my ad hoc tests, it seems
much faster. I do not know the in-and-outs of encryption, but is there
any compelling reason to use md5 over md4 in this case? We don't care
if someo
Justin Erenkrantz wrote:
MD5 has the possibility for collisions, too. What do squid or other
proxies do?
True. I'll see what others do.
On one hand, I think doing MD5 is sort of silly - just use the URL
itself. *shrug* -- justin
With mod_disk_cache, how about urls such as:
http://www.cnn.com
--On Monday, August 2, 2004 2:49 PM -0400 Bill Stoddard
<[EMAIL PROTECTED]> wrote:
To get mod_cache/mod_mem_cache (I know little or nothing about
mod_disk_cache) really performing competatively against best-of-breed
caches will require bypassing output filters (and prebuilding headers)
and possib
--On Monday, August 2, 2004 3:01 PM -0400 Brian Akins
<[EMAIL PROTECTED]> wrote:
(and prebuilding headers)
Is there a way to send pre built headers? mod_asis uses
ap_scan_script_header_er which is fairly slow.
That was what I fixed with apr_file_gets() last night. The code there was
really nai
--On Monday, August 2, 2004 2:54 PM -0400 Brian Akins
<[EMAIL PROTECTED]> wrote:
The other bottleneck I looked at was MD5 as the on-disk naming
scheme. I think MD5 is a poor choice here because it's not very
fast. Ideally, switching to a variant of the times-33 hash might
work out better. *shr
Bill Stoddard wrote:
To get mod_cache/mod_mem_cache (I know little or nothing about
mod_disk_cache) really performing competatively against best-of-breed
caches will require bypassing output filters
Yes.
(and prebuilding headers)
Is there a way to send pre built headers? mod_asis uses
ap_
Justin Erenkrantz wrote:
The other bottleneck I looked at was MD5 as the on-disk naming
scheme. I think MD5 is a poor choice here because it's not very
fast. Ideally, switching to a variant of the times-33 hash might
work out better. *shrug*
How to handle collisions?
--
Brian Akins
Senior Sy
Justin Erenkrantz wrote:
mod_cache isn't doing anything stupid or damaging
performance-wise, I'd like to start being more aggressive about what we
can cache. From my perspective, these patches I've posted (and started
to commit) are just the beginning of trying to get mod_cache on more
solid gr
Justin Erenkrantz wrote:
post_config's not a bad place for that. But, I've yet to get a good
handle on my thoughts for the storage mechanism that mod_disk_cache is
using. My hunch so far is that it's really inefficient. (The reading
of the headers one-byte at a time with the brain-dead apr_fi
--On Monday, August 2, 2004 9:21 AM -0400 Brian Akins <[EMAIL PROTECTED]>
wrote:
In a low traffic site, yes. In a very high traffic site, with lots of
objects, the mkdir's kill you. After a "while," most of the directories
will be created. However, bringing up a "fresh" server behind a very b
--On Monday, August 2, 2004 8:46 AM -0400 Brian Akins <[EMAIL PROTECTED]>
wrote:
Another big speed-up may be to pre-make all of the directories. A simple
script could use CacheRoot, |CacheDirLength|, and |CacheDirLevels to create
them all. Just require that this script be ran before starting a
Jeff Trawick wrote:
Wouldn't assuming that the directory is already there be sufficient
(and then create the directory structure on the error path)? It looks
to me that we only assume the directory structure exists if we had
this very same file cached previously.
In a low traffic site, yes. In
On Mon, 02 Aug 2004 08:46:07 -0400, Brian Akins <[EMAIL PROTECTED]> wrote:
> Justin Erenkrantz wrote:
>
> > -
> > static void mkdir_structure(disk_cache_conf *conf, char *file,
> > apr_pool_t *pool)
>
> Another big speed-up may be to pre-make all of the directories. A simple script
> could use
Justin Erenkrantz wrote:
-
static void mkdir_structure(disk_cache_conf *conf, char *file,
apr_pool_t *pool)
Another big speed-up may be to pre-make all of the directories. A simple script
could use CacheRoot, |CacheDirLength|, and |CacheDirLevels to create them all. Just
require that this scr
--On Sunday, August 1, 2004 11:25 AM -0400 Bill Stoddard <[EMAIL PROTECTED]>
wrote:
Too many changes in one patch. Break this up into multiple consumable in 15
minute patches and I'll review them.
Some more work, analysis, and tests yielded apr_file_gets() and MD5 as two
more bottlenecks.
I've
20 matches
Mail list logo