I'm trying to get reliable values for memory usage on my php scripts,
some of which are making big pdfs using PDFlib.
I've set up a CustomLog in my apache config file using the PHP memory
usage note to apache and I'm trying to understand the discrepancies
between the peak memory usage value logged via
"%{mod_php_memory_usage}n" and the values that are displayed via SIZE
and RES when using the unix command top.
EXAMPLES of mod_php_memory_usage VS top:
php_log_peak_mem top(SIZE) top(RES)
72408 12404K 7156K
371464 12404K 7900K
45240 12576K 8160K
126112 12452K 7968K
1073304 13840K 9596K
2678008 17620K 12724K
3094000 18796K 13892K
I've had a hard time finding resources to perform proper testing and
capacity planning for my Virtual Private Server hosting setup. For the
most part, I've been successful getting some guidelines with load
testing using JMeter but I've been requested to provide some sort of
outlook on my scripts memory consumption and how that translates into
scaling numbers of concurrent users.
Any info on the reliability of utilizing the top command would be
helpful because I have some non-php CGI scripts that I need to track
as well. It's easy to throw hardware/memory/money at things but I'd
like to provide options that are sensible and affordable AND have the
reasons why changes are necessary.
The big question:
Why is there a difference and what is the difference between the
figures logged by mod_php_memory_usage and the numbers reported by the
SIZE and RES columns of the top command?
Any suggestions on how I can get up to speed on planning server
requirements around my existing scripts would be extremely helpful as
well. (Useful testing experts/how-to's/books/websites etc.)
Thanks much.
Eric
--
PHP General Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
