Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:21757
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Message-ID: <200602012224.RAA21594@bellerophon.concentric.com>
Errors-To: <jluedke@concentric.com>
To: Rasmus Lerdorf <rasmus@lerdorf.com>, <jluedke@concentric.com>
Reply-To: jluedke@concentric.com
Cc: <internals@lists.php.net>
Date: Wed, 01 Feb 2006 17:24:21 -0500 (EST)
In-Reply-To: <200602011939.OAA10156@alexander.concentric.com>
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
Subject: Re: [PHP-DEV] reducing open calls in php
From: jluedke@concentric.com (James M Luedke)
References: <200602011939.OAA10156@alexander.concentric.com>

I tested this and you are correct. This should make a big difference.
I learn something new every day :)=20


Thanks a million.

-James


---- Rasmus Lerdorf <rasmus@lerdorf.com> wrote:
>
> I think the main hint you need is to never use require_once/include_once=
=20
> with an opcode cache.  Those are the only ones that will cause an open()=
=20
> call.  You should only be seeing a single stat() call per file under APC=
=20
> (needed to get the inode+device hash key for the shared memory lookup).
>=20
> -Rasmus
>=20
> James M Luedke wrote:
> > Hello:
> >     We have a very high-volume site that runs across a cluster of sever=
al
> > machines which needs to have access to our back end storage platform, w=
hich
> > needless to say is very busy. I have been given the task of reducing lo=
ad to
> > the filer. Due to the fact that our application is rather large and inc=
ludes
> > lots of files which reside on our filer, php would be a good place to c=
ut down load.
> >=20
> >     We currently have a rather unique setup. We are running php with
> > fast-cgi support via sapi/cgi/cgi_main.c as well as APC for opcode/user
> > caching**. This allows us to run php as a daemon which reduces system c=
alls
> > and retains persistence with APC. Our setup has been working quite well=
.
> > However, there is one behavior that is less than optimal.
> > =20
> >     Whenever one of our pages is accessed, it appears*** that the reque=
sted
> > file and all included files are opened, which results in several open c=
alls to our filer. As I mentioned, we are running APC which seems to be doi=
ng a
> > nice job of op-code caching. It would be ideal if somehow I could modif=
y php
> > to first check APC cache for the pre-compiled opcode before attempting =
to
> > open the file. This would be a huge win on our cluster, since each hit
> > currently generates 5-10 open calls.
> >=20
> >     Now please bear with me because I am new to the Zend / main tree of=
=20
> > the php source. If my understanding is correct, the basic flow of=20
> > cgi/php goes something like this:                   =20
> >=20
> >    request
> >    some variable init.
> >    module_init <-apc overide compile_file w/my_compile_file
> >    php_fopen_primary_script(&file_handle TSRMLS_CC);
> >    php_execute_script(&file_handle TSRMLS_CC);
> >        VCWD_CHDIR_FILE(primary_file->filename);
> >        # i think included_files gets opened here someplace
> >        # i do not understand that magic yet.
> >        zend_execute_scripts()
> >            zend_compile_file()
> >                compile_file || my_compile_file() <-- apc
> >                # it looks like the std compile_file
> >                # may do some zend_open cmds here via
> >                # open_file_for_scaning()
> >            zend_destroy_file_handle() <-- destroy before execute?
> >            zend_execute()
> >        php_request_shutdown()
> >     FCGX_Finish_r;
> >=20
> > I have a few questions:
> > 1. When exactly do the files get opened/read? And is this a necessary=
=20
> > step if you already have the op-code.
> > 2. Is it possible to check apc_cache before opening files? If it=20
> > is possible, how/where do you think it could be best implemented?=20
> > Is there some similar interface to the my_compile_file trick used=20
> > to override zend_compile_file by apc? Perhaps something like=20
> > my_open_file, which I could then add support for into the apc module?
> > 3. would changing utility_functions->fopen_function; to point to=20
> > my_fopen_function in place of zend_open work.
> >=20
> > Any input would be greatly appreciated.
> >=20
> > -James
> >=20
> > ** ( mod_php and the likes are not an option for us as we have our own
> > custom web server which we cannot do without  :)  )
> >=20
> > *** when i truss one of the php daemon process i can clearly see that
> > it opens and reads the requested file. It also opens all the=20
> > included/required files but does not seem read them.
> >=20
>=20
> --=20
> PHP Internals - PHP Runtime Development Mailing List
> To unsubscribe, visit: http://www.php.net/unsub.php
>=20
>=20
>=20