Hi all, is it possible to set persistent connection for Propel to speed up DB? I played with config file, but no any effect on speed of app. All time i got ~54req/sec. If it is possible to set persistent connection can u share configuration file how to do it. Thanks!
On Mon, Jul 9, 2007 at 2:00 PM, <[EMAIL PROTECTED]> wrote: > Send users mailing list submissions to > [email protected] > > To subscribe or unsubscribe via the World Wide Web, visit > http://lists.agavi.org/mailman/listinfo/users > or, via email, send a message with subject or body 'help' to > [EMAIL PROTECTED] > > You can reach the person managing the list at > [EMAIL PROTECTED] > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of users digest..." > > > Today's Topics: > > 1. Re: How to Speed Up Agavi app to >500req/sec? (Mike Seth) > 2. Re: How to Speed Up Agavi app to >500req/sec? (David Z?lke) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Sun, 08 Jul 2007 16:15:42 +0200 > From: Mike Seth <[EMAIL PROTECTED]> > Subject: Re: [Agavi-Users] How to Speed Up Agavi app to >500req/sec? > To: Agavi Users Mailing List <[email protected]> > Message-ID: <[EMAIL PROTECTED]> > Content-Type: text/plain; charset=ISO-8859-1; format=flowed > > Liutauras wrote: > > Hi all, > > > > i want to speed up agavi app and reach >500 requests per second. > > use database is false, > > development is false, > > page execution time filter shows about 0.08s > > > > i try for testing : > > > > # http_load -parallel 5 -fetches 100 urls_file.txt > > (urls_file.txt content just one line > http://localhost/~code/0.11/samples/pub/index.php/en/) > > > > and got poor performance output is here: > > > ------------------------------------------------------------------------------ > > 100 fetches, 5 max parallel, 512800 bytes, in 5.50459 seconds > > 5128 mean bytes/connection > > 18.1667 fetches/sec, 93158.7 bytes/sec > > msecs/connect: 0.09106 mean, 0.21 max, 0.021 min > > msecs/first-response: 272.438 mean, 877.924 max, 104.075 min > > HTTP response codes: > > code 200 -- 100 > > > --------------------------------------------------------------------------------- > > as you can see there is only 18 requests made per second :( what else > > can i do? > > > > When i set use_translation == false i got ~43 requests per second. > > Shoud i turn off everything for best performance? > > I tested old old mojavi framework :) he reached 30req/s in sample page > > with DB and smarty on. > > If i run this test on a sample php file with just one line of php code, > > i got 2350req/sec. > > > > > > How you achieved : > > "Our largest request / second rate on an Agavi powered application is > > approx. 2500 reqs/sec on a single server." > > and > > "Currently managing about 500 reqs/sec with an ease." > > Info taken from: http://trac.agavi.org/wiki/ShowCase/MiDaSi > > > > I run test on notebook which has: Intel Core2 CPU T7200 2.00GHz, with > > 1 GB ram. > > Thanks for any advice > > > > > Well, the advice I can give you is not Agavi specific (though you can > probably improve performance by enabling per-action caching) > > First and foremost, play with Apache settings, namely > MaxRequestsPerChild and other resource allocation settings. > MaxRequestsPerChild defines how many requests an Apache child process > handles. You should monitor memory usage closely while running tests and > find a value that wouldn't cause a lot of overhead but won't consume all > the memory either (assuming PHP leaks, which it does). > > Second, of course, is database optimization. Ensure that indices are > present on the columsn that you search on. If you are running something > more advanced than MySQL (i.e. PostgreSQL) you can additionally play > with resource settings to achieve better performance. > > Third, if you have .htaccess mod_rewrite rules, move them to the Apache > config file, and disable .htaccess across the entire application. Kill > Apache modules that aren't in use. > > Beyond that, I suppose, you should add more RAM and switch to a faster > server implementation e.g Lighty + FastCGI > > _______________________________________________ > > users mailing list > > [email protected] > > http://lists.agavi.org/mailman/listinfo/users > > > > > > > ------------------------------ > > Message: 2 > Date: Sun, 8 Jul 2007 15:26:33 +0200 > From: David Z?lke <[EMAIL PROTECTED]> > Subject: Re: [Agavi-Users] How to Speed Up Agavi app to >500req/sec? > To: Agavi Users Mailing List <[email protected]> > Message-ID: <[EMAIL PROTECTED]> > Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed > > Am 07.07.2007 um 15:43 schrieb Liutauras: > > > Hi all, > > > > i want to speed up agavi app and reach >500 requests per second. > > use database is false, > > development is false, > > page execution time filter shows about 0.08s > > > > When i set use_translation == false i got ~43 requests per second. > > Shoud i turn off everything for best performance? > > I tested old old mojavi framework :) he reached 30req/s in sample page > > with DB and smarty on. > > If i run this test on a sample php file with just one line of php > > code, > > i got 2350req/sec. > > > > How you achieved : > > "Our largest request / second rate on an Agavi powered application is > > approx. 2500 reqs/sec on a single server." > > and > > "Currently managing about 500 reqs/sec with an ease." > > Info taken from: http://trac.agavi.org/wiki/ShowCase/MiDaSi > > > > I run test on notebook which has: Intel Core2 CPU T7200 2.00GHz, > > with > > 1 GB ram. > > Thanks for any advice > > First of all, it obviously is pretty difficult to reach 2500 requests > per second with Agavi if you only get 2500 requests per second from a > plain PHP file. > > Other than that, those 500req/s are likely on some huge quad CPU xeon > servers, running linux and so on. You'll never get the same > performance on a local machine, especially not on a laptop where the > hard drive is another limiting factor. > > You really should consider installing APC as well, that improves > performance drastically. But as I said, a dedicated server will > always yield far better performance, especially if your local machine > is running Windows. > > Also, are you really sure you have debug off? > > David > > P.S: excecution time filter only measures action execution time, not > agavi startup or context initialization or the global dispatch code, > but you can add the filter to global_filters.xml too to get a > slightly more accurate result. To see the real exec time, measure in > index.php, but if you want to output the time, set > "send_content_length" to false for AgaviWebResponse, or you won't see > it in the browser. > > > > ------------------------------ > > _______________________________________________ > users mailing list > [email protected] > http://lists.agavi.org/mailman/listinfo/users > > > End of users Digest, Vol 10, Issue 5 > ************************************ > _______________________________________________ users mailing list [email protected] http://lists.agavi.org/mailman/listinfo/users
