#23953 [Opn]: es.php.net is so slow it is unusable

From: Date: Tue, 03 Jun 2003 14:29:53 +0000
Subject: #23953 [Opn]: es.php.net is so slow it is unusable
References: 1  Groups: php.mirrors 
Request: Send a blank email to php-mirrors+get-18451@lists.php.net to get a copy of this message
ID: 23953 User updated by: pete at foxcreekleather dot com Reported By: pete at foxcreekleather dot com Status: Open Bug Type: Website problem Operating System: NA PHP Version: 4.3.2 New Comment: I wasn't suggesting that automic apachebenches be done from a central location. I was suggesting that the potential mirror site do it themselves from a computer in the community they would serve giving them and you a better idea of whether or not they can meet the estimated 150mb per day load. Given the current situation it appears that, es.php.net is either receiving more traffic than they anticipated or they underestimated their ability to handle an extra 150mb a day on whatever machine they're using. Perhaps the 150mb a day expectation should be upped for 2003 or if the later is true they could use a load tool to get a better idea of whether or not they can handle the estimated traffic. I hope that is clearer. Previous Comments: ------------------------------------------------------------------------ [2003-06-03 09:05:52] goba@php.net We cannot do automatical apachebences from one place as said before, it would not be accurate data. We suppose the those who offer mirror hosting has the bandwidth (this is also emphasized on /mirroring.php) ------------------------------------------------------------------------ [2003-06-03 08:51:42] pete at foxcreekleather dot com es2.php.net is definitely acceptable, and much faster than es.php.net. However I compared a few function searches on es2.php.net to us2.php.net and us2 (just by my own counting) appears to be at least 3 to 4 times faster (I am in barcelona, high speed access). Perhaps mirrors should be responsible for measuring it themselves when they apply for certification? Whoever the maintainer at the mirror is just goes home at the end of the day or to the local internet cafe and runs apachebench or something similar with predefined parameters. If it can't carry a certain load it's a no go. Or based on the apachebench numbers the redirects could be weighted. es2.php.net gets 5 to es.php.net's 1. ------------------------------------------------------------------------ [2003-06-03 08:30:26] goba@php.net After consulting the maintainer, he sent me this message: | We're sorry to hear that! | We have some congestion with our outgoing traffic, | but we are working on it. | | We're planning to move to another physical location | before July 6th. In this new place, we will have | much more bandwidth, and this problem will be | solved. I am going to disable the mirror with agreement with the maintainer until they can get better bandwidth. BTW the question of measuring mirror speeds is still up, in case someone has an idea. :) ------------------------------------------------------------------------ [2003-06-03 08:24:43] goba@php.net Ups, disregard my last paragraph, that is not in connection with this bug report. But the first para is correct. BTW is the speed of es2.php.net OK? ------------------------------------------------------------------------ [2003-06-03 07:57:43] goba@php.net I can perfectly understand your reasoning Pete, but we have no method to measure mirror speeds from a central administration machine. We are able to check if it is up, if it is updated, and check several other parameters it provides about itself. But that it is slow if accessed by our bot does not mean that it is slow when accessed from a local machine. There are several ISPs, who have very weak international connections but a strong local connection. We can only act on user reports. At least I don't know of a solution... BTW the ee2.php.net is now free, as the Korean mirror which was there is removed now. So if some ISP comes with an offer to set up a mirror in Estonia, we won't turn it down. ------------------------------------------------------------------------ The remainder of the comments for this report are too long. To view the rest of the comments, please view the bug report online at http://bugs.php.net/23953 -- Edit this bug report at http://bugs.php.net/?id=23953&edit=1

« previous php.mirrors (#18451) next »