Everything was working fine until I upgraded from PHP 5.2.0 to PHP 5.2.1. I compiled from source to upgrade, and I installed PHP 5.2.0 from source. Once I upgraded, eAccelerator complained that I had to recompile it again for it to work on 5.2.1. I did and it gave me the same error. I recompiled maybe 10 times and it has still complained about it. How do I fix this error?
Does anyone have a recommendation for the eaccelerator eaccelerator.shm_ttl setting? If it was left at the default of "0", this means no newer scripts would ever be cached once the memory is full. Or would it do additional caching on disk as well and not show up in the phpino.php file?
We have alloted 32 Megs of Ram for eA. Maybe something like this would help move some old scripts out of the cache?
When eAccelerator doesn't have enough free shared memory to cache a new script it will remove all scripts from shared memory cache that haven't been accessed in at least shm_ttl seconds. By default this value is set to "0" which means that eAccelerator won't try to remove any old scripts from shared memory.
Does that mean that once the shared memory is full, that even if a script that is in shared memory hasn't been accessed in 5 months, any new scripts will have to be ran off of the compiled versions on disk instead of being read out of the much-faster RAM?
i installed eAccelerator, everything look good, exept i have a problem with Wifidog-auth server that i have installed, it gives me errors that it cant connect to database.... but not all the times, could eaccelarator be doing this? it was fine before i installed EA
Fatal error: Uncaught exception 'Exception' with message 'Unable to connect to database on localhost' in /home/lavals/public_html/auth1/wifidog-auth/wifidog/classes/AbstractDb.php:96 Stack trace: #0 /home/lavals/public_html/auth1/wifidog-auth/wifidog/classes/AbstractDb.php(321): AbstractDb->connect(NULL) #1 /home/lavals/public_html/auth1/wifidog-auth/wifidog/classes/Server.php(227): AbstractDb->execSqlUniqueRes('SELECT server_i...', NULL, false, true) #2 /home/lavals/public_html/auth1/wifidog-auth/wifidog/include/common.php(107): Server::getCurrentServer(true) #3 /home/lavals/public_html/auth1/wifidog-auth/wifidog/hotspot_status.php(53): require_once('/home/lavals/pu...') #4 {main} thrown in /home/lavals/public_html/auth1/wifidog-auth/wifidog/classes/AbstractDb.php on line 96
I have a Linux server for shared hosting in which I am using Cpanel/WHM. I have PHP running as suPHP which I believe is for security. The problem I am facing is a lot of PHP based websites create load on the server and consume as much as 10% of the CPU and sometimes some script even consumes 50% CPU. I think I can reduce the load caused by the PHP scripts by installing eAccelerator. However, it does not work with PHP running as suPHP. Can anybody tell me which one should I choose of the both? Is there any other way to reduce the load on the server?
I successfully installed eAccelerator and its already working fine then I decided to update my PHP version which is 5.1.6 to 5.2.6 after the upgrade eAccelerator is no longer loaded I tried recompiling eAccelerator but to no avail it doesnt load anymore I double checked eaccelerator.ini and it was already set. Someone said that the phpize that was installed in my server is not compatible to PHP5.2.6 and it needs to be updated also but I cant seem to update it since it always says that it is already updated.
I read on some threads here, that installing Zend Optimizer and Eaccelerator, speeds up the box.
I already installed ZEND optimizer on my box, can someone explain benefit's of Eaccelerator , and how to install it and tweak it for best performance ?
Current Specs -
Quote:
PHP 4.4.8 (cli) (built: Feb 21 2008 01:17:23) Copyright (c) 1997-2008 The PHP Group Zend Engine v1.3.0, Copyright (c) 1998-2004 Zend Technologies with Zend Extension Manager v1.2.2, Copyright (c) 2003-2007, by Zend Technologies with Zend Optimizer v3.3.3, Copyright (c) 1998-2007, by Zend Technologies
I tested eaccelerator with php 5.1.6 but I noticed just a very minor performance improvement. Eaccelerator is known for his incompatiblities, also php 5.1 is a big boy now and no longer need an accelerator.
Code: wget [url] #### replace below with the version tree you have (I will use 4.4) cp ioncube_loader_lin_4.4.so /usr/local/lib/php/extensions Add this line BEFORE any other zend_extension (e.g. Zend Optimizer or eAccelerator). (#### replace below with the version tree you have (I will use 4.4))
Using version 0.9.5 with the default settings. The cache fills up the shared memory in less than a day. I noticed the cached script is stuck at 176, what happens after this? Will it cache content to disk in the temporary folder (/tmp/eaccelerator/) when the shared memory is full?
I'm sure all of you who use eaccelerator know how big the cache can get and that it needs emptied manually. Well at least to my knowledge it does, the only thing I have seen that you can configure is the shm pruning. If anyone does know such a feature with eaccelerator please share. But I also notice a performance decrease and a few php errors mostly related to memory allocations here and there.
Anyway I was piddling around and came up with a command to disable eaccelerator from php.ini, delete the cache folder and then enable it back. I figured this would be good as a daily cron. I would like some input if anyone knows any better ways.
So I wanted to share this in case someone else here has the same problems with eaccelerator cache folder getting huge. If anyone knows a way to make this better or shorter please reply with solution.
Of course this will vary depending on where your php.ini is and where your eaccelerator cache is. Just replace those values with yours. I'm sure there is also a way to use similar commands to find and input these values, again if anyone knows please share.
I have been trying to figure out which one I should be using for my wordpress site. I have read some discussion but after reading it all I am on the fence. Will Xcache work with cpanel?
The server is running XEN, with cpanel/whm, php5
The server spec is: 3.0Ghz - 8 cpu's 2048MB RAM, and 2048MB SWAP
I have also heard some people using litespeed server, would this be any good?
Server : pentium IV 3.0 Ghz debian 3.1 kernel 2.6 apache2 php4.3.10
Some months ago I installed php4 on this server , and added eaccelerator.net , mostly with this site as guide. Everything worked fine.
Few days ago I realize I hadn't updated this server for a while. I run 'apt-get update' and 'apt-get dist-upgrade' . A whole lot of things get updated, and after that I perform a reboot, as adviced by the update script.
The next day I experiment a bit with the profiler of xdebug . I manage to install it , play around with the results, but decide to uncomment it from php.ini again as it is not that helpfull to me and it makes quiet big log files too.
Today I notice this server is getting very big loads. Normally on peak evenings it would float around 2-4 , with sometime jumps to 10-12 but always quickly recovering and working well. But on this evening it kept rising, stayed around 30-60 for most of the time, and even hit up to 85 . It only lowers when the busy hours are really over. So I look around everywhere , and then after a few hours I notice that eaccelerator is not visible in phpinfo(); anymore. This is the only weird thing I could find, and I assume this caused a big difference on the performance.
I dont know why exactly it went away. Maybe it was the server update, wich touched a lot of files. But I thought it was not supposed to update any configuration files, not without warning at least. I could also have been the xdebug script, wich also uses 'phpize' to install and also required editing in php.ini . I really wish I noticed the lack of eaccelerator after the update but before trying xdebug.
So I try to install eaccelerator again. Unpack from source, phpize, make, etc etc, just as last time. I edit php.ini to make sure the load command and settings of eaccelerator are still there and reload apache. I check phpinfo(); , but nothing changes. I try it all again to be sure, but still nothing. I check the location of the .so files, yes they are still there and the rights of the files look ok.
I edit php.ini again and uncomment 'extension=mysql.so', and reload apache. I notice a mysql error on my site as expected. So that means I'm really editing the right php.ini file and that I can influence the extensions that are loaded. I add the mysql extension again and reload apache.
I load an old php.ini backup file, from right after first installation of this server. The eaccelerator settings are there too already, but if I try it and reload apache : no difference.
I reboot my server again, because .. well sometimes it helps. But not with this. phpinfo(); still says no eaccelerator.
I check another server, exact same specs, same php+eaccelerator installation on the same day as the first one. I compare versions of apache/php/php extensions/zend extensions and all numbers I can find. Everything still looks the same, except server 1 does not have eaccelerator and server 2 still has.
I'm shomehow out of things to check now. I'm pretty sure eaccelerator has a big influence on the performance of my heavy php site, and I want it back. Everything looks ok, the files are there, the settings in php.ini are there, it should work ... but it doesnt. No errors or warnings anywhere, just phpinfo(); saying that eaccelerator has not been loaded and the high loads are confirming this.
It will either be a very complicated solution, or a very simple one, but I'm just looking over it. Does anyone has any tips on where to look or what to try ? Thanks.
Does it make sense to enable zlib compression if eAccelerator is already installed? I'm running PHP 5.2.1 (FastCGI) with lighttpd 1.4.13 and eAccelerator 0.9.5. Also, here's a more general question. Since I manually compiled these programs, how would I go about updating them with the source files? Would I just repeat the normal installation process with the new source files?
Running gallery2 on a box with 4GB RAM. Tried PHP 5.2 w/XCache then with eAccelerator, both caused segmentation faults after a couple days.
"In a previous article, I stressed a common problem with all PHP op-code caches/accelerators: they die with segmentation faults every once in a while." [url]
Anybody using logwatcher? How many faults per day do you get? If you're using a cache and never get segfaults how busy is your site and how do you prevent them?