In error log i found warnings something like that
WARNING: [pool www] server reached pm.max_children setting (10), consider raising it
So I changed it in conf file /etc/php/7.0/fpm/pool.d/www.conf to 52 after some recommended calculations.
After this I restarted apache and I thought that problem was resolved. But right now I see again in error log file same warning with same value 10.
So I have a question - how can I check current pm.max_children? If it is not 52 then I will try to inspect why it is not applying.
It is an old post, surely you have already solved it in the meantime. In any case, maybe the problem was that you also had to restart php-fpm to make the changes effective with something like this:
sudo service php7.0-fpm restart
I try to run a Yii based PHP application in a Docker container using the official php-fpm image.
It's common practice for docker containers to write all log messages to stdout / stderr. So I do the same from my container with a code that basically looks like this:
$fp = #fopen('php://stdout','w');
fwrite($fp, $message);
fclose($fp);
Unfortunately PHP-FPM now prefixes all my log messages with ugly warnings:
[21-Mar-2016 14:10:02] WARNING: [pool www] child 12 said into stdout: "2016-03-21 14:10:02 [x.x.x.x][-][-][trace][yii\base\Application::bootstrap] Bootstrap with yii\debug\Module::bootstrap()"
[21-Mar-2016 14:10:02] WARNING: [pool www] child 12 said into stdout: "2016-03-21 14:10:02 [x.x.x.x][-][-][info][yii\web\Session::open] Session started"
...
Why is this and is there a way to prevent PHP-FPM from doing this?
UPDATE:
I've also opened an issue at the docker-library/php project site.
They already include the catch_workers_output = yes in their configuration. So that's not the solution.
According to these changes the problem should be fixed in PHP 7.3
It is correct. you are running php-fpm and error logs are from php-fpm , not from PHP .
you should expect php-fpm log , not php log. fortunately php-fpm will write every log that received from PHP in end of its logs.
I'm trying to setup an php-fpm daemon to send error_log to syslog.
My config php-fpm.conf looks thereby.
error_log = syslog
syslog.facility = local4
syslog.ident = php-fpm
Restarting php-fpm daemon I can see in the remote syslog server.
Mar 22 00:32:08 192.168.33.14 php-fpm[20919]: [NOTICE] configuration file /etc/php5/fpm/php-fpm.conf test is successful
Good!
But when I test this with a real php script with errors, message change to..
Mar 22 00:05:59 192.168.33.14 ool www[20889]: PHP Parse error: syntax error, unexpected ''api'' (T_CONSTANT_ENCAPSED_STRING) in /var/api/public/index.php on line 2
ool www is pool + default pool name (www). So two questions:
Why I see ool www instead of php-fpm (syslog.ident value)?
Why is pool title is truncated?
I think its a PHP bug :(
See please
https://bugs.php.net/bug.php?id=67764
This issue will become more prevalent as we start to upgrade older systems to php 5.5+, so posting for future reference.
Here is the fix we used, put it as low level as you can so all your scripts can benefit:
// This was needed for upgrade to php5.5+ with php-fpm
// Source: https://gist.github.com/gjuric/e0c9e45efb3d15e3b949
// Bug History: http://stackoverflow.com/questions/22575152/sending-errors-to-syslog-in-php-fpm
openlog('php', LOG_ODELAY, LOG_USER);
I found many errors like this one
[Wed Nov 06 14:34:01 2013] [warn-phpd] mmap cache can't open C:\www\somefile.php (pid 4484 th 1668)
in my Apache error.log file. I tried to pinpoint the source of the error for some time but with no luck so far.
I find out that PHP Opcache is not the culprit.
error_log did not help. I think that my PHP source codes do not affect the error.
My stack: Apache 2.4.6, Windows, PHP 5.4.20
Did anyone encounter the same error?
Note: The error message I get is not the same as, for example, the error:
Mon Dec 1 21:08:20 2008] [warn-phpd] mmap cache can't open /var/www/vhosts/domain.com/httpdocs/file.php - Permission denied (pid 7831)
where there is a reason why mmap can't open the file.
This is caused by the total number of files that are opened by the server. If this is on a hosting company then they would be able to resolve this for you, if you are on your own system then try these steps:
Edit the apache startup script, \Program Files\Apache Software Foundation\Apache2.2\etc\init.d\httpd (may be different on your system) and add this before anything else:
ulimit -n 20480 #Raise the ulimit to a higher value then you have
Then Restart apache using httpd.exe restart
Hope this points you in a general direction
Disable MMAP. It's not supported on Windows.
It's an efficient method to map files to memory, to work on their content. Similar story with sendfile, an efficient method to send the content of a file as a response.
# https://httpd.apache.org/docs/2.4/en/mod/core.html#enablemmap
EnableMMAP On
EnableSendfile Off
Mon Dec 1 21:08:20 2008] [warn-phpd] mmap cache can't open
/var/www/vhosts/domain.com/httpdocs/file.php - Permission denied (pid
7831)
It seems that mmap doesn't has the rights to open the file, check file's folder rights .
Check file's folder properties .
I am having issue on PHP where my app is trying to run a php backup file and suddenly getting HTTP Error 500 Code. I have checked the logs and this what it saying.
[Tue Aug 28 14:17:28 2012] [warn] [client x.x.x.x] (104)Connection reset by peer: mod_fcgid: error reading data from FastCGI server, referer: http://example.com/backup/backup.php
[Tue Aug 28 14:17:28 2012] [error] [client x.x.x.x] Premature end of script headers: backup.php, referer: http://example.com/backup/backup.php
Anyone knows how to fix this? I'm really stuck in here and can't find solution in internet.
Hope anyone could share their knowledge.
Thanks.
James
I managed to solved this by adding FcgidBusyTimeout . Just in case if anyone have similar issue with me.
Here is my settings on my apache.conf:
<VirtualHost *:80>
.......
<IfModule mod_fcgid.c>
FcgidBusyTimeout 3600
</IfModule>
</VirtualHost>
I had very similar errors in the Apache2 log files:
(104)Connection reset by peer: mod_fcgid: error reading data from FastCGI server
Premature end of script headers: phpinfo.php
After checking the wrapper scripts and Apache2 settings, I realized that /var/www/ did not have accordant permissions. Thus the FCGId Wrapper scripts could not be read at all.
ls -la /var/www
drwxrws--- 5 www-data www-data 4096 Oct 7 11:17 .
For my scenario chmod -o+rx /var/www was required of course, since the used SuExec users are not member of www-data user group - and they should not be member for security reasons of course.
if you want to install a PHP version < 5.3.0, you must replace
--enable-cgi
with:
--enable-fastcgi
in your ./configure statement, excerpt from the php.net doc:
--enable-fastcgi
If this is enabled, the CGI module will be built with support for FastCGI also. Available since PHP 4.3.0
As of PHP 5.3.0 this argument no longer exists and is enabled by --enable-cgi instead. After the compilation the ./php-cgi -v should look like this:
PHP 5.2.17 (cgi-fcgi) (built: Jul 9 2013 18:28:12)
Copyright (c) 1997-2010 The PHP Group
Zend Engine v2.2.0, Copyright (c) 1998-2010 Zend Technologies
NOTICE THE (cgi-fcgi)
I had this issue and realized that the file cgi-bin/php-fcgi had no execution rights.
It had 644 mode while is should have 755 mode.
Setting the correct mode was impossible (probably because the file was opened or sth), so I copied that file from another domain directory where it had proper rights already set and that fixed everything.
I had the same problem with a different and simple solution.
Problem
I installed PHP 5.6 following the accepted answer to this question on Ask Ubuntu.
After using Virtualmin to switch a particular virtual server from PHP 5.5 to PHP 5.6, I received a 500 Internal Server Error and had the same entries in the apache error log:
[Tue Jul 03 16:15:22.131051 2018] [fcgid:warn] [pid 24262] (104)Connection reset by peer: [client 10.20.30.40:23700] mod_fcgid: error reading data from FastCGI server
[Tue Jul 03 16:15:22.131101 2018] [core:error] [pid 24262] [client 10.20.30.40:23700] End of script output before headers: index.php
Cause
Simple: I didn't install the php5.6-cgi packet.
Fix
Installing the packet and reloading apache solved the problem:
sudo apt-get install php5.6-cgi if you are using PHP 5.6
sudo apt-get install php5-cgi if you are using a different PHP 5 version
sudo apt-get install php7.0-cgi if you are using PHP 7
Then use service apache2 reload to apply the configuration.
The famous Moodle "replace.php" script can generate this situation too.
For me it was taking ages to run and then failed with a 500 message in the browser and also with the above error message in my apache error log file.
I followed up on #james-wise answer:
FcgidBusy is readably described in the Apache documentation. I tried this: doubled the amount of time which apache would give my script to run, by inserting the following line in /etc/apache2/mods-available/fcgid.conf
FcgidBusyTimeout 600
Then I restarted Apache and tried to run my replace.php script again.
Fortunately this time the script instance ran to completion, so for my purposes this served as a solution.
I came across this one while debugging a virtualmin/apache related error.
In my case, I am running virtualmin and had in my virtual machine's php.ini
safe_mode=On.
In my Virtual Machine's error log, I was getting the fcgi Connection reset by peer: mod_fcgid: error reading data from FastCGI server
In my main apache error log I was getting:
PHP Fatal error: Directive 'safe_mode' is no longer available in PHP in Unknown on line 0
In my case, I simply set safe_mode = Off in my php.ini and restarted apache.
stackoverflow.com/questions/18683177/where-to-start-with-deprecated-directive-safe-mode-on-line-0-in-apache-error
Not in this questions askers case but often:
What does the "premature end of script headers" error mean?
That error means that the FCGI call was exited unexpectedly.
In some cases it means that the script "backup.php" did crash.
How to fix this?
If the crash of a script was the cause, fix the script so that it does not crash. Then this error is fixed, too. To find out if and why a script crashes, you need to debug it. For example you can check the PHP error log. Errors logged to STDERR normally go into the error handler of the FCGI.
I had the same problem with long-running scripts with the error messages
"Premature end of script headers: index.php" and "Connection reset by peer: mod_fcgid: error reading data from FastCGI server" in error_log.
After hours of testing this helps for me (CentOS 6, PHP-FPM 7, Plesk 12.5.30):
edit the config file:
/etc/httpd/conf.d/fcgid.conf
Set a higher running time. In my case 600 seconds
create the new entry:
FcgidBusyTimeout 600
adapt following entries:
FcgidIOTimeout 600
FcgidConnectTimeout 600
restart httpd:
service httpd restart
In CentOS releases suexec is compiled to run only in /var/www. If you try to set a DocumentRoot somewhere else you have to recompile it - the error in apache log are:
(104)Connection reset by peer: mod_fcgid: error reading data from FastCGI server
Premature end of script headers: php5.fcgi
Just install php5-cgi
in debian
sudo apt-get install php5-cgi
in Centos
sudo yum install php5-cgi
Check /var/lib/php/session and its permissions. This dir should be writable by user so the session can be stored
As already mentioned this could be happening due to fcgi handler permission issues. If you're using suexec - don't forget to check if apache has this module enabled.
I increased max execution time to 600 seconds job done !
If you're on a shared server like me the host said it was a result of hitting memory limits, so they kill scripts which results in the "Premature end of script headers" seen in this error. They referred me to this:
https://help.dreamhost.com/hc/en-us/articles/216540488-Why-did-procwatch-kill-processes-on-my-Shared-serv
Given an increase in memory, the issues went. I think a backup plugin Updraft on wordpress was perhaps over zealous in its duty/settings.
In my case I was using a custom extension for my PHP files and I had to edit /etc/apache2/conf-available/php7.2-fpm.conf and add the following code:
<FilesMatch ".+\.YOUR_CUSTOM_EXTENSION$">
SetHandler "proxy:unix:/run/php/php7.2-fpm.sock|fcgi://localhost"
</FilesMatch>
I've tried the majority of answers that I've found on this issue. My issue was with wp-cron.php executing a particular function.
I'm working on Plesk CentOS7, Apache server.
I used this related question and suggested answer to help me find out how to adjust the fcgid.conf memory limit utilizing the command line.
Upon trying to troubleshoot the limits on the fcgid.conf file (/etc/httpd/conf.d/fcgid.conf) and restarting apache gracefully, I found that there was no change.
After seeing that other cron jobs were running properly, I decided to refer back to my function declared in functions.php and found that some of the arguments declared were not being specified properly, so there was a loop occurring that would eventually lead to the timeout.
Upon fixing this and running the cron again, it ran as it should.
Hope this helps someone else in a similar position!
I got the same problem (with Plesk 12 installed).
However, when i switched from execute PHP as FastCGI to Apache Module the website worked.
Checked my suexec log:
$ cd /var/log/apache2/
$ less suexec.log
When you find something like this:
[2015-03-22 10:49:00]: directory is writable by others: (/var/www/cgi-bin/cgi_wrapper)
[2015-03-22 10:49:05]: uid: (10004/gb) gid: (1005/1005) cmd: cgi_wrapper
try this commands
$ chown root:root /var/www/cgi-bin/cgi_wrapper
$ chmod 755 /var/www/cgi-bin/cgi_wrapper
$ shutdown -r now
as root.
I hope it can help you.