PHP 5.3.3 cgi and fast-cgi broke down after adding php 5.5.13

JasperV · Jul 16, 2014

Hello Parallels Forum,

According to de Plesk Administrator Guide and according to the kb article: http://kb.parallels.com/en/118378, Plesk Panel is suitable to enable multiple php versions After I performed the steps mentioned in the kb article, all seemed fine until somehow php-cgi and php-fcgi broke down after

Code:

# service httpd restart

After this event, all websites hosted on this particular the server only work running php as an apache module.

I am running Pleks Panel 12.0.18 #7 on top of CenOs 6.5 on this particular server, which means php version 5.3.3 is installed by default. PHP 5.3.3 also is was the php version for the "custom: false" php handlers (id's: cgi, fastcgi, module). Somehow all php handlers look like this:

Code:

# /usr/local/psa/bin/php_handler --list
      id:  display name:  full version:  version:    type:                           cgi-bin:  php-cli:                           php.ini:  custom:
      cgi         5.5.13         5.5.13       5.5      cgi                   /usr/bin/php-cgi                                 /etc/php.ini    false
  fastcgi         5.5.13         5.5.13       5.5  fastcgi                   /usr/bin/php-cgi                                 /etc/php.ini    false
   module         5.5.13         5.5.13       5.5   module                   /usr/bin/php-cgi                                 /etc/php.ini    false
    php55         5.5.13         5.5.13       5.5  fastcgi  /usr/local/php-5.5.13/bin/php-cgi            /usr/local/php-5.5.13/etc/php.ini     true

What did I do wrong? What logs do I need to analyse? What to resolve to make the php handlers work again?

Maybe relevant to this situation is that the php mailer has broke down as well. Stating:

Code:

Reporting-MTA: dns; myserver.domain.tld
X-Postfix-Queue-ID: 41C4023C2D
X-Postfix-Sender: rfc822; [email protected]
Arrival-Date: Wed, 16 Jul 2014 03:35:40 +0200 (CEST)

Final-Recipient: rfc822; [email protected]
Original-Recipient: rfc822; [email protected]
Action: failed
Status: 5.7.1
Diagnostic-Code: X-Postfix; Command rejected

Note that php-fpm support for nginx won't start either. Giving the error message:

Code:

Error: Unable to make action: Unable to manage service by phpinimng: [16-Jul-2014 21:29:05] ERROR: unable to bind listening socket for address '127.0.0.1:9000': Address already in use (98)
[16-Jul-2014 21:29:05] ERROR: FPM initialization failed
Failed to start php-fpm service
('--type=fpm', 'php-fpm')

I sincerely hope that someone can help me out.

JasperV · Jul 17, 2014

php mailer has broke down as well

JasperV said:

Maybe relevant to this situation is that the php mailer has broke down as well. Stating:

Code:

Reporting-MTA: dns; myserver.domain.tld
X-Postfix-Queue-ID: 41C4023C2D
X-Postfix-Sender: rfc822; [email protected]
Arrival-Date: Wed, 16 Jul 2014 03:35:40 +0200 (CEST)

Final-Recipient: rfc822; [email protected]
Original-Recipient: rfc822; [email protected]
Action: failed
Status: 5.7.1
Diagnostic-Code: X-Postfix; Command rejected

Can be solved with the solution in post 1736083 @ thread: http://forum.parallels.com/showthre...-of-wrong-HELO&p=736083&viewfull=1#post736083

igor.A. said:
remove whole non_smtpd_milters from postfix main.cf in case you do not use commtouch.
If you would like to use commtouch then you can remove only "inet:127.0.0.1:12768" from non_smtpd_milters.
restart postfix.

JasperV · Jul 17, 2014

Plesk version 12.0.18 Update #8

The related server has been updated to: Plesk version 12.0.18 Update #8

JasperV · Jul 19, 2014

Is there anyone having any thoughts related to this issue?
Any comment would be appreciated.

Thanks in advance.

JasperV · Jul 25, 2014

http://kb.parallels.com/114601 does not seem to apply. Alle VPS optimized componenst are off:

Code:

# /usr/local/psa/bin/vps_optimized --status
VPS Optimized mysql-innodb: off
[B]VPS Optimized apache-modules-all: off[/B]
VPS Optimized apache-modules-except-php: off
VPS Optimized apache-modules-except-scripting: off

http://kb.parallels.com/120785
Symptom 1: A secondary PHP is not working while switching the PHP mode to fastcgi

Code:

tailf /var/www/vhosts/domain.tld/logs/error_log

(104)Connection reset by peer: mod_fcgid: error reading data from FastCGI server
Premature end of script headers: index.php

is appearing at the moment:

Code:

# tailf /var/www/vhosts/lesstof.net/logs/error_log
[Fri Jul 25 09:44:25 2014] [warn] [client 37.153.249.244] (104)Connection reset by peer: mod_fcgid: error reading data from FastCGI server 
[Fri Jul 25 09:44:25 2014] [error] [client 37.153.249.244] Premature end of script headers: index.php

Symptom 2: [Errno 12] Cannot allocate memory
does not seem to apply. Error message does not appear on server.

Symptom 3: The Suexec log shows an error:

Code:

# tailf /var/log/httpd/suexec_log
[2014-07-25 08:15:44]: uid: (10014/lesstof) gid: (505/505) cmd: cgi_wrapper
[2014-07-25 08:23:45]: uid: (10014/lesstof) gid: (505/505) cmd: cgi_wrapper
[2014-07-25 08:29:29]: uid: (10014/lesstof) gid: (505/505) cmd: cgi_wrapper
[2014-07-25 08:37:07]: uid: (10014/lesstof) gid: (505/505) cmd: cgi_wrapper
[2014-07-25 08:50:15]: uid: (10014/lesstof) gid: (505/505) cmd: cgi_wrapper
[2014-07-25 08:53:50]: uid: (10014/lesstof) gid: (505/505) cmd: cgi_wrapper
[2014-07-25 08:57:01]: uid: (10014/lesstof) gid: (505/505) cmd: cgi_wrapper
[2014-07-25 09:07:04]: uid: (10014/lesstof) gid: (505/505) cmd: cgi_wrapper
[2014-07-25 09:07:47]: uid: (10014/lesstof) gid: (505/505) cmd: cgi_wrapper
[2014-07-25 09:11:04]: uid: (10014/lesstof) gid: (505/505) cmd: cgi_wrapper

Error: unknown error in docroot check (10033)

does not appear in the log.

Resolutions:
1) PHP is not compiled with fastcgi. ==> PHP is compiled with fastcgi

Code:

# /usr/local/php-5.5.13/bin/php-cgi -v
PHP 5.5.13 (cgi-fcgi) (built: Jul 12 2014 14:24:50)
Copyright (c) 1997-2014 The PHP Group
Zend Engine v2.5.0, Copyright (c) 1998-2014 Zend Technologies

2) There isenough memory on the server.
3) The domain php.ini is customized, does not contains misprinted errors or an invalid configuration.
4) Permissions on the /etc/passwd file are correct.

JasperV · Jul 25, 2014

http://kb.parallels.com/115704 does not apply. Proper permissions and ownership settings on /var/lib/php/session

Code:

$ cd /var/lib/php/
$ ls -l
total 14468
drwx-wx-wt. 2 root root 14811136 Jul 23 07:01 session

JasperV · Jul 25, 2014

http://kb.parallels.com/116801. not having any of the symptoms.

Code:
Symptoms
I cannot open the settings of a subscription in the Parallels Plesk Panel (Plesk) control panel. I get this error:

Code:

Unable to create or update the subscription: The PHP handler "module" is not registered. It was probably deleted from the server. Register a handler with the same properties and try again.

Or I get the following error:

Code:

Unable to register the PHP handler: Handler with id = `fastcgi' already exists

Or this one:

Code:

Unable to create or update the subscription: The PHP handler "" is not registered

JasperV · Jul 27, 2014

After opening a Paralles Support ticket and a long session with a couple of support engineers, this issue was fixed by re-installing php 5.3.3 as default.

The assessment was that somehow the default php 5.3.3 installation got overwritten by the php 5.5.13 installation, causing the problems decribed.

Case closed.

PHP 5.3.3 cgi and fast-cgi broke down after adding php 5.5.13

JasperV

Basic Pleskian

JasperV

Basic Pleskian

JasperV

Basic Pleskian

JasperV

Basic Pleskian

JasperV

Basic Pleskian

JasperV

Basic Pleskian

JasperV

Basic Pleskian

JasperV

Basic Pleskian

Similar threads