Potential FPM errors after update MU25 - solution

[trialotto]

@Everyone,

It can be the case that issues arise after micro-update #25, see symptoms below.

ISSUE: wrong versions of packages libapache2-mod-proxy-psa and psa-phpfpm-configurator

DESCRIPTION: the package versions, required or working with MU #25, are incorrect and should be equal to

libapache2-mod-proxy-psa-2.4.7-ubuntu14.04.16022012
psa-phpfpm-configurator-1.0.0-ubuntu14.04.build1205160218.11.amd64

and a similar issue (of incompatibility) can occur on other OSes, in which case the name will differ.

APPLIES TO: Ubuntu 14.04.x (every subversion)
NOTE: if I am not mistaken, this also applies to other OSes (if the package versions are not correct)

SYMPTOMS: one symptom of this FPM related issue is that pages, rendered by PHP script, are offered as a download, instead of being opened as a regulare web page.
SYMPTOMS: installation of WordPress can result in a "remarkable" error notification and the installation process will exit without installing the WordPress instance.

SOLUTION: do the following (in chronological order)

a) install the correct packages:

- open a SSH console, go to the directory /opt/psa/admin/sbin (on Ubuntu)
- run the command: ./autoinstaller
- click on N (a lot of times) and the packages will be updated to the appropriate version

b) restart ALL running FPM handlers:

- restart the "OS vendor" FPM handler: service php5-fpm restart
- check status of other FPM handlers: service plesk-php<xx>-fpm status
- if running, execute the command: service plesk-php<xx>-fpm restart

and note that

- the restart of FPM is REQUIRED (for many reasons, I will not elaborate on this),
- <xx> is one of the strings "55", "56" and "57",
- the commands are "Ubuntu" specific, but similar commands apply for other OSes.

Hope the above helps,

Regards....

 

[trialotto]

@Everyone,

If you run Plesk on a VPS, it can be the case that

SYMPTOM: web pages are not loading
SYMPTOM: "Nginx 502 gateway error" is occurring

after the MU25 update has been applied.

CAUSE: refused resource allocations and/or resource overusage (of some other kind)

SOLUTION: Reboot the VPS

NOTE: the file /proc/user_beancounters should be present, if you are using a Virtuozzo based VPS. Running the command

cat /proc/user_beancounters

in a (full screen) SSH console will give you a visible insight into the number of refused resource allocations (read: the last column, called "failcnt"). This number should be zero.

Hope the above helps a bit.

Regards......

 

[Kingsley]

@trialotto

[root@XXXX ~]# cat /proc/user_beancounters
cat: /proc/user_beancounters: No such file or directory

 

[IgorG]

@trialotto

[root@XXXX ~]# cat /proc/user_beancounters
cat: /proc/user_beancounters: No such file or directory

This command works only on Virtuozzo VPSes.

 

[Kingsley]

This command works only on Virtuozzo VPSes.

ok