Question How can sw-engine-fpm be restarted?

[Bitpalast]

I'd like to restart the sw-engine-fpm service. According to the process list

root 8736 0.0 0.0 501808 39028 ? Ss Oct01 0:00 sw-engine-fpm: master process (/etc/sw-engine/sw-engine-fpm.conf)
psaadm 27200 1.8 0.0 539288 71948 ? S 10:59 0:06 sw-engine-fpm: pool plesk

there should be a service named "sw-engine-fpm". But:

# service sw-engine-fpm status
Redirecting to /bin/systemctl status sw-engine-fpm.service
Unit sw-engine-fpm.service could not be found.

What's the right name?

 

[mizar]

Hello,

The service name is "sw-engine". In general, binary name is not same as service name, for instance mariadb.service executes mysql_safe binary

 

[Bitpalast]

Hmm, o.k., I already tried restarting sw-engine and other Plesk related services, but I feel that since a month sw-engine-fpm (the PHP-FPM part of the construction) is quite slow for unknown reason on this one machine. All transactions that sw-engine does on configuration changes are slow, although other processes (like database, webserver Apache and Nginx, other PHP modules, disk access etc.) are working fast as always. No database issues either. So I thought that restartig the PHP service that is responsible for sw-engine transactions must be the culprit.

Anyway, I'll continue to monitor it then.

 

[pleskpanel]

Hmm, o.k., I already tried restarting sw-engine and other Plesk related services, but I feel that since a month sw-engine-fpm (the PHP-FPM part of the construction) is quite slow for unknown reason on this one machine. All transactions that sw-engine does on configuration changes are slow, although other processes (like database, webserver Apache and Nginx, other PHP modules, disk access etc.) are working fast as always. No database issues either. So I thought that restartig the PHP service that is responsible for sw-engine transactions must be the culprit.

Anyway, I'll continue to monitor it then.

Anecdotally I feel that it could be age -

I get the feeling some days that the sw-engine on one of our servers has more creaks and takes longer to get up in the morning than it should despite all other services moving within expected operating parameters

 

[Bitpalast]

I should not have used the word "feel". It is observance, and the lack of speed is very clear. For example on all of our machines creating a new domain or subdomain takes a few seconds. On that very one machine it takes a minute or longer while the overall cpu load is somewhere around 0.5 on a 12-core machine. The long process times are not only limited to this example. They occur on basically all things where Plesk does a configuration change of any type.

 

[pleskpanel]

Just after that behavior occurs, is anything meaningful written to the /var/log/sw-cp-server/sw-engine.log file?

 

[Bitpalast]

Nothing in that file but the stop/start sequence with no errors. I also have check all other system files like /var/log/messages, the disk status. No database inconsistencies either. It all looks normal, only the Plesk operations are much slower than on other systems. It's not so bad that users start complaining, but there is a noticeable difference, and normally that means: problem.