Question Mail server is not working till restart

[MrPleskLearner]

Server operating system version

Ubuntu 22.04

Plesk version and microupdate number

18.0.56

Hello everyone,

my web pages are working without any issue.

But somehow sometimes my webmail.mywebpageadress.com is not working or mails are not coming. After i restart the server via plesk everything is working again. I feel like, i need to restart server every day, if not every 2 days.

What can be the problem?

 

[Peter Debik]

Could you please describe in more detail what symptoms you mean by "webmail... is not working"? Can you not login? Is an error message displayed when trying to login? Which error message? Can you not see the login page at all? Or can you login, but expected mails do not show up in the mailbox?

 

[MrPleskLearner]

Could you please describe in more detail what symptoms you mean by "webmail... is not working"? Can you not login? Is an error message displayed when trying to login? Which error message? Can you not see the login page at all? Or can you login, but expected mails do not show up in the mailbox?

I cannot even reach the webmail page while my webpages are working. After restarting my server, webmail works again.

 

[Peter Debik]

This symptom can occur for a number of reasons. My guess is that PHP processes that your webmail uses, got stuck. To find out more about it, we'll first need to know which PHP and interface webmail is using. The default configuration is located in
/usr/local/psa/admin/conf/templates/default/webmail/roundcube.php
But maybe you have a custom configuration in
/usr/local/psa/admin/conf/templates/custom/webmail/roundcube.php
Which one is correct?

In that file you will find a line similar or equal to
FcgidInitialEnv PP_CUSTOM_PHP_CGI_INDEX [B]plesk-php74-fastcgi[/B]
with the last parameter indicating the PHP version and interface. Which is the one you are using?
See details on options here: https://support.plesk.com/hc/en-us/...e-PHP-version-for-webmail-in-Plesk-for-Linux-

If you are using FastCGI, then it is possible that all FastCGI processing slots are "full" by PHP processes so that no new connect can be made. This would result in an unresponsive webmail page. You could check this situation by running ps aux | grep roundcube and simply checking if there are "many" such processes related to roundcube. "many" is relative. Maybe you can check how your Apache web server is configured, to find out, what it will accept as the maximum number of concurrent FCGI requests. This information is stored in /etc/httpd/conf.d/fcgid.conf, for example:

FcgidMaxProcesses 100
FcgidMaxProcessesPerClass 10

Maybe you want to increase these values a little? Let's say FcgidMaxProcesses to 500 and FcgidMaxProcessesPerClass to 40? You can tell from the number of concurrent webmail processes you see if this could help. If you change the fcgid.conf settings, restart Apache afterwards systemctl restart apache2.

Another approach could be to change the handler from Fastcgi to FPM. How to do that can be derived from the above mentioned KB article.

 

[MrPleskLearner]

This symptom can occur for a number of reasons. My guess is that PHP processes that your webmail uses, got stuck. To find out more about it, we'll first need to know which PHP and interface webmail is using. The default configuration is located in
/usr/local/psa/admin/conf/templates/default/webmail/roundcube.php
But maybe you have a custom configuration in
/usr/local/psa/admin/conf/templates/custom/webmail/roundcube.php
Which one is correct?

In that file you will find a line similar or equal to
FcgidInitialEnv PP_CUSTOM_PHP_CGI_INDEX [B]plesk-php74-fastcgi[/B]
with the last parameter indicating the PHP version and interface. Which is the one you are using?
See details on options here: https://support.plesk.com/hc/en-us/...e-PHP-version-for-webmail-in-Plesk-for-Linux-

If you are using FastCGI, then it is possible that all FastCGI processing slots are "full" by PHP processes so that no new connect can be made. This would result in an unresponsive webmail page. You could check this situation by running ps aux | grep roundcube and simply checking if there are "many" such processes related to roundcube. "many" is relative. Maybe you can check how your Apache web server is configured, to find out, what it will accept as the maximum number of concurrent FCGI requests. This information is stored in /etc/httpd/conf.d/fcgid.conf, for example:

FcgidMaxProcesses 100
FcgidMaxProcessesPerClass 10

Maybe you want to increase these values a little? Let's say FcgidMaxProcesses to 500 and FcgidMaxProcessesPerClass to 40? You can tell from the number of concurrent webmail processes you see if this could help. If you change the fcgid.conf settings, restart Apache afterwards systemctl restart apache2.

Another approach could be to change the handler from Fastcgi to FPM. How to do that can be derived from the above mentioned KB article.

/usr/local/psa/admin/conf/templates/default/webmail/roundcube.php is the correct one, i don`t have folder named "custom".

I don`t use roundcube, i like horde. I opened horde.php via "vi" command and inside looks like this:



I could not find /etc/htppd folder. I am using ubuntu. But i found /etc/apache2/mods-available/fcgid.conf



Should i change FcgidMaxProcesses 20 to 500 in this file?

Thank you very much for your help!

 

[Peter Debik]

I could not find /etc/htppd folder. I am using ubuntu. But i found /etc/apache2/mods-available/fcgid.conf
Should i change FcgidMaxProcesses 20 to 500 in this file?

yes, also change FcgidMaxProcessesPerClass to something like 15 or 20.

 

[MrPleskLearner]

Looks like, i don't have any issue anymore. Thank you so much @Peter Debik !