SOLVED [CPANEL-28306] webmail and/or phpmyAdmin stuck at loading screen

[kabatak]

I have an issue similar to this but I couldn't reply to that old topic anymore.

CSF/LFD reports something like this:

lfd on xxx: Excessive resource usage: xxx

Time: Sat Jul 6 16:02:00 2019
Account: xxx
Resource: Process Time
Exceeded: 84921 > 1800 (seconds)
Executable: /usr/local/cpanel/3rdparty/php/72/sbin/php-fpm
Command Line: php-fpm: pool user_xxx
PID: 2567226 (Parent PID:1722110)
Killed: No​

As replied by @cPanelMichael in that old topic, PHP-FPM doesn't have the feature of showing full path.

My question is, what are other ways to find out the full script path of that process that's being reported?

 

[cPanelMichael]

Hello @kabatak,

You could try searching the Apache access logs for the date/time included in the CSF notification to see if you notice access attempts on specific PHP scripts under the account. Here's an example of a command you could use:

grep "06/Jul/2019:16:02" /home/username/access-logs/* |awk -F'\"' '{print $2}' |sort |uniq -c |sort -nr |head

Replace "username" with the name of the cPanel account included in the CSF notification.

Thank you.

 

[kabatak]

Hi cPanelMichael Thanks for the reply however I couldn't find anything in the access-logs/ folder. The account is relatively new so even manually checking the logs inside that folder doesn't show anything related.

We have since have the end-customer message us about the issue, turns out it was phpMyAdmin. It appears to happen if phpMyAdmin is stuck in a "loading" screen (perhaps when it couldn't properly read a table or something like that). Those "php-fpm: pool user_xxx" processes appears to linger for days long after closing the phpMyAdmin browser. You have to manually kill the process so phpMyAdmin can be usable again. It can also be tricky if you kill the wrong process.

Thing is, we would've never found out if the end-customer didn't inform us. Would be nice if there's a way to know without manual guess work. Thank you.

 

[cPanelMichael]

Hello @kabatak,

Thank you for sharing the additional information. Can you confirm if the end-user was accessing a custom phpMyAdmin installation on their website? Or, are you referring to phpMyAdmin as accessible through the cPanel interface?

Thank you.

 

[kabatak]

It was the default phpMyAdmin from Cpanel.

 

[cPanelMichael]

Hello @kabatak,

. It appears to happen if phpMyAdmin is stuck in a "loading" screen (perhaps when it couldn't properly read a table or something like that). Those "php-fpm: pool user_xxx" processes appears to linger for days long after closing the phpMyAdmin browser. You have to manually kill the process so phpMyAdmin can be usable again. It can also be tricky if you kill the wrong process.

This is worth further investigation to rule out any potential defects in the product. Do you mind reporting this issue in a support ticket so we can take a closer look at the server? You can post the ticket number here once it's opened and I'll link this thread to it.

Thank you.

 

[cPanelMichael]

Hello @kabatak,

Can you confirm if your server uses cPanel & WHM version 82? You can use the following command to determine the version:

cat /usr/local/cpanel/version

If your server uses version 82, then it's likely this relates to internal case CPANEL-28306. This case was opened to address an issue where an automatic task to clear the Horde cache files upon the update to cPanel & WHM version 82 (ref: CPANEL-27452) fails to restart the cPanel PHP-FPM service. This can lead to features such as phpMyAdmin or Webmail failing to load. I'll monitor this case and update this thread with more information on it's status as it becomes available.

In the meantime, the temporary workaround is to restart the cPanel PHP-FPM service using the following command:

 /scripts/restartsrv_cpanel_php_fpm

Thank you.

 

[kabatak]

The version is 11.80.0.20, (server set to update on RELEASE tier).

Another thing I noticed is that this issue also caused the WHM backup scheduled process to pause for days unless I kill the related process.

Right now we are not experiencing the issue since the tasks was killed. Should I still open a ticket?

 

[cPanelMichael]

The version is 11.80.0.20, (server set to update on RELEASE tier).

Another thing I noticed is that this issue also caused the WHM backup scheduled process to pause for days unless I kill the related process.

Right now we are not experiencing the issue since the tasks was killed. Should I still open a ticket?

Hello @kabatak,

Yes, could you open a support ticket so we can take a closer look at your system and review the logs to see what could have lead to this issue? You can post the ticket number here and we'll link this thread to it.

Thank you.

 

[kabatak]

Just created a ticket 12861973, Thanks!

 

[cPanelMichael]

Hello,

To update, we were unable to reproduce the behavior reported as part of ticket 12861973.

Regarding case CPANEL-28306, this is fixed in cPanel & WHM version 82.0.3:

Fixed case CPANEL-28306: Restart FPM after clear_horde_cache task if it is enabled.

Let us know of any additional issues after updating to this version.

Thank you.