spamd processes stuck causing high load

[stormy]

We are seeing this for the last 2-3 days on two different servers. Multiple spamd processeses by a user that has little or no incoming mail, taking up 100% of CPU usage in top.

No matter what we try, the processes reappear. We have tried deleting the bayes databases for the user, restarting exim several times, suspending and unsuspending the account, killing the processes, nothing works.

Any ideas?

 

[cPanelMichael]

Hello

Please review the following thread and let us know if it's related:

'spamd' bottleneck after update

Thank you.

 

[stormy]

I'm familiar with that thread. I was suffering the "spamd failed" problem a few weeks back. I did disable the KAM rules just in case. cPanel support determined it was not related, and added maxspare=7 to /etc/cpspamd.conf in order to deal with this. This increased load a bit too much so I lowered it to maxspare=4.

That was a few weeks back, and everything was normal until a couple days.

However, the server that is most affected by this problem today wasn't affected by this other problem, and has no custom spamd configuration.

 

[cPanelMichael]

Please verify the version of cPanel installed on your system, and if any of the following options are enabled in "WHM >> Exim Configuration Manager >> Basic Editor":

Enable BAYES_POISON_DEFENSE Apache SpamAssassin™ ruleset
Enable KAM Apache SpamAssassin™ ruleset
Enable the Apache SpamAssassin™ ruleset that cPanel uses on cpanel.net

Thank you.

 

[stormy]

I'm on WHM 11.52.2 (build 4).

• BAYES_POISON_DEFENSE is on
  
• KAM is off
• SpamAssassin™ ruleset that cPanel is on

 

[cPanelMichael]

Could you verify if the issue continues with the additional rulesets disabled?

Thank you.

 

[stormy]

I disabled all three and haven't seen the issue reoccur today. cPanel support had also advised to add this to the spamd configuration, though it didn't solve the issue:
maxchildren=10
timeoutchild=120
timeouttcp=60