DirectAdmin process high CPU loading every minute

tim874536

tim874536

Verified User
Joined
Nov 24, 2006
Messages
71
Hi All,

I am DirectAdmin big fans and never using other panel in production environment.

Recently, I am migrating servers into newly deployed Directadmin server due to CentOS 6 EOL
I find that both two new servers' DirectAdmin process keep having high CPU loading every minute.
Therefore I would like to look for help on how to identify the issue and resolve them.

During the Directadmin process high CPU loading, none of the new log can be found inside /var/log/directadmin/*

I attached two screen captures.
Htop_DirectAdmin_process.PNG is the htop screen showing number of DirectAdmin process when high loading start. The loading takes end after 10-15 seconds. This high loading happens every minute.
Htop_DirectAdmin_process.PNG
Htop_DirectAdmin_strace.PNG is the strace of the top loading directadmin process. The process seems keep looping the same thing as shows in the strace screen capture.
Htop_DirectAdmin_strace.PNG

Would you please advise me how I can identify the issue?
Let me know if any further information is needed.

Thank you very much.

Here is the detail environment:
  • Proxmox VM platform, KVM virtualization, SSD Storage
  • CloudLinux 8.7, CageFS, alt-php, MySQL Governor enabled
  • DirectAdmin: 1.653 build c08a1d88b28d4253a8664310ba7429e884c3f4db
  • Apache: 2.4.57
  • Pure-FTPd: 1.0.51
  • Redis: 7.2.1
  • WP-CLI: 2.8.1
  • Imapsync: 2.229
  • mod_lsapi: 1.1-68
  • mod_proctitle: 0.1-7
  • mod_hostinglimits: 1.0-41
  • Dovecot: 2.3.21
  • dovecot.conf: 0.4
  • Exim: 4.96.1
  • Lego: 953d5c85145b6a2b9a52f2d919faf23e04a359b3
  • exim.conf: 4.5.44
  • BlockCracking: 1.12
  • Easy Spam Fighter: 1.32
  • Rspamd Config: 0.4
  • MariaDB: 10.6.15
  • PHP 8.1: 8.1.23
  • RoundCube: 1.6.3
  • phpMyAdmin: 5.2.1-all-languages
 
Hi all,

Looking forward to your feedback and getting some hints to identify and resolve the issue.

Thank you very much.
 
After change the dataskq_run_interval=1m to dataskq_run_interval=30s, the DirectAdmin high loading for every 30 seconds.

When using the second links for debug ( /usr/local/directadmin/dataskq d2000 ), it shows the following
Code: 
# /usr/local/directadmin/dataskq d2000
Debug mode. Level 2000

get_process_list_from_status: running: /usr/bin/systemctl status clamav-freshclam.service 2>&1
get_process_list_from_status: running: /usr/bin/systemctl status [email protected] 2>&1
get_process_list_from_status: running: /usr/bin/systemctl status rspamd.service 2>&1
Ssl::get_admin_ssl_poll_domain_list:user1:xxxxx.com: domain does not have ssl enabled.
Ssl::get_admin_ssl_poll_domain_list:user2:yyyyy.com: domain does not have ssl enabled.
Ssl::get_admin_ssl_poll_domain_list:user3:zzzzz.com: domain does not have ssl enabled.
....... # Skip output for all user # ......
....... # Skip output for all user # ......
Ssl::get_admin_ssl_poll_domain_list:zyxuser:zyxuser.com: domain does not have ssl enabled.
Ssl::admin_retry_domains: About to work on 0 values.
#
 
How many RAM do you have and have you create a swap partition on that VM too? It's required.
I ask for the RAM as I get the impression that the high load might be caused by AutoSSL.
 
Richard G said:
How many RAM do you have and have you create a swap partition on that VM too? It's required.
I ask for the RAM as I get the impression that the high load might be caused by AutoSSL.
Click to expand...
# free -m
total used free shared buff/cache available
Mem: 15631 5392 189 409 10048 9398
Swap: 16379 5559 10820


I just disabled AutoSSL by the following commands and the dataskq complete within 2 seconds when compare with 10-15 seconds CPU high loading before.
Code: 
/usr/local/directadmin/directadmin set admin_ssl_check_retries 0
service directadmin restart
 
tim874536 said:
and the dataskq complete within 2 seconds when compare with 10-15 seconds CPU high loading before.
Click to expand...
Should be enough ram, swap is also enough.
So yes then indeed it's caused by the AutoSSL with that time difference. I just don't know how to fix this in another way.
 
There is another problem that happening. Once disabled admin_ssl_check_retries, the normal user SSL that does not used AutoSSL before is now not renewing anymore. Many of the client report their website / SSL was expired these two weeks.
I have to enable admin_ssl_check_retries again (However, the directadmin process keep loading CPU every minute again)

I have to manual renew SSL by the following post script as well.

How to force renew all Let's Encrypt cert?

I just got this mail "We recently discovered a bug in the Let's Encrypt certificate authority code, described here: https://community.letsencrypt.org/t/2020-02-29-caa-rechecking-bug/114591 Unfortunately, this means we need to revoke the certificates that were affected by this bug, which...
forum.directadmin.com forum.directadmin.com
 
ikkeben said:
Any antivirus ?
Click to expand...
Not antivirus related.
It is DirectAdmin and SSL related as said in my #4 post

DirectAdmin process high CPU loading every minute

Hi All, I am DirectAdmin big fans and never using other panel in production environment. Recently, I am migrating servers into newly deployed Directadmin server due to CentOS 6 EOL I find that both two new servers' DirectAdmin process keep having high CPU loading every minute. Therefore I...
forum.directadmin.com forum.directadmin.com
 
You must log in or register to reply here.
Back
Top