4

Recently I asked this question and this one about systemd and a few related processes using 100% CPU just 5 minutes after every boot.

I have a quadcore Intel processor, each core has two virtual cores for a total of 8 CPU cores. Three or more of these being at 100%+ utilisation causes problems for my ancient laptop, namely overheating, and to preserve itself it powers off under such temperatures.

However, I noticed the system runs totally normally until a few minutes after I -- my user -- logs in - only then does systemd go haywire, and logging out does not stop it.

What's causing this? What about my user's login specifically would cause this?

It only happens when my account, cat, logs in: not the guest session, not root in a TTY or graphical session -- I logged in to a virtual TTY (Ctrl-Alt-F1) as root and the problem hasn't happened so far.

from a small amount of googling the problem seems to be like these two:

and a lot of the rest of these google results have information pertaining to my problem but no real solution, just ancient bugfixes which are surely in upstream already.

This is not a one time thing, it's 5 or fewer minutes after logging in as myself, every time.

What's going on with systemd after I log in?


Incidentally, this is Ubuntu 15.10 (functionally Debian), and all my packages are up to date. Some logs / examples:


systemctl -al. (long)

journalctl | tail -n400. (less long, but long)


Here are some examples of this happening again (right now):

  PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND                                                                                                                      
    1 root      20   0  185480   6080   3900 R 100.0  0.1   1:27.54 systemd                                                                                                                      
  699 message+  20   0   44316   5148   3500 R  93.7  0.1   1:25.50 dbus-daemon                                                                                                                  
  300 root      20   0   34716   4400   3964 R  56.2  0.1   0:49.01 systemd-journal                                                                                                              
  848 syslog    20   0  256372   4288   2648 S  43.7  0.1   0:33.05 rsyslogd                                                                                                                     
  835 root      20   0   28628   3204   2816 R  18.7  0.1   0:22.72 systemd-logind                                                                                                               

or

    1 root      20   0  186928   7632   3988 R  99.9  0.1   5:37.49 systemd                                                         
  806 message+  20   0   44212   5216   3540 R  99.6  0.1   5:34.16 dbus-daemon                                                     
  298 root      20   0   34712   6824   6360 R  53.1  0.1   3:05.28 systemd-journal                                                 

or

    1 root      20   0  186928   7632   3988 R 100.0  0.1   6:29.31 systemd                                                         
  298 root      20   0   34712   7784   7320 R  50.8  0.1   3:33.11 systemd-journal                                                 
 8165 root      20   0   90668  23408  12636 R  27.7  0.4   0:00.84 debtags                                                         
  789 root      20   0   28636   3156   2780 S  21.8  0.1   1:29.49 systemd-logind                                                  

A pretty picture:

enter image description here

  • 1
    If you'd like help, please provide as much information as you can(and as you think is needed) in this single post. – MatthewRock Mar 27 '16 at 23:37
  • Yeah asking for help then saying we need to go through their previous posts to get all the information they know we need is kind of a dick move. – Bratchley Mar 27 '16 at 23:38
  • 1
    Have you tried unmasking the units it's talking about in the journal? The CPU usage is probably due to systemd continually trying to (re)start services that are failing due to dependencies. Not sure why one would mask though services, tbh. – Bratchley Mar 28 '16 at 1:16
  • 1
    Regardless of how they got like that you might try to see if unmasking fixes your issue. For the other thing, the "in the body of the post" is the rule of thumb. For data that long, putting it into the question's body is just impractical. – Bratchley Mar 28 '16 at 1:40
  • 1
    What about another regular (non-guest, non-root), freshly created user account? – L. Levrel Mar 29 '16 at 10:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.