Near-100% CPU utilization about docker-fail2ban HOT 7 CLOSED

Sure. When I was experiencing high CPU usage, the only env vars I had in my compose were:

      - TZ=America/Chicago
      - F2B_IPTABLES_CHAIN=DOCKER-USER
      - F2B_MAX_RETRY=10

When I added the rest of the vars (and changing the action) the CPU usage went down to normal levels:

      - TZ=America/Chicago
      - F2B_IPTABLES_CHAIN=DOCKER-USER
      - F2B_MAX_RETRY=25
      - F2B_DEST_EMAIL=*******@*****
      - F2B_SENDER=******@******
      - F2B_ACTION=%(action_mw)s
      - SSMTP_PORT=***
      - SSMTP_HOST=*********
      - SSMTP_USER=*********
      - SSMTP_PASSWORD=********
      - SSMTP_TLS=YES

Again, I cannot explain the why, as you can see the log data above shows absolutely no information as to why it would be using high CPU, but this seems to have fixed it.

from docker-fail2ban.

Seems adding additional environment variables corrected this problem.

from docker-fail2ban.

@bkraul can you be a little more specific just incase other people have this problem?

from docker-fail2ban.

I know this issue is old. Seems I kept having the same problem even with those config changes. The only thing that worked for me was to use compose version 2.2, and put hard resource limits on the container. It still shows as 100%, but it does not bring the host to its knees....

This is the internal command that is causing the high CPU:

/usr/bin/python /usr/bin/fail2ban-server -f -x -v start 

There are 2 threads with the same command and both are permanently 100% CPU. The stdout log returns no helpful information.

from docker-fail2ban.

@bkraul, I think this is an issue with fail2ban itself. often linked to the backend you use / active jails.
So first if you can try fail2ban outside Docker with all your jails / cfg.

I also updated the Docker image yesterday. Fail2ban now runs with Python 3.7 and you have a new variable to choose the backend (auto / pyinotify / polling).

from docker-fail2ban.

Downloaded the new image, and will test with it. Thanks man!

from docker-fail2ban.

Which option in F2B_BACKEND is the default (I know auto, but which one from the 2 is actually used by auto? I am still getting 100% CPU utilization. I've had to institute resource limits on the compose file for this...

  1  [|||                                                                               2.6%]   Tasks: 284, 1116 thr; 2 running
  2  [||||||                                                                            5.8%]   Load average: 1.65 0.92 1.50
  3  [|||||||||||||||||||||||||||||||                                                  35.7%]   Uptime: 00:26:30
  4  [||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||100.0%]
  Mem[|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||2.12G/7.79G]
  Swp[|                                                                           1.51M/976M]

  PID USER      PRI  NI  VIRT   RES   SHR S CPU% MEM%   TIME+  Command
29104 root       20   0 21908 17560  6096 R 100.  0.2  2:27.50 /usr/bin/python3 /usr/bin/fail2ban-server -f -x -v start

from docker-fail2ban.