Menu

#5047 /usr/share/webm 100% CPU

1.860
open
nobody
None
5
2017-12-04
2017-11-27
borja
No

Webmin process is always 100% cpu. Process /usr/share/webm is always at 93-100% (one core). Sometimes appear more than one instance and each one eating 90-100% of a core. So i have to restart webmin to free resources. Always have, at least, one core dedicated to webmin.

SO is debian stretch

Discussion

  • Jamie Cameron

    Jamie Cameron - 2017-11-27

    Does this happen even when you're not accessing webmin in your browser?

     
  • borja

    borja - 2017-11-27

    Hi Jamie,

    Yes, i restart webmin and after that check top and is already there. I think could be related to letsencrypt? Im unable to get a new certificate for one of my websites, it start to retrieve it but never finish. When trying to get again it creates another thread at 100%. Could be possible?

     
  • Jamie Cameron

    Jamie Cameron - 2017-11-28

    SSL shouldn't be related if the CPU load is high even when you're not using Webmin.

    What type of system are you running Webmin on there?

     
  • borja

    borja - 2017-12-02

    Problem is related to let'sencrypt module. I have a problem with one of my virtualservers and I cant generate certificate correctly.

    Script seems to be in an infinite loop, no erros, can be running for hours at 90-100% cpu.
    In browser only appears this: "Requesting a certificate for cloud.exampleserver.com from Let's Encrypt .."
    If i reload the page script is kept running in background. If i try to run the script eight time (number of cores of my CPU) it overloads my server. I have to restart webmin to destroy the processes or detroy them one by one. But one core is always at 100% after restarting and im pretty sure is trying to get certificate in background even after restarting webmin.

    I couldnt find letsencrypt log in /var/log/letscrypt, in my old debian 7 server (also webmin-virtualmin) is there.

    I get certificates for other virtual servers in the same server without problem, before and after i have discovered this problem. I dont know whats the problem because it not fail. no errors shown.

    Thanks for the help

     
  • borja

    borja - 2017-12-02

    I have discovered the problem, it was related to IPv6. It was configured on domain, but not in virtualserver. I think it should have a timeout but now is resolved.

    Thanks!

     
  • Jamie Cameron

    Jamie Cameron - 2017-12-04

    That Let's Encrypt bug will be fixed in the next Webmin release (1.870)

     

Log in to post a comment.

MongoDB Logo MongoDB