# Problems with Fedora 22 and Mathematica 10.2/10.3

I'm using Fedora 22 Linux distribution and since I switched to Mathematica 10.2 I'm experiencing some very weird behavior of the frontend. From time to time, the frontend would freenze for a few seconds because WolframKernel process eats almost 100% of the CPU time. The strange thing is that it happens when nothing is evaluated, i.e. the frontend is actually idle. Also, these freezes are not really reproducable, since they appear to occur quite randomly, but usually 5-10 minutes after I start working with Mathematica.

Mathematica 10.0.2 is not affected, while Mathmeatica 10.3 behaves the same way as 10.2.

I'm pretty sure that this is somehow related to Fedora, because since the last two Fedora package updates the situation has become even worse. Now, sometimes the kernel would start eating 100% CPU time for unlimited time. If I kill the kernel with "kill -9 WolframKernel", then the its the frontend that starts to consume the CPU, so there is really no help.

I'm not doing anything fancy with my Linux distibution. It's just a standard Fedora installation + some standard packages for coding and LaTeX, but nothing really special and no self-compiled kernels.

The reason I'm posting this question here is that I don't really expect Wolfram support to help me debugging this very specific problem (may be I'm really the only one affected), but possibly someone from the community has encountered something similar?

I'm also wondering if there is a way to diagnose what Mathematica kernel is actually doing while it is consuming all of my CPU time? If there would be a possibility to squeeze out some debugging output out of it, then I could provide it to the support and this would probably help to solve my problem.

Here is the list of packages from the two last updates. Is there something among them that is very important for Mathematica?

1)

Aktualisiert  git-2.4.3-6.fc22.x86_64              @updates
Aktualisiert  libebml-1.3.1-1.fc22.x86_64          @fedora
Aktualisiert  libmatroska-1.4.2-1.fc22.x86_64      @fedora


2)

Aktualisiert  byobu-5.92-1.fc22.noarch                           @fedora
Aktualisiert  netcf-libs-0.2.8-1.fc22.x86_64                     (unknown)
Aktualisiert  opus-1.1.1-0.2.beta.fc22.x86_64                    (unknown)
Aktualisiert  rpm-4.12.0.1-12.fc22.x86_64                        (unknown)
Aktualisiert  rpm-build-4.12.0.1-12.fc22.x86_64                  (unknown)
Aktualisiert  rpm-build-libs-4.12.0.1-12.fc22.x86_64             (unknown)
Aktualisiert  rpm-libs-4.12.0.1-12.fc22.x86_64                   (unknown)
Aktualisiert  rpm-plugin-selinux-4.12.0.1-12.fc22.x86_64         (unknown)
Aktualisiert  rpm-plugin-systemd-inhibit-4.12.0.1-12.fc22.x86_64 (unknown)
Aktualisiert  rpm-python-4.12.0.1-12.fc22.x86_64                 (unknown)
Aktualisiert  rpm-python3-4.12.0.1-12.fc22.x86_64                (unknown)

• Which kernel is using 100% CPU? There should be two kernel processes. One does the computation and the other is used by the front end for various tasks, e.g. auto-completion. You can find out which one is the compute kernel in many different ways, e.g. by checking $ProcessID. – Szabolcs Nov 14 '15 at 21:30 • @b.gatessucks Thanks. Now I'm really wondering what's wrong with my system :( @Szabolcs I believe that this is the one used by the front end, since the freezes usually occur when nothing is evaluated. Unfortunately, once a freeze starts, the front end stops responding, so I can't evaluate anything there. But I can just get $ProcessID at the start and then wait for the freeze to occur. – vsht Nov 14 '15 at 21:58
• Now I can confirm that it's the "front end kernel" that goes crazy. I removed everything from my .Mathematica directory, except for Licensing and Applications. Let's see if it helps. – vsht Nov 15 '15 at 21:17