Version-Release number of selected component: policycoreutils-python-utils-2.9-1.fc30 Additional info: reporter: libreport-2.10.1 cmdline: /usr/bin/python3 -Es /sbin/semanage fcontext -a -t rpm_exec_t /usr/share/dnfdaemon/dnfdaemon-system crash_function: module exception_type: ImportError executable: /sbin/semanage interpreter: python3-3.7.4-5.fc31.x86_64 kernel: 5.2.8-200.fc30.x86_64 runlevel: N 5 type: Python3 uid: 0 Truncated backtrace: #1 [/usr/lib64/python3.7/site-packages/setools/diff/types.py:26] <module> #2 [/usr/lib64/python3.7/site-packages/setools/diff/bounds.py:25] <module> #3 [/usr/lib64/python3.7/site-packages/setools/diff/__init__.py:20] <module> #4 [/usr/lib64/python3.7/site-packages/setools/__init__.py:79] <module> #5 [/usr/lib/python3.7/site-packages/sepolicy/__init__.py:7] <module> #6 [/usr/lib/python3.7/site-packages/seobject.py:33] <module> #7 [/sbin/semanage:877] createCommandParser #8 [/sbin/semanage:947] do_parser #9 [/sbin/semanage:970] <module>
Created attachment 1625110 [details] File: backtrace
Created attachment 1625111 [details] File: cgroup
Created attachment 1625112 [details] File: cpuinfo
Created attachment 1625113 [details] File: environ
Created attachment 1625114 [details] File: mountinfo
Created attachment 1625115 [details] File: namespaces
Created attachment 1625116 [details] File: open_fds
(In reply to Mai Ling from comment #0) > Version-Release number of selected component: > policycoreutils-python-utils-2.9-1.fc30 ... > interpreter: python3-3.7.4-5.fc31.x86_64 > kernel: 5.2.8-200.fc30.x86_64 Interpret is from Fedora 31 while policycoreutils and kernel from Fedora 30. Do you know when it happened? Are you able to reproduce it again? What is the version of python3-setools - rpm -q python3-setools ?
python3-setools-4.2.2-1.fc31.x86_64 I think I booted with old kernel because I had an issue with new kernel. Will report if it ever occurs again. I don't remember what happened in this particular case, I usually report all stuff that pops up in the systray as crashes.
Thanks. Given that I'm not able to reproduce it and it's the only one reported bug with this particular backtrace I'm going to close this bug now. If you have more information in future, feel free to reopen or file a new bug.
FWIW, I just saw this exact traceback show up during a `dnf system-upgrade` from F30 to F31, during the upgrade of bind. When the machine comes up, I'll grab the relevant chunk of journal and attach it.
Created attachment 1632312 [details] `journalctl -b -1` snippet from F30->F31 upgrade
I have another report where the reporter states that it can be reproduced using the following steps: dnf install -y chromium dnf-plugin-system-upgrade dnf system-upgrade download -y --releasever=31 dnf system-upgrade reboot # wait for upgrade to finish... # look at /var/log/dnf.rpm.log
The problem seems to be related to the fact that setools-4.4.2 uses /usr/lib64/python3.7/site-packages/setools/policyrep.cpython-37m-x86_64-linux-gnu.so instead of /usr/lib64/python3.7/site-packages/setools/policyrep/ and it looks like that during the transaction are both modules available.
This just was reported again in bug 1777001 on current F31.
*** Bug 1889125 has been marked as a duplicate of this bug. ***
*** Bug 1889126 has been marked as a duplicate of this bug. ***
I wasn't able to fix it until now and given that it affects upgrades from unsupported systems I'm not going to try anymore. I'm sorry.