Bug 1315106

Description of problem:
All `dnf ...` commands stall with no output.  Some rpm commands also stall.

Version-Release number of selected component (if applicable):
`dnf --version` stalls

How reproducible:
every time (dnf was working fine 1-2 months ago...)

Steps to Reproduce:
# dnf -v help
[waited 5 minutes for anything to happen, then forced to background: ^Z]
# kill -KILL `pgrep dnf`
[1]+  Killed                  dnf -v help
# cat /var/log/dnf.*.log /var/log/dnf.log
# rm -fr /var/cache/dnf
# dnf -v help
  [new terminal]
  # netstat --wide --numeric --program --timers --all|grep -e `pgrep dnf` -
  (standard input):unix  3      [ ]         STREAM     CONNECTED     480908   21698/python3        
[waited 5 minutes for anything to happen, then forced to background: ^Z]
# kill -KILL `pgrep dnf`
[1]+  Killed                  dnf -v help
# cat /var/log/dnf.*.log /var/log/dnf.log
# rm -fr /var/cache/dnf
# rpm -q librepo curl dnf
  [new terminal]
  # top -p `pgrep rpm`
  [shows rpm always in a sleeping state with no cpu usage: ^C]
[waited 5 minutes for anything to happen, then forced to background: ^Z]
# kill -KILL `pgrep rpm`
[1]+  Killed                  rpm -q librepo curl dnf
# cat /etc/redhat-release
Fedora release 23 (Twenty Three)



Expected results:
No stalls, some cpu activity, any spawned processes to place the blame on, some relevant error messages in any log files...