[neutron] Neutron server init script kills itself while stopping service

[Artem Panchenko]

  2014.1.1-5.0.2:
    VERSION:
      api: '1.0'
      astute_sha: 6db5f5031b74e67b92fcac1f7998eaa296d68025
      build_id: 2014-09-09_10-19-11
      build_number: '72'
      feature_groups:
      - mirantis
      fuellib_sha: b538a2baa1cfc3c345e6f7d52f5ca191abdd2b74
      fuelmain_sha: b1e46c8c3ac86a8eafac225507294226322160ea
      nailgun_sha: c18a21381843dffe807b254a4ff96eec259953cb
      ostf_sha: 2969c1ad443b632e815bb1f01149c3800cd7aa3a
      production: docker
      release: 5.0.2

During OS cluster update (5.0 -> 5.0.2) puppet doesn't able to restart neutron-server:

http://paste.openstack.org/show/109539/

Because 'stop' function inside init script (/etc/rc.d/init.d/neutron-server) kills itself using 'killall' command:

http://paste.openstack.org/show/109562/
http://paste.openstack.org/show/109568/

I think 'pgrep $prog-server &>/dev/null && killall $prog-server' should be replaced by 'pgrep -u neutron $prog-server &>/dev/null && killall -u neutron $prog-server', so /etc/init.d/neutron-server process which is owned by root will not be killed.

Also, this issue doesn't affect 5.0.1 or 5.1 releases, because their openstack-neutron packages doesn't contain such code inside init scripts.

[Sergey Vasilenko (xenolog)]

diplicate of https://bugs.launchpad.net/fuel/+bug/1362839