unstack.sh will not kill the n-cpu spawned in Ubuntu12.04
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
devstack |
Invalid
|
Undecided
|
yalei wang |
Bug Description
n-cpu is spawned by this command now if you use libvirtd:
sg libvirtd '/usr/local/
it runs in libvirtd group, and there will be two processes spawned for n-cpu.
$ ps aux |grep compu
root 28490 0.0 0.0 15028 1024 pts/20 S+ 14:18 0:00 sg libvirtd /usr/local/
neutron 28491 1.1 2.5 1851612 101432 pts/20 Sl+ 14:18 0:06 /usr/bin/python /usr/local/
and when u call unstack.sh, the stop_process in functions-common will be called . BUT
pkill -g 28490 will only kill 28490, and process 28491 will be left.
some error will emit, because that process will still occupy the source , like database(you can drop database if you use postgresql).
I use the ubuntu 12.04 and the pkill 's version is:
$ pkill -V
pkill (procps version 3.2.8)
and kill 28491 twice, is can be killed immediately.
Changed in devstack: | |
assignee: | nobody → yalei wang (yalei-wang) |
description: | updated |
Changed in devstack: | |
status: | New → Incomplete |
description: | updated |
description: | updated |
Changed in devstack: | |
status: | Incomplete → New |
the TERM signal will be handled in nova in function _wait_for_ exit_or_ signal in file nova/openstack/ common/ service. py
and will never return currently.