I have tested nova-api master too (as well as neutron) the following way:
1. Added logging to nova.service.reset method:
ash@ubuntu:/opt/stack/nova$ git diff diff --git a/nova/service.py b/nova/service.py index 3ae307c..5f4e952 100644 --- a/nova/service.py +++ b/nova/service.py @@ -366,6 +366,7 @@ class WSGIService(object): :returns: None
""" + LOG.exception("Calling reset!") self.server.reset()
def _get_manager(self):
2. Changed logging format in nova.conf to show pids
3. Started nova-api in daemon mode with the default number of workers (metadata_workers = 2, ec2_workers = 2, osapi_compute_workers = 2)
4. Verified that all processes are running:
ash@ubuntu:~/devstack$ ps -ef | grep nova-api ash 11999 9714 0 13:00 pts/10 00:00:04 /usr/bin/python /usr/local/bin/nova-api ash 12008 11999 0 13:00 pts/10 00:00:00 /usr/bin/python /usr/local/bin/nova-api ash 12009 11999 0 13:00 pts/10 00:00:00 /usr/bin/python /usr/local/bin/nova-api ash 12010 11999 0 13:00 pts/10 00:00:00 /usr/bin/python /usr/local/bin/nova-api ash 12011 11999 0 13:00 pts/10 00:00:00 /usr/bin/python /usr/local/bin/nova-api ash 12018 11999 0 13:00 pts/10 00:00:00 /usr/bin/python /usr/local/bin/nova-api ash 12019 11999 0 13:00 pts/10 00:00:00 /usr/bin/python /usr/local/bin/nova-api ash 12245 8600 0 13:10 pts/32 00:00:00 grep --color=auto nova-api
5. Sent SIGHUP to the parent process:
ash@ubuntu:~/devstack$ kill -HUP 11999
6. Checked nova-api's logs: http://paste.openstack.org/show/195347/
Please note, that the parent process (pid = 11999) didn't call reset().
The situation is the same for all services that use ProcessLauncher including neutron. Please, correct me, if there is something I'm missing.
I have tested nova-api master too (as well as neutron) the following way:
1. Added logging to nova.service.reset method:
ash@ubuntu: /opt/stack/ nova$ git diff object) :
diff --git a/nova/service.py b/nova/service.py
index 3ae307c..5f4e952 100644
--- a/nova/service.py
+++ b/nova/service.py
@@ -366,6 +366,7 @@ class WSGIService(
:returns: None
""" "Calling reset!")
self. server. reset()
+ LOG.exception(
def _get_manager(self):
2. Changed logging format in nova.conf to show pids
3. Started nova-api in daemon mode with the default number of workers (metadata_workers = 2, ec2_workers = 2, osapi_compute_ workers = 2)
4. Verified that all processes are running:
ash@ubuntu: ~/devstack$ ps -ef | grep nova-api bin/nova- api bin/nova- api bin/nova- api bin/nova- api bin/nova- api bin/nova- api bin/nova- api
ash 11999 9714 0 13:00 pts/10 00:00:04 /usr/bin/python /usr/local/
ash 12008 11999 0 13:00 pts/10 00:00:00 /usr/bin/python /usr/local/
ash 12009 11999 0 13:00 pts/10 00:00:00 /usr/bin/python /usr/local/
ash 12010 11999 0 13:00 pts/10 00:00:00 /usr/bin/python /usr/local/
ash 12011 11999 0 13:00 pts/10 00:00:00 /usr/bin/python /usr/local/
ash 12018 11999 0 13:00 pts/10 00:00:00 /usr/bin/python /usr/local/
ash 12019 11999 0 13:00 pts/10 00:00:00 /usr/bin/python /usr/local/
ash 12245 8600 0 13:10 pts/32 00:00:00 grep --color=auto nova-api
5. Sent SIGHUP to the parent process:
ash@ubuntu: ~/devstack$ kill -HUP 11999
6. Checked nova-api's logs: http:// paste.openstack .org/show/ 195347/
Please note, that the parent process (pid = 11999) didn't call reset().
The situation is the same for all services that use ProcessLauncher including neutron. Please, correct me, if there is something I'm missing.